Wikipedia:Administrators' noticeboard/Incidents/Rich Farmbrough/January 2009-September 2010

A few days ago Rich started using his bot, User:SmackBot, to add {{Ibid|date=January 2009}} to articles. As you make mass automated edits which are bad, in lieu of a bot reverting them, no human can be expected to use anything but rollback nor can see from his smackbot's talk page, thanks to me as he blanks it before discussions end [1], at least 6 users came to complain about it. I used rollback to revert this, much to my own displeasure (as I had intended to use my wiki time yesterday to write articles) and I thought it would end there. Today however Rich, as User:Rich Farmbrough, is rollbacking all the reverts of this tag made yesterday by me. By the looks of my talk page, he is using the fact that I got a few non {{Ibid|date=January 2009}}s as an excuse to do this (such as changing ref code to {{Reflist}}; the alternative was me previewing each revert, which would be ridiculous given the numbers), when he knows fine well he can have his bot do this. Deacon of Pndapetzim (Talk) 16:53, 22 January 2009 (UTC)[reply]

It should also be noted that SmackBot (BRFA · contribs · actions log · block log · flag log · user rights) doesn't appear to have the mandate to unilaterally add this tag, only to date it if it were added by a human. Correct me if I'm wrong. Furthermore that Rich Farmbrough (talk · contribs) seemed to use some kind of mass-rollback script to complete this at an alarming epm rate for someone without a bot flag. –xeno (talk) 16:59, 22 January 2009 (UTC)[reply]

Please note that Deacon rollbacked these edits without discussing with me. It is not an issue that he has reverted several hundred recent edits that are not those he is targeting (although they have a different edit summary) - recent stuff can by and large be redone - the problem with rollback is it undoes all the consecutive edits by that user to the article. So for example, edits the bot made in 2006, using code which will no longer run could be reverted. Adn there is no way to know which articles this applies to. Had deacon come to my talk page as clearly requested on the bot's talk page and discussed the matter there, we could have avoided a lot of work for both of us. I have rollbacked as much of Deacons rollbacks as I can, and am re-applying the removal of the template he finds so disquieting. I will be left with probably several hundred articles to go and check the history of manually. Deacon, you really needed to talk to me about this, rather than just apply rollback which is for anti-vandalism purposes only. Rich Farmbrough, 17:43 22 January 2009 (UTC).

Sorry but I find it highly unlikely that an edit the bot made in 2006 has not been followed up by another editor in the meantime. –xeno (talk) 17:49, 22 January 2009 (UTC)[reply]

Xeno, he was [ab]using automated tools to add a controversial tags to thousands of articles, and now, after spending most of his day rolling back the edits of another user, he's trying to complain that another user used rollback to attempt to undo his efforts. Please, Rich, just remove this tag from all these article and use your bot more wisely in future. I don't want to have to spend another evening trying to fix your bot's mess. Incidentally, I did put a notice on the bot's talk, and all I got was you blanking my comment along with 5 others who agreed with me. Thanks, Deacon of Pndapetzim (Talk) 17:51, 22 January 2009 (UTC)[reply]

Yep, I was just pointing out that his objection to the use of rollback ("...you may have rolled back more than the ibid edit, like an edit from 2006") didn't really hold much water. I agree that the bot ought stick to tasks it has been approved for and are inline with current consensus. –xeno (talk) 17:59, 22 January 2009 (UTC)[reply]

What tags does that concern? I've seen the bot add recently the "ibid., op.cit. and so an are bad" tags with which I agree... --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:01, 22 January 2009 (UTC)[reply]

As far as I know, the bot has not been approved for adding tags, merely dating them. –xeno (talk) 18:02, 22 January 2009 (UTC)[reply]

Xeno is correct. It's not a question of whether it's correct or not to add those tags but whether the bot is allowed to run a task that it was not approved for. I, too, think it is not and Deacon is correct to complain about it. Rich should have gotten approval first... SoWhy 18:06, 22 January 2009 (UTC)[reply]

I was one of the people who contacted Rich about the tags. He responded extremely promptly and constructively and I considered the matter resolved. His changes for future runs are listed in this version of the talk page. [2]. It's hardly his fault that WP:FN says not to use ibid. in footnotes... — Carl (CBM · talk) 18:09, 22 January 2009 (UTC)[reply]

Lucky you. I and other just got blanked (contrary to Wikipedia:BOT#Good_communication). Whatever WP:FN says this day of the week, and read its talk page too, this has nothing to do with the issues here. Regards, Deacon of Pndapetzim (Talk) 18:17, 22 January 2009 (UTC)[reply]

YOU did not get blanked, don't be disengenous. I left a mesasge on your talk page. The bot's talk page is for stopping the bot. It is very clearly stated there that that is how it works. There was also a clear message specifically relating to this issue asking for messages to go to my talk page. The bot's talk page is archived as soon as I find messages on it - again clearly stated on the talk page. Rich Farmbrough, 19:37 22 January 2009 (UTC).

The bottom line is that he should file another BRFA if he feels there is consensus for this task. –xeno (talk) 18:15, 22 January 2009 (UTC)[reply]

I guess I could hunt this information down myself, but I'm sure others in this thread already know the answers, I'll just ask.

1. Rich: was this task approved by BAG?

   I did this as part of general tag/fixup work. Rich Farmbrough, 19:37 22 January 2009 (UTC).

   So this template is getting automatically put in due to "general fix options" in AWB? If so that should probably be disabled while the template's usage is in contention. If its something you programmed in yourself, it is stretching the boundaries of what the bot was approved for. –xeno (talk) 19:45, 22 January 2009 (UTC)[reply]

2. Deacon: how long did you wait for a reply to your question before you started rolling back?
3. Deacon: is there a deadline I don't know about? Rather than waste hours of your time rolling back edits, why not discuss with Rich, and see if he could undo things with his Bot instead? In particular, did your rollbakcs come before or after the talk page blanking that distressed you so much?
4. Rich: Do I understand correctly that you have either completed, or are in the process of, going back to the status quo before the Ibid run? That this consists of rolling back Deacon, then undoing the addition of the tag?

   Yes perfectly correct. Rich Farmbrough, 19:37 22 January 2009 (UTC).

5. Deacon: Assuming I'm correct, did you understand that this is what Rich is doing right now?
6. Rich: did you explain to Deacon what you were doing, so he wouldn't misinterpret your rollbacks?

   I left him a message, explaining that is was going to be hard to fix up his rollbacks, and expressing a desire that he would have talked to me about it. Rich Farmbrough, 19:37 22 January 2009 (UTC).

7. Both: Is it really true you each think that the blame for this solely rests with the other?
8. All: Am I the only one who detects a surprising lack of assumption of good faith from both users, people that have each been here long enough to know better?

   I hope I AGF, but the rollback tool is dangerous, and a blunt instrument. Talking is far more useful, because it is (was) easy for me to revert those pages accurately, now will take me some considerable time. Also all the edits had distinct edit summaries, so the mass rollback including edits which has quite different summaries seems to be taking the blunt instrument and using it with a blindfold. Rich Farmbrough, 19:37 22 January 2009 (UTC).

thanks. --barneca (talk) 18:27, 22 January 2009 (UTC)[reply]

All I can say is that it is making my watchlists almost useless right now and I'm fed up with it. dougweller (talk) 18:35, 22 January 2009 (UTC)[reply]

@ Barneca, thanks for your mediation attempt, but this isn't a personal conflict; it will be resolved if and only if Rich/SmackBot reverts these edits (looks like this might happen) and uses his bot properly in future. There should never be a question of a bot adding controversial tags. These edits were started at least 4 days ago (that's how far back I saw him doing this), the reversions [by me] started earlier today/yesterday (though a bunch of ad hoc reversions were occurring before this). Rich was not responding, and only eventual response was to blank the page [really really helpful considering others may have had the same issue]. Smackbot doesn't keep talk archives so that's another cause for concern. Who knows how many problems like this have occurred. Regards, Deacon of Pndapetzim (Talk) 18:43, 22 January 2009 (UTC)[reply]

It does keep talk archives. There is a pointer on the talk page. There was a message on your talk page and I replied to others too. Furthermore as you know full well having done the mass rollback, there was only one run, so again saying "started 4 days ago" smacks of disingenuity. Rich Farmbrough, 19:37 22 January 2009 (UTC).

I keep both SmackBot and Rich's pages watchlisted. Rich responds promptly and politely and is very easy to work with. I believe every concern I have brought to his attention (4 or 5 at least), have been handled this way, even though all (or at least most) of my concerns were completely unfounded. SmackBot's talk page is very easy to use: leave a message on the page itself to halt the bot, leave a message for its operator on the operator's page. SmackBot often is the only editor on pages for years at a time in my experience. A lot of cruft was created a long time ago and marked for cleanup, and SmackBot adjusts the cleanup messages every few years as the standards change. I only see such changes because I have entire categories watch-listed. I don't think SmackBot should add the {{ibid}} tag, the conensus seems to be that he should not add the tag, and SmackBot is now fixing that mistake. JackSchmidt (talk) 19:06, 22 January 2009 (UTC)[reply]

Summary, from my point of view. This is a supposed to be a collaboration, while this may not have met with universal approval :), it would have been better to discuss and find a solution rather than engage on such a massive revert, Deacon's time and mine would both have been saved. Futhermore bringing the matter here just wasted more time and energy. Lets get on with the project and leave this behind. Rich Farmbrough, 19:37 22 January 2009 (UTC).

This thread appears to hinge on barneca's #6 above. Now that we know that Rich is removing the tags and the rollback was a precursor to that, I gather we can mark this as resolved. –xeno (talk) 19:16, 22 January 2009 (UTC)[reply]

Rich, sorry I didn't see that bot kept an archive. My mistake. Though you shouldn't be removing posts so quickly. They aren't just messages to your bot, but also in this case public discussions. That I decided to take reverting your disruptive automated edits out of your hands was down to the ill-considered nature of the edits and your unresponsiveness, ignoring the complaints being a reasonable sign of uncooperativeness. That you subsequently decided to roll them all back with your own account was bad [besides being adversarial and immature, it also added to the server space and watchlist issues unnecessarily as well as violating rollback policy], but then to proceed to complain about rollback afterwards was really unconvincing and incidentally, since everyone here is an intelligent adult, counter-productive as it is obviously hypocritical. If you were planning before this ANI thread to remove all these tags, I'd be delighted to know the "technical" reason why rollback with a non-bot admin account was good first stage ... not communicating your intention being just the cherry on top. In any case, if you make mass automated edits which are bad, in lieu of a bot reverting them, no human charged with reversing their impact can be expected to use anything but rollback nor to care if he removes a few minor decent fixes in the process. Anyways, it's nothing personal and I don't mind you wanting to put a good face on it as long as you carry out your self-reversions and don't do it again. Regards, Deacon of Pndapetzim (Talk) 02:46, 23 January 2009 (UTC)[reply]

And just out of curiosity, Rich, why did you continue using automated tool to place the tag when User:Dbachmann had already complained about it at Template_talk:Ibid, a discussion you saw? Deacon of Pndapetzim (Talk) 03:09, 23 January 2009 (UTC)[reply]

• When they showed up on my watchlist (mostly opera and classical music), I sometimes moved them to the References section, if there were several ibids/op cits to fix. Otherwise I just fixed them before removing the tag. Just as a demonstration of how completely unsuitable this tag is for a bot to apply, it was plastered at the top of (and thereby defaced in my view) a featured article Dmitri Shostakovich when there was a only one instance of an op cit. out of 48 footnotes. Voceditenore (talk) 16:41, 23 January 2009 (UTC)[reply]

• There are still a fair number of articles that smackbot applied this tag to that have not been removed: [3] . –xeno (talk) 16:10, 23 January 2009 (UTC)[reply]

  I went ahead and removed these tags. –xeno (talk) 03:03, 24 January 2009 (UTC)[reply]

Rich Farmbrough (talk · contribs · blocks · protections · deletions · page moves · rights · RfA) is still at it, changing categories out-of-process without consensus. He not only should know better, he's already participated in the earlier complaints. What can be done?

Probably missing some, where he fails to provide an edit summary:

1. 2009-05-24T00:03:00 Template:Failed verification (edit | talk | history | links | watch | logs) (top)
2. 2009-05-24T00:08:47 Template:Original research (edit | talk | history | links | watch | logs) (since=>from and simplify) (top)
3. 2009-05-24T00:13:13 Template:Or (edit | talk | history | links | watch | logs) (top)
4. 2009-05-24T20:44:21 Template:Expand-section (edit | talk | history | links | watch | logs)
5. 2009-05-24T21:09:55 Template:Article issues (edit | talk | history | links | watch | logs) (since => from)
6. 2009-05-24T22:15:51 Template:Mergefrom (edit | talk | history | links | watch | logs) (since=> from) (top)
7. 2009-05-24T22:18:34 Template:Mergeto (edit | talk | history | links | watch | logs) (since=> from : rmeover the category parameter: not likely to be used in article space.) (top)
8. 2009-05-24T22:24:33 Template:Mergefrom-multiple (edit | talk | history | links | watch | logs) (clean up using AWB) (top)
9. 2009-05-24T22:25:04 Template:Merge JRRT (edit | talk | history | links | watch | logs) (Since => from using AWB) (top)
10. 2009-05-24T22:25:42 Template:Merge FJC (edit | talk | history | links | watch | logs) (Since => from, Replaced: since → from, using AWB) (top)
11. 2009-05-24T22:26:38 Template:Merge-school (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
12. 2009-05-24T22:27:07 Template:Portalmerge (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
13. 2009-05-24T22:27:23 Template:NorthAmMergeto (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
14. 2009-05-24T22:27:40 Template:Multiplemergefrom (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
15. 2009-05-24T22:28:56 Template:Merging (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
16. 2009-05-24T22:29:15 Template:Mergetomultiple-with (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
17. 2009-05-24T22:30:32 Template:Mergeto2 (edit | talk | history | links | watch | logs) (Since => from or other cleanup using AWB) (top)
18. 2009-05-24T22:30:41 Template:Mergeto-multiple (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
19. 2009-05-24T22:30:55 Template:Mergesections (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
20. 2009-05-24T22:31:05 Template:Mergesection (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
21. 2009-05-24T22:32:35 Template:Mergefrom-category (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
22. 2009-05-24T22:32:55 Template:Merge-multiple-to (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
23. 2009-05-24T22:33:00 Template:Merge-multiple (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
24. 2009-05-24T22:33:06 Template:Merge (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
25. 2009-05-24T22:33:19 Template:Afd-mergeto (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
26. 2009-05-24T22:33:25 Template:Afd-mergefrom (edit | talk | history | links | watch | logs) (Since => from or other cleanup, Replaced: since → from, using AWB) (top)
27. 2009-05-24T22:34:57 Template:Expert-verify (edit | talk | history | links | watch | logs)
28. 2009-05-24T22:35:22 Template:Expert-verify (edit | talk | history | links | watch | logs) (top)
29. 2009-05-24T22:36:25 Template:Expert-subject (edit | talk | history | links | watch | logs) (top)
30. 2009-05-24T22:48:35 Template:Tdeprecated (edit | talk | history | links | watch | logs) (since=> from. Stop self include.)
31. 2009-05-24T22:49:52 Template:Tdeprecated (edit | talk | history | links | watch | logs) (top)

--William Allen Simpson (talk) 08:56, 25 May 2009 (UTC)[reply]

I reverted all of those edits (and probably a few more related edits in sequence, which may have been sensible.) — Arthur Rubin (talk) 10:03, 25 May 2009 (UTC)[reply]

These were trivial edits, changing only "since" to "from", standardising all 42 Wikipedia maintenance categories involving dated categories. Making changes is not sensible. Why are you undoing another admins actions for no good reason? Debresser (talk) 11:37, 25 May 2009 (UTC)[reply]

Arthur Rubin wrote Rich to his talk page here. But didn't await his reply or actions. And see my reply there that Arthur Rubin was non-specific and did not take the most logical course of action. Debresser (talk) 12:06, 25 May 2009 (UTC)[reply]

See also above and this diff, where Arthur Rubin admits he might have reverted some sensible changes. Debresser (talk) 12:24, 25 May 2009 (UTC)[reply]

People should know better than to respond to these sort of trollings. The reversion breaks maybe a zillion articles, maybe two zillion. Rich Farmbrough, 13:36 25 May 2009 (UTC).

People should know better than to change a template/category pattern without discussing it in the relevant WikiProject or on TfD or CfD. As I pointed on in a smaller rename (about 38 decade names), we need to make sure that all the links are done correctly. — Arthur Rubin (talk) 14:42, 25 May 2009 (UTC)[reply]

Since either "since" needed to be changed to "from" or "from" to "since", both causing a significant disruption, you should have discussed which it was to be before making the changes. I don't think I have the tools you constructed to reduce the auxilliary errors caused by the process. I suppose, at this point, the good of Wikipedia suggests I allow to continue as you wish, as I don't know how else to mitigate the damage you caused.

For the most part, From is just wrong. I suppose I had better revert my corrections, as I can't figure out else to repair the damage. May I suggest that you rename all the generated categories to "since YYYY-MM", as that makes automated processing easier? — Arthur Rubin (talk) 14:48, 25 May 2009 (UTC)[reply]

See Wikipedia_talk:Categories_for_discussion#More_out_of_process_category_renames for linguistic arguments to choose "from" rather than "since". But that can be discussed and taken care of later. When we had 32 categories at "from" and only 9 at "since" (and 1 at "as of") the obvious choice was to go to "from". Debresser (talk) 15:21, 25 May 2009 (UTC)[reply]

• The above demonstrate at least 30 changes in 1 day "since" to "from" — your count must be inaccurate. Perhaps you meant 32 at "since" and only 9 at "from".
  --William Allen Simpson (talk) 19:38, 25 May 2009 (UTC)[reply]

Actually, changing all categories to another format will be a lot easier after Rich finishes. His edits are well though through and take care of all loose ends. See Category:Articles with invalid date parameter in template which started geting crowded right after Arthur Rubin's actions and is now again depopulated. Debresser (talk) 15:33, 25 May 2009 (UTC)[reply]

• Points to Rich Farmbrough for being bold on this one, but since it's obvious that these changes have encountered opposition, he should stop his unilateral changes and submit them to CFD, which is the process established for exactly this purpose. I have no opinion on whether any of his changes are actually a good idea, but wholesale changes of this sort generally set somebody's nose out of joint. Best to send it through the process created for the purpose.--Aervanath (talk) 16:45, 25 May 2009 (UTC)[reply]

• Farmbrough deliberately defied the folks at WT:CFD#CfD categories renamed, where similarly bad edits were previously discussed.
  --William Allen Simpson (talk) 19:38, 25 May 2009 (UTC)[reply]

Not convinced of the need of process as the 32 vs 9 argument goes well with WP:IAR in particular as the discussion is under discussion and a final resting place can easier be dug with those changes already in place. Agathoclea (talk) 17:27, 25 May 2009 (UTC)[reply]

• The above demonstrate at least 30 changes in 1 day — "since" to "from" — your count must be inaccurate. WP:IAR is inapplicable, as Farmbrough's edits were not improving or maintaining anything. Indeed, as Arthur Rubin learned, Farmbrough actually made it difficult to revert, an essential maintenance function.
  --William Allen Simpson (talk) 19:38, 25 May 2009 (UTC)[reply]

1. I said 32 categories, not templates.
2. As I said before, it will be easy to make changes after you let Rich finish. Debresser (talk) 19:52, 25 May 2009 (UTC)[reply]

Who are "the folks at WT:CFD#CfD categories renamed"? Is that some Wikipedia subgroup with special rights somewhere? Trying to own part of Wikipedia? Or is that you and me and Rich and anybody else who wants to contribute to Wikipedia in good faith? Debresser (talk) 19:55, 25 May 2009 (UTC)[reply]

Of course it means "everybody who wants to contribute to Wikipedia in good faith"...which means consulting other editors and being willing to seek consensus for your actions once objections have become known.--Aervanath (talk) 07:19, 26 May 2009 (UTC)[reply]

You and me know that, but do all editors involved in this discussion know that? I have a reasonable doubt as to that. Debresser (talk) 09:50, 26 May 2009 (UTC)[reply]

So if we agree on this, why don't we agree that Rich should stop and let his changes be discussed more thoroughly with all interested parties before he continues with the wholesale changes?--Aervanath (talk) 16:55, 26 May 2009 (UTC)[reply]

See e.g. Wikipedia:Categories_for_discussion/Log/2009_May_26#Category:Pages_for_deletion where William Allen Simpson uses language such as "Obviously, we decided by consensus" (without any reference to that discussion, btw). Debresser (talk) 13:59, 26 May 2009 (UTC)[reply]

Actually, in that discussion he does reference a discussion from 2006 in the nomination.--Aervanath (talk) 16:55, 26 May 2009 (UTC)[reply]

Thank you. My fault completely. You might be interested in Wikipedia:Wikiquette_alerts#User:William_Allen_Simpson. Debresser (talk) 19:12, 26 May 2009 (UTC)[reply]

As I have pointed out a couple of weeks ago, some users have been using bots and scripts to impose their own preferred style of referencing, the "named" references, on articles previously not using it.

This system (the same footnote re-used again and again) is common in some fields and used by many science journals. It is, however, absolutely non-standard in the humanities. Many contributors, not just me, do not like it and do not want to have this system imposed on all articles.

See discussion at Wikipedia talk:Citing sources#Replacing duplicate footnotes with named footnotes. User:Postdlf makes a good job there of summarizing the reasons against this style in three points. A fourth point is that of usability. Named refs makes references dependent on each other, which makes it more cumbersome to edit them, for instance to correct a page or page range, to add an additional source with a contrasting view, or to clarify how a reference supports the claim made. This point was touched on by me, earlier in that discussion, and expressed very clearly by User:Golbez in a previous (now archived) discussion (from July 2009).

The article Charles Boit, which I used as an example, had at that point been hit three times by this:

• First (30 July 2009) by User:Falcon8765, using a script called AWB.
• Second (10 August 2009) by User:DrilBot.
• Third (16 September 2009) by User:SmackBot.

I reverted this every time.

• It has now been hit again, a fourth time, again by SmackBot.

SmackBot, or rather its keeper, User:Rich Farmbrough, has previously been warned by the administrator User:CBM for this behaviour. CBM blocked SmackBot, then unblocked it on the condition that the feature was disabled. Rich Farmbrough agreed to this. (See edit link earlier in paragraph, it's all there.)

Thanks for your attention. --Hegvald (talk) 10:01, 5 December 2009 (UTC)[reply]

I have blocked temporarily again, and will unblock again once this is fixed. — Carl (CBM · talk) 13:11, 5 December 2009 (UTC)[reply]

Considering SmackBot/RF has already been warned about this and that these edits were never appropriate to begin with (as there is no general agreement that named refs are better), it would only be appropriate for SmackBot to be given the task of reverting its own previous edits. Who else is going to do this? --Hegvald (talk) 17:11, 5 December 2009 (UTC)[reply]

I don't suppose User:SmackBot/References Log Log of ref runs would be of any assist as evidence?

When I've been hit by this it's made me assume I'd been a lazy/awful/terrible editor for non memorizing every last work of article guidelines. If I think that way, who knows how many others have been discouraged? This has covered an insane number of articles and as far as we know it could have started edit wars from article creators... especially since the edit summaries given have nothing to do with what was changed. ♪ daTheisen(talk) 18:32, 5 December 2009 (UTC)[reply]

It should rather make you think "Thank goodness I don't have to worry about the niff-naff and trivia". Rich Farmbrough, 20:50, 6 December 2009 (UTC).[reply]

These bot edits of yours have had the misleading edit summary "general fixes". The problem here is that your way of "fixing" what you call the "niff-naff and trivia" results in a referencing style that is non-standard for many contributors and contrary to the way they are used to work, and want to continue working, with footnotes and references. While you may think that you are just polishing the formatting of these articles, you actually create an editing environment that is going to discourage some contributors from doing any additional work on these articles. --Hegvald (talk) 01:09, 7 December 2009 (UTC)[reply]

I've unarchived this, it's only supposed to be archived after 24 hours of inactivity but the bot is doing it after 18. I'm also unhappy about this and find named references often a pain. It needs more discussion. Dougweller (talk) 18:12, 6 December 2009 (UTC)[reply]

For what it's worth, I've changed archiving on this page back to 24 hours; I couldn't find any discussion of the change to faster archiving, and anything less than 24 hours risks missing input from those far off the most active time zones. — Gavia immer (talk) 18:26, 6 December 2009 (UTC)[reply]

I'm not sure what more there is do here. I'm informed User:Rich Farmbrough in case he wasn't aware that SmackBot had been blocked. As to the references style, this is ANI. Village pump, MOS, (particularly Wikipedia:Citing sources) are better places for that discussion than here. Like the British/American spelling disputes we sometimes see, I don't think this is really resolvable. Current policy is "follow the style already established in an article, if it has one; where there is disagreement, the style used by the first editor to use one should be respected" but people just have different preferences. -- Ricky81682 (talk) 20:02, 6 December 2009 (UTC)[reply]

• But note that while editors making changes can do so across a reasonable spectrum of interpretation, bots cannot. Protonk (talk) 20:15, 6 December 2009 (UTC)[reply]
• Well, if Rich has been told before, and can't seem to run his bot appropriately without causing disruption, perhaps he should have to go through another bot approval process before he's allowed to run it again.--Crossmr (talk) 00:31, 7 December 2009 (UTC)[reply]
  • It appears to have been a good faith error and minor in the context of the number of edits the bot is performing. The damage done is not too significant. The best way forward might be to approach the AWB project about the fact that this is (incorrectly) classified as a general fix, which is the root of the problem. Christopher Parham (talk) 02:35, 7 December 2009 (UTC)[reply]
    • was this or was this not brought up before? Rich is an experienced editor and bot owner. If this was brought up before and decided to be an inappropriate change for his bot to make, I would expect him to no longer make those changes.--Crossmr (talk) 05:09, 7 December 2009 (UTC)[reply]
      • I don't think that's a reasonable expectation - people make mistakes, including bot operators. One mistake in the context of 2 million edits is not surprising. It is reasonable to expect Rich not to deliberately make those changes, but there is no indication that this was anything but an accident. Christopher Parham (talk) 14:53, 7 December 2009 (UTC)[reply]

Trouble is, this isn't one mistake. RF has been asked before not to use his bots to change the formatting of references - Wikipedia supports more than one style of referencing, and there is no obligation to use the bot's style. It's not the bot's error - it's the programmer's error.Elen of the Roads (talk) 18:17, 7 December 2009 (UTC)[reply]

I have to repeat my question: Shouldn't SmackBot now be given the task of reverting its own previous edits? To clean up after a bot you need a bot, or it will take a week to revert what the bot did in a couple of hours. --Hegvald (talk) 01:09, 7 December 2009 (UTC)[reply]

Reverting it would practically be impossible at this point, unless the magic feature listed at meta could handle it. Given it's been spread out and many articles changed since I have no clue.with the personal style of one user poisoning it. It's arguably the largest possible case of vandalism since it's deliberately removing content and replacing it with a personal POV with a deliberately misleading edit summary on top and in practically a hidden manner. Very depressing form of WP:OWN, but will probably have to be left at never permitting it again unless the bot can be given orders to only act places where it's still top/previous change, assuming compliance. If an editor chooses a certain method of style, it's disruptive to change it, period. Bot operator needs to form a community consensus on the sole form of referencing if such views are so fervently held, as was mentioned above.♪ daTheisen(talk) 21:48, 7 December 2009 (UTC)[reply]

Excuse me, I don't see any evidence above of SmackBot removing content. Would you please either supply a diff or withdraw the accusation of vandalism? Thanks. --SarekOfVulcan (talk) 21:55, 7 December 2009 (UTC)[reply]

It's not removing content (Datheisen seems to have got a bit carried away here I agree) but if someone repeatedly replaced the reference style in use in the article with a different style, against consensus and having been asked to stop; after the fourth time I think even you might term it vandalism. Elen of the Roads (talk) 16:15, 8 December 2009 (UTC)[reply]

Poor wording. It's removing formatting (only "content in terms of characters, not actual article substance which is the most basic form of what we call "content"), but it's still removing what someone entered with their own two hands and replacing it, and my resentment was surrounded on that. ... apologies, please. ... None to the any bot again seen trying this. Right. Head into desk repeatedly time. ♪ daTheisen(talk) 03:12, 9 December 2009 (UTC)[reply]

If it has really happened this much I can't see any reason Rich shouldn't have to go through the bot approval process again before being allowed to run Smack bot.--Crossmr (talk) 21:06, 8 December 2009 (UTC)[reply]

Since the discussion with CBM it happened that I made a short run using some old settings. It is probable that no more than a handful of articles, if that, were actually changed the way described. The only visible change in the article in question would have been in superscripts and the reduction of the notes from 23 to [16]. Given the number of articles that probably have had names added by various AWB users, and that the only complaint has been from one user on one article (although of course it would be that one that got changed again) , your suggestion seems a little over the top. Rich Farmbrough, 23:27, 8 December 2009 (UTC).[reply]

SmackBot is the highest profile user of AWB, but certainly there are a lot of others going around doing the same. If there are other bots, they should also be blocked until the problematic behavior is corrected. Non-bot users are harder to deal with, but a more stringent line could convince the writers of AWB to bring their software into compliance with the relevant guidelines. Christopher Parham (talk) 23:41, 8 December 2009 (UTC)[reply]

As far as that goes I had put a feature request in to avoid applying this change in cases such as the above example, but I have not been following it. The discussion above suggests that citation style has been changed, and this is a hot-button issue. Of course that is not what is meant by the word style in the guidelines, that is the five citation styles. GRB 970508 uses the shortened footnote style, just as the article above, and uses de-duplicated named references. The citation style is the same, simply the repetitive columns of "Schilling 2002 p126" have been reduced, in that case from 53 entries to 34. If this were paper it would be a different matter, there would be a finger or ruler following the cites, and a another in the end matter, while the feet were used to control a book wheel (do Amazon do those?). Rich Farmbrough, 00:45, 9 December 2009 (UTC).[reply]

Presenting the same information to the reader in two different ways constitutes a change in style, for the purposes of WP:CITE. The "general fix" in AWB encourages users to falsely believe that there is consensus that named references are preferred over sequentially numbered footnotes. Christopher Parham (talk) 13:29, 9 December 2009 (UTC)[reply]

(unindent) rev 5730 Change AWB genfixes not to implement named references for an article if there are currently no named references in the article. This should appease editors who believe this constitutes a change of reference style, hence contravening WP:CITE rules. Rjwilmsi 16:06, 9 December 2009 (UTC)[reply]

Over the last 8 months, Rich Farmbrough (talk · contribs · blocks · protections · deletions · page moves · rights · RfA) has made approximately 400,000 edits [4], rocketing themselves to the setting themself high atop of the List of Wikipedians by number of edits. The vast majority of these edits were done with AutoWikiBrowser and furthermore, many were primarily cosmetic. Irrespective of that, they are seemingly using some kind of macro or modified AWB build to automate the saving process unless we are to believe that they really are sitting in front of their PC for several-hour-long stretches hitting "save" at close to 30 edits per minute [5].

Most recently, they had been running a task that made three edits in sequence to articles in order to complete a task that could conceivably be done in a single edit. [6] Had the proper steps been taken and Bot Approvals Group was engaged, they surely would've denied a bot using such an inefficient method of editing. I note Rich has since discovered a workaround for this issue after my prodding.

My concern is that the user is blatantly flouting the WP:AWB#Rules of use and furthermore flooding recent changes with tasks that really should be run with a bot flag (if at all). Upon querying the user, their stated reason for running bot tasks from their main account was uncompelling [7] and they resumed the task from their main account.

I invite additional scrutiny and advice as to how to convince or compel this user to respect Wikipedia:Bot policy and AWB's rules of use. –xeno^talk 17:56, 31 December 2009 (UTC)[reply]

Hardly in the last 8 months. There's a serious amount of low level dirt that needs cleaning, BRFA is seriously slow, I prefer to fix a problem rather than file a request for someone else to do it. Wiki - if anyone had forgotten - means quick. Rich Farmbrough, 18:05, 31 December 2009 (UTC).[reply]

I'm not asking you to ask someone else to do it, I'm asking you to ask BAG for approval to do it from a bot account rather than your main account. The tasks you are running are not in any way mission-critical such that you can't wait a few days or a week for BAG approval. –xeno^talk 18:06, 31 December 2009 (UTC)[reply]

You can't just ignore bot policy because WP:BRFA "is seriously slow". Singularity42 (talk) 18:08, 31 December 2009 (UTC)[reply]

I have a lot of respect for Rich and he does a lot of excellent work, however, as an occasional recent changes patroller, I can tell you the AWB edits do flood the recent changes- he can be in the same list 4 or 5 or even more times and the standard 50 edit display only shows the last 1 or 2 minutes on a normal day (perhaps someone can give a statistic for how many edits are made a minute to the whole wiki?). I have to say, it would be preferable for these edits to be made from a separate, bot account, though they do need to be made. HJMitchell You rang? 18:16, 31 December 2009 (UTC)[reply]

There are roughly 12.57 every second. This is the total number of edits divided by the total number of seconds that Wikipedia has existed (I rounded up to 9 years here). I agree with Xeno though in that Rich has way to many edits. He has already opted out of the list of edits here, and until this recent dump, his name was on the top. He has 250,000 more edits than the guy right behind him as well. I don't see anything wrong with a ton of edits, but the flooding of recent changes and other things is a concern. Also, he practically is a bot with the number of edits that he has. I don't even know of a bot that can make over 106,000 edits in one month, as he did this past month. Kevin Rutherford (talk) 18:55, 31 December 2009 (UTC)[reply]

OK guys point made. Rich Farmbrough, 19:28, 31 December 2009 (UTC).[reply]

"My concern is that the user is blatantly flouting the WP:AWB#Rules of use ". Darn this guy is doing a good thing. He makes good edits that nobody else can be bothered to make!!. Guys we shouldn't be sitting around moaning about Rich. His edits in my book have been of great help to this project and I seem no harm in him using his own account to do so. I think this is a case of jealousy over his edit count above anything else. Who cares about edit count? What the frick does it matter whether he notches up 400,000 edits with his own account or under a different account? The same tasks will still need doing either way so who cares? There are far more serious things to be worrying about than Rich Farm having a high edit count LOL!! I will always support whatever Rich wants to do, we should be thanking him for his dedication to making such repetitive edits, he doesn't have to bother making a single edit on here. If you add um the sum of his edits this year they have made a massive difference to the encyclopedia in terms of considtency, formatting and cosmetics. He is clearly content to do so using his own account so why stop him. Have a great new year Rich I appreciate every edit you make even if these people don't. Above all is this ANI report really necessary? I strongly dislike the way such decent editors get reported here like vandals. Dr. Blofeld ^{White cat} 19:40, 31 December 2009 (UTC)[reply]

Please do try to assume good faith. The problem is clearly stated and has nothing to do with jealousy or relative edit counts, but with the reason for the edit count: namely, running unapproved bot tasks from a main account contrary to bot policy and AWB rules of use. I'm not saying the edits aren't useful, but they should ultimately be done from a approved, flagged bot account. The ANI thread was necessary because I've brought this up to Rich a couple times, and it did not appear that he was going to modify his approach. I've given him a suggestion as to how he could have a fairly open-ended bot approved for tasks like these and hopefully this can now be marked {{resolved}}. Thanks for your input. –xeno^talk 19:44, 31 December 2009 (UTC)[reply]

I did assume good faith but the way you started your post saying "he has made approximately 400,000 edits, rocketing themselves to the top of the List of Wikipedians by number of edits" makes it look as if you resent this fact and the way he achieved it... Dr. Blofeld ^{White cat} 19:53, 31 December 2009 (UTC)[reply]

It was more to highlight the dramatic and exponential increase in Rich's editing rate. –xeno^talk 19:56, 31 December 2009 (UTC)[reply]

Does anyone have any concern about the content of the edits? Chillum ^{(Need help? Ask me)} 19:42, 31 December 2009 (UTC)[reply]

Have you seen his bot's block log? He is prone to making controversial changes (moving tags, mucking about with named refs, etc.) which is why BAG should have oversight. –xeno^talk 19:44, 31 December 2009 (UTC)[reply]

In my experience I'd say the vast majority of his edits have run relatively smoothly and he is responsible enough to sort any major issues out. I wouldn't go as far as to say he is prone to making controversial changes, he seems to get most things right, unlike BetaCommandBot.... As for Smackbot, when it has performed the level of edits it has, one can imagine that issues crop up everynow and then... Actually I have advised Rich to run a bot instead to do certain big tasks as it would be more efficient, but he seems content to do thinks using AWB with his own account even if it takes longer. Do you think he is suffering from editcounteritis? Dr. Blofeld ^{White cat} 19:52, 31 December 2009 (UTC)[reply]

The tasks that should be done, should be done from an approved, flagged bot. The procedures in place should be followed rather than ignored without good reason. And it looks like Rich has got the message (see above at 19:28), so there's probably no need to draw this out any longer. –xeno^talk 19:54, 31 December 2009 (UTC)[reply]

• Sigh - We had a shit long discussion about this a year ago, and the community decided that it was against policy to run a bot on the main admin account, unless it was still being developed, so he is technically not following policy. However that being said, BAG has been insanely slow at approving anything, for the past few weeks/months, so I totally cannot blame him for continuing to run it. Unless you plan on getting BAG to start fixing and approving shit faster, then I recommend we allow him to continue running the bot. --Coffee // have a cup // ark // 20:01, 31 December 2009 (UTC)[reply]

  The tasks he are running are mainly cosmetic and are not time-sensitive in the least. He can wait for BAG approval just like the rest of us. –xeno^talk 20:03, 31 December 2009 (UTC)[reply]

I have to agree with Coffee the problem is that BAG is too slow to authorise things as its projetc members have many other committments on here. Given that Rich often performs a mass of different tasks every day I think it is a tall order to expect him to file requests for each one. Everybody is free to edit and do what they think is an improvement to the encyclopedia. But for the especially big runs that may be sene as "flooding the recent changes" at a rate of 20 odd edits a minute I'd say that might need some discussion if it is a prolonged activity... But to date he has made over 600,000 edits to english wikipedia and is still using the method he has used for a long time and seems to mostly be a success... Dr. Blofeld ^{White cat} 20:06, 31 December 2009 (UTC)[reply]

I see Rich is in agreement. However, I would like to point out that the general consensus with Rich's last BRFA was that it should be approved, even though not technically necessary. At that point, Rich withdrew the request. It's actually faster if you leave the request up to get approved. There's no hurry for general fixes. They can be approved in a fairly timely fashion. --IP69.226.103.13 20:16, 31 December 2009 (UTC)

I don't think proposing an "open-ended bot" is a good idea. That sounds equivalent to writing a blank check that allows the bot to do whatever Rich (in his good judgment) thinks needs to be done. That bypasses the checks-and-balances that the BAG is there for. On a different note, I'd prefer to see a reduction in unnecessary edits, such as removal of optional spaces and blank lines that has no effect on presentation, but clutters up diffs – particularly multi-version diffs where changes by other editors are obfuscated because the diff engine has difficulty matching paragraphs. Lastly, I think running an unapproved bot because the BAG is too slow to respond sets a bad precedent – particularly so when the bot behavior (partial-date delinking) id different from the bot proposal (full-date delinking). -- Tom N (tcncv) talk/contrib 20:15, 31 December 2009 (UTC)[reply]

Over the past eight months Rich has been "rocketing" to the top of the List of Wikipedians by number of edits? Ah, I think he's occupied No. 1 position for years. He has gained the immanent trust of the community. BAG should get its act into order. Tony (talk) 20:38, 31 December 2009 (UTC)[reply]

Are you sure? [8] Anyhow, this is peripheral to the true issue which I noted was the lack of a bot flag for the vast number of edits being made. –xeno^talk 01:59, 1 January 2010 (UTC)[reply]

Agreed, BAG has been less than helpful to get crap done for the past few months (other than MBisanz). It looks like we might need more people or different people. --Coffee // have a cup // ark // 20:42, 31 December 2009 (UTC)[reply]

I'd love to see more people commenting on the BRfAs and more BAG members, however, atm there's a very small number of people active there, and they seem to receive very little thanks. But just because the process takes a long time, does not mean it's okay to ignore it, as Rich appears to have done here. There are good reasons that automated editing is supposed to be done on a separate account which has the bot flag. There are also good reasons to go through the BRfA, for example, the task gets input and ideas. Rich has gone against policy, and what appears to me to be the sensible path. I don't think it's unreasonable to except Rich to always go through a BRfA, and he should do so in the future. - Kingpin¹³ (talk) 21:12, 31 December 2009 (UTC)[reply]

Obviously Rich is running what amount to multiple bots on his main account. From the comments here it appears that the bot tasks he's been running have been largely unproblematic. (Indeed, if someone had made a half-million or so controversial edits, there would already be blood on the floor.) As noted, there are a number of good reasons to want automated tasks to run under proper bot accounts, and there are also at least a few minor concerns about a few of the tasks. Since bot-flagged accounts tend to draw less scrutiny when in action, it is also proper for us to want the BAG to approve the tasks in advance. Tcncv's point about offering any editor a blank cheque is also well-taken; it wouldn't be a good idea to give Rich a free hand to run any bot task he wants, just because even the best of us occasionally make mistakes.

So, a compromise. Since it appears Rich's bot tasks have generally appeared helpful and responsible, could we get the BAG (and/or the community) to offer a blanket presumption of permission to Rich for running tasks, and only demand the full BAG process for tasks for which an explicit objection is raised? Here's how I envision the process:

1. Rich posts an 'expedited request' at Wikipedia:Bots/Requests for approval. This request will include a brief description of what task is to be performed. As part of the 'presumption of expertise', this step may also include a link to a batch of test edits (fifty perhaps?) demonstrating the principle.
2. A brief comment period will elapse. (Three days? Five days? No more than seven days, certainly. How long should it take for some reasonable number of BAGgers to skim the new request?) During this time, any editor may raise an explicit objection to the expedited process. If an objection is filed, Rich will have to go through the regular, full BRFA process.
3. If the comment period elapses without any comment (or with only positive comments) then Rich may proceed with the new bot task.
4. If any unforeseen problems arise, any member of the BAG can request the suspension of any of these 'presumed permissible' tasks pending review by the normal BRFA process.

It's nice and lightweight; it takes into account Rich's general competence; it avoids overburdening the BAG with additional paperwork; it means that there will be a second set of eyes on Rich's bot edits; and it gets these bot edits under a proper bot account. The BAG is free to add any other standard conditions they feel appropriate (edit throttles and the like) which would apply universally to all of these 'presumed permissible' requests. What do people think of that process? TenOfAllTrades(talk) 21:48, 31 December 2009 (UTC)[reply]

This sounds better than other possible ideas. Thanks for being the sane one here. Kevin Rutherford (talk) 22:00, 31 December 2009 (UTC)[reply]

I think adding a speedy approval process for proven bot operators is a good idea, but I think this would be a policy decision within the BAG, preferably with at least one BAG member granting the speedy approval. If we can't get even one BAG member to review the request (and either grant, deny, or hold for further discussion), we definitely need to increase the ranks. Requests such as SmackBot XXII should not sit for weeks at-a-time with no activity. I would further suggest that requests for speedy approvals be limited to clearly uncontroversial operations and that and have at least some independent discussion as evidence of consensus and need. -- Tom N (tcncv) talk/contrib 22:49, 31 December 2009 (UTC)[reply]

The six weeks it took for SmackBot XXII not to go anywhere is impressive, by any standards. Ohconfucius ^¡digame!

Agreed, good idea. Chillum ^{(Need help? Ask me)} 00:51, 1 January 2010 (UTC)[reply]

Sure, something like this would be fine. Anything to have those innocuous cosmetic edits start to be made with the bot flag would be an improvement. –xeno^talk 02:04, 1 January 2010 (UTC)[reply]

• I'm pleased to see this issue close to being resolved. Rich is a janitor of the first order, like a big machine gobbling up selectively all the garbage in the sea without doing harm to the ecosystem. His making scale edits in his dynamic way (when he perceives a job needs doing) is obviously counter-culture to the bureaucratic functioning of the BAG. We are all agreed that Rich is responsible and responsive with AWB, and the only concern I note from the above discussion is the potential for problems if another editor starts making non-controversial edits on a large scale. The above proposal is good, in that it streamlines the bureaucratic process and allows Rich to get on with cleaning up the crap floating in our ocean. Ohconfucius ^¡digame! 02:31, 1 January 2010 (UTC)[reply]
  • Me too. A bot should be running with a bot flag, and there seems to be general agreement that Rich does good work (with a few reservations). So there must be some way for BAG to nod this one through quickly, and I hope that the solution above is acceptable all round. --BrownHairedGirl (talk) • (contribs) 03:23, 1 January 2010 (UTC)[reply]

I'm finding this discussion about the BAG moving to slow ironic. While I do think it does drag out a bit, people need to remember that earlier this year, the BAG took quite a bit of heat for moving too fast. The criticism was they were not waiting long enough for community input. And it was over items that on the surface seemed to be just as routine and non-controversial, but turned out not to be. Well, the community got what it wanted... -- JLaTondre (talk) 03:43, 1 January 2010 (UTC)[reply]

I've tried to move Rich's tasks through as quickly as possible. But the problem has been and remains that very few people comment on BAGs. I could quite easily approve most of the pending bots, and then in 2 months get my head handed to me for approving a bot without consensus. So I wait long enough to be able to plead that it looked like silence was consensus. MBisanz ^talk 04:26, 1 January 2010 (UTC)[reply]

Yes it is true that BAG has taken heat, and also that individual members (and others) have been great with their work, whether considered opinions on fine detail, or swift approval of requests. It just happens that it is voluntary, therefore sometimes requests languish. As I say, WP:SNOW requests have been approved or approved for trial swiftly, and xeno has made some helpful suggestions. I don't think this is a big deal, maybe I should get a little more involved with BAG, and spend a little less time with reg-exes. I thank everyone for their suggestions and kind words and wish them a great new year. Rich Farmbrough, 11:07, 1 January 2010 (UTC).[reply]

Isnt it the job of the bot op to advertise the bot? Go to wikiprojects, the village pump, even (dare I say it) IRC? If the bot-op does his/her job correctly, then we wouldn't have the problem we have now. Tim1357 (talk) 16:23, 1 January 2010 (UTC)[reply]

If people weren't so wonky we wouldn't have problems, either. But I suppose that's another discussion for another day. I don't see a problem with Rich's edits. But I also am not one to look for rules that are being broken for the sake of it. Don't say that isn't what people are doing. If the edits are fine, and the only complaint is that they did not go through BAG, then that is what they're doing. But of course, we're all free to spend their time however we want to. - Rjd0060 (talk) 17:02, 1 January 2010 (UTC)[reply]

If nobody comments on a BAG request then silence is consent, BAG can use their own judgment in the lack of community input. If someone complains about it tell them they are welcome to contribute to future discussions instead of not contributing and complaining about it later. If the community decides a bot should not be approved after it has been approved then unapprove it. People are going to complain no matter what you do, don't try to please everyone or you will accomplish very little. Chillum ^{(Need help? Ask me)} 17:07, 1 January 2010 (UTC)[reply]

Silence is not always consent. BAG had already had serious problems with treating silence as consent. It has been discussed at WP:BOT. BAG members weigh the community impact of a request. Rich's bot request probably would have been approved and was moving towards being flagged even thought it might not require a flag when Rich removed the request.

If the proposer withdraws the BRFA, it won't get a request. There's no need established with the broader wikipedia community for a special set of policies for Rich. If there is, link to it. Rich can post a BRFA. If he uses the bots responsibly and he's good at coding and he's a communicative editor, and the task is well supported or uncontroversial he can make an argument at BRFA in the BRFA for speedy approval. As can anyone else. BAG needs more members. A useful bot operator is currently running. Others can also offer their services. --IP69.226.103.13 | Talk about me. 17:50, 1 January 2010 (UTC)[reply]

If people don't choose to participate in BAG discussions then yes, that means BAG has to use their own judgment. People need to speak if they want to be heard and if they stay silent then those making the decisions need to assume they are not objected. Chillum ^{(Need help? Ask me)} 18:47, 1 January 2010 (UTC)[reply]

Well I'm glad that most of us here are in agreement that Rich's edits to date have been good and that he is considered a reliable and responsible editor. I'm not saying that Rich shouldreceive special treatment above anybody but I do think that given the scope of his tasks on a daily basis that it would be a good idea for him to forms a good agreement with people like M Bisanz and is given freedom, and reviewed every few weeks or something, I dunno. I also sympathise with Bisanz that the input at BAG is very limited and how he could be blamed down the line for authorising a bot to run with little discussion. Hopefully we can reach a conclusion on this but I'm glad to see that all here are with Rich so to speak and not against what he does in practice in his edits. I believe his edits have overwhelmingly been uncontroversial and are mostly acceptable (and often much needed). A good New Year to you all. Dr. Blofeld ^{White cat} 18:24, 1 January 2010 (UTC)[reply]

Several strange edits on my watchlist recently (such as [9] [10]) alerted me to a massive undiscussed change made by User:Rich Farmbrough to Template:Portal and Template:Portal box. The change involves not just a change to the templates, but edits to thousands of articles, without any public announcement apart from a brief discussion on Template talk:Portal box in which Rich said that he was editing "a few hundred articles". For the edits, see Rich's contributions over the past few days.

I pointed out yesterday at Template_talk:Portal_box#Named_parameters that some discussion and announcement is necessary before the usage of a highly-used, protected template is changed on thousands of articles. Without responding to this post, Rich Farmbrough ran an unapproved bot task tonight to change the syntax of approximately a thousand articles, using AWB on his main account.

I am strongly considering reverting the AWB edits from this evening, as they represent not only an undiscussed and unapproved bot job, but they also seem to include severe errors. In particular, the edit on my watchlist removed see-also links for no apparent reason [11]. This sort of error is one thing that bot approval is intended to catch.

Could someone else contact Rich to impress the importance of discussion for large-scale bot jobs? Access to AWB does not grant automatic permission to edit thousands of articles without discussion. — Carl (CBM · talk) 00:21, 10 May 2010 (UTC)[reply]

On the contrary the initial changes to {{Portal}} were made by User:Svick, and discussed on Template talk:Portal. One of the consequences is that the syntax of {{Portal}} has become considerably simplified. Further to that additional changes to the internal workings of the templates were made and {{Portal box}} was simplified massively both in use and implementation. Discussion of the templates welcome on their talk pages. Rich Farmbrough, 00:29, 10 May 2010 (UTC).[reply]

...and what of seeking bot approval for the task? And the errors CBM mentioned? Equazcion ^(talk) 00:30, 10 May 2010 (UTC)[reply]

I was aware of the error and tried to correct it in the following edit. Rich Farmbrough, 00:46, 10 May 2010 (UTC).[reply]

And you neglected to seek the bot approval that might have avoided the error to begin with because...? Equazcion ^(talk) 00:49, 10 May 2010 (UTC)[reply]

Because it was a manual run of about 120 edits over 2 hours. Followed by a more substantial run of about 500 a few hours later. Rich Farmbrough, 01:28, 10 May 2010 (UTC).[reply]

According to your contribs, today you made 763 edits with summary "Update portal box using AWB" and another 661 with summary "new portal syntax using AWB". There are another couple hundred with edit sumaries that look like section edits. In total you made over 1600 edits with a summary including the word "portal". — Carl (CBM · talk) 01:36, 10 May 2010 (UTC)[reply]

The next edit did not restore the see also links, though ([12]). They were still missing hours later when I noticed the page on my watchlist. — Carl (CBM · talk) 01:04, 10 May 2010 (UTC)[reply]

Perfectly correct, my correction was wrong. I am glad someone picked it up. Rich Farmbrough, 01:30, 10 May 2010 (UTC).[reply]

• While we're waiting for Rich to actually offer a candid response above, I'd like to propose that something be done to assure that he can't run bot tasks without approval again. I don't know if that's possible, since he's been using his main account for automated tasks... but as much as Rich has been invaluable to Wikipedia, he seems to be taking shortcuts lately and not respecting process. Equazcion ^(talk) 00:52, 10 May 2010 (UTC)[reply]

• Pardon me, stupid question alert, but why would Rich need bot approval for something he is doing manually with AutoWikiBrowser (AWB)? You still have to actually hit the "Save Page" button on AWB, so it isn't a bot and wouldn't need bot approval....right? - NeutralHomer • Talk • 00:53, 10 May 2010 (UTC)[reply]

  It is quite obvious to this observer that Rich has modified the AWB code to allow him to run in auto mode from his main account. –xeno^talk 00:56, 10 May 2010 (UTC)[reply]

  Ah, then if it is modified, then yeah, that is bad and would need approval. I was assuming (and you know what they say about that) that he was just using the standard AWB program. I didn't read the code or anything (wouldn't know where to go and wouldn't understand it if I found it). Doesn't Rich have a bot of his own? (User Edit: After looking, he does...User:SmackBot.) Since he is one of the more prolific users here, he could easily get bot approval for anything and do it correctly with his own bot. - NeutralHomer • Talk • 01:00, 10 May 2010 (UTC)[reply]

  This thread may help lend some insight as to why Rich doesn't always engage the BRFA process (execute summary: it can, at times, be akin to molasses). Since that thread, I've been BAGged myself, so maybe Rich could prod me into reviewing his BRFAs as punishment for my incessant nagging =) –xeno^talk 01:04, 10 May 2010 (UTC)[reply]

  In some cases the slowness is a good thing. For example, the issue here is a wholesale rewriting of a highly-used template. That sort of thing should be announced on a village pump and people should have time to comment. The BRFA helps ensure that such notice is given, because the BAG members should be looking for announcement like that before they approve bots. We don't want people to fundamentally change a highly-used template without discussion. — Carl (CBM · talk) 01:15, 10 May 2010 (UTC)[reply]

  I don't disagree with anything you said. –xeno^talk 12:34, 10 May 2010 (UTC)[reply]

• Rich has been running bot tasks, both approved and not, sometimes insignificant, from his main account for some time. I've approached him several times with the suggestion that he move it onto an actual flagged bot to no avail. I eventually brought this to this noticeboard, but the rough consensus seemed to be to leave him be, so I left it at that and haven't bothered him since - but I still think he should move these various tasks onto a bot, and seek approval. –xeno^talk 00:56, 10 May 2010 (UTC)[reply]

Most of my AWB edits are on SmackBot, and I have submitted a fair few BRFA's, and intend to continue to do so. Xeno's advice on the matter was well taken. Rich Farmbrough, 01:28, 10 May 2010 (UTC).[reply]

"was well taken"... Does that mean you'll be seeking bot approval for all automated tasks in the future, and running them from your bot account -- or does it mean you see yourself as having taken his advice already, and don't see what you've done now as having opposed that advice? Equazcion ^(talk) 01:51, 10 May 2010 (UTC)[reply]

As per my original post, I rolled back the edits to namespace 0. Will look into the ones in the category: namespace tomorrow. — Carl (CBM · talk) 02:40, 10 May 2010 (UTC)[reply]

What a waste of time. Rich Farmbrough, 08:12, 10 May 2010 (UTC).[reply]

Wait. If most of the edits were correct why they were rolled back? This one is a very back example of rolling back. The new portal box syntax was discussed from April 5 to May 5 to its talk page and they were no disagreements. If they are problems with some templates we can fix them too. Per Wikipedia:Rollback feature "The rollback feature is a fast method of undoing blatantly unproductive edits, such as vandalism and nonsense." I don't see any vandalism nor nonsense in Rich's edits.

Carl, I strongly disagree with these rollbacks even if you are right in some other points. I noticed that you did the same when an anonymous IP started replacing otheruses4 with about. The edits we correct even if there were not supposed to be done with this way. After the rollbacks the watchlists were alerted for 2nd time at the same day.

On the new portal box syntax: I think we had consensus on that. If people think that this should be done by a bot or more slowly this is something we can discuss. -- Magioladitis (talk) 08:54, 10 May 2010 (UTC)[reply]

I agree, this is a, dare I say, misuse of Rollback by User:CBM. When some of these edits are clearly not vandalism or nonsense and are not unproductive. I would recommend that CBM quickly revert his rollbacks and not misuse the Rollback application again, else it be taken away. - NeutralHomer • Talk • 09:08, 10 May 2010 (UTC)[reply]

I don't think you can remove rollback from an admin. Jafeluv (talk) 11:16, 10 May 2010 (UTC)[reply]

Re Neutralhomer: see the section Wikipedia:Rollback_feature#Mass_rollbacks. Rollback is a standard tool for reversing bot edits. — Carl (CBM · talk) 12:32, 10 May 2010 (UTC)[reply]

Have we actually confirmed that Rich was running an unapproved bot or we just assuming? You know what they say when you assume. - NeutralHomer • Talk • 02:17, 11 May 2010 (UTC)[reply]

Re Magioladitis: Yes, the issue is that it needed to be done more slowly and thoughtfully. I asked Rich for evidence that the changes had been well announced as early as May 6 [13] [14]; he didn't give any. He also did not respond to my note on Template talk:Portal box, or to this note from another user on May 6.

In this case, the changes were not to support a new syntax – they were to make existing features not only deprecated, but actually forbidden. That sort of change, if it affects thousands of articles, requires more than a template talk page discussion. The procedure that was followed for the ambox rollout is an example of better practices: use village pump posts, keep everyone informed, make sue that the right features are implemented.

Now, I don't agree that most of the changes were "correct". At best, they represent a stylistic change. Worse, many of the edits removed named parameters, such as "break", that had been intentionally placed by article editors. Worst, at least one edit erroneously removed "see also" links; probably the same happened to other articles that used col-break around a portal.

The edits were also flawed in that they attempted to force a large-scale change without any sign of community-wide consensus. Running large bot jobs like this presents users with a fait accompli – if all the articles have been changed, the new syntax starts to look like the default. The use of unapproved bots to implement such changes without consensus is not appropriate (see Wikipedia:Requests_for_arbitration/Date_delinking#Fait_accompli and the other findings there). The simplest way to respond when someone makes this sort of attempt is to simply to roll back the edits, restoring the status quo.

I agree that it's unfortunate to have to roll them back, but if the unannounced, an unapproved bot is like BRD then this is the R part of it. The scale is larger, but that's exactly why it should be discussed completely before the edits are made. That is: bot edits are not supposed to be Bold, but when they are, they can still be Reverted like other edits.

To be clear, I am not focused on the bot approval for bureaucratic reasons. The point of bot approval is to guarantee public discussion and announcement of changes, and reduce the chance of errors. When bot operators make unapproved bot runs, they cannot be too surprised if the edits are rolled back. The point of the rollbacks here is to remove any errors unintentionally added by the bot, and to restore the status quo so that the changes can be discussed properly. — Carl (CBM · talk) 12:06, 10 May 2010 (UTC)[reply]

I don't like this use of terms like "forcing through without consensus". There's far too much pointless bureaucracy in templatespace already. There was rough consensus and a belief that the new code worked, and that's all that should be required. To me, the only issue here is that the task looked automated and that there may be discrepancies in the way that it was carried out. I certainly don't see that running to ANI to admonish people should be the first response to what amounts to a red tape problem outside of a few technical edge cases which nobody had previously spotted. Chris Cunningham (not at work) - talk 14:02, 10 May 2010 (UTC)[reply]

This entire change could be described as "pointless bureaucracy" – the previous template worked, and the only reason for these edits seems to be that a few editors felt that it required some sort of cleanup. Still, I would be fine with the change if it was actually supported by a village pump discussion that agreed that the previous behavior should be forbidden (as with the ambox changes). But in this case, it is not obvious that the "break" and "left" parameters should be forbidden. Instead, it looks like a small number of editors may have decided that their preferred implementation of the template was better and then forced that change on thousands of articles via unapproved bot jobs. If they had made the new code backwards-compatible, that would also have been fine, but then they would not have needed to run a bot job in the first place.

The reason that I brought this to ANI is that Rich has a history of doing this sort of thing. Had it been the first time, or had he made any effort to reply to the notices on his talk page of the template page, I would not have come here. The key point is that even when an editor has the technical ability to make changes to thousands of articles, via AWB or otherwise, such massive changes should not be done without widespread agreement that they are actually worthwhile. — Carl (CBM · talk) 14:26, 10 May 2010 (UTC)[reply]

From what I can see, your talk page comment didn't bring up any of the technical problems; that might have given Rich a bigger pause for concern than concentrating on the bureaucratic angle. That said, apologies for insinuating that you hadn't pinged him first. Chris Cunningham (not at work) - talk 14:41, 10 May 2010 (UTC)[reply]

Rich Farmbrough (talk · contribs · count · logs · block log · lu · rfas · rfb · arb · rfc · lta · socks)

On May 10, I brought a thread to ANI about Rich Farmbrough's use of AWB to run large-scale, unapproved bot jobs from his main account (archived thread). Today Rich is running the same job again, despite the comments on the ANI thread that he should avoid this. Seriously. — Carl (CBM · talk) 15:46, 26 May 2010 (UTC)[reply]

I don't see consensus for anything there. Why is what Rich is doing a problem? I don't mean on the macro scale of "He can't DO that", I mean are the edits improving the project or arent' they?--Wehwalt (talk) 15:51, 26 May 2010 (UTC)[reply]

He's breaking the bot policy in numerous ways, remember the bot policy is there for a reason, not just because we like red tape. For example, to name just two parts of the policy he has seemingly ignored, tasks like this should be run on a separate account (so as to be marked with the bot flag, and so as to not clog up a user's contributions page), they should be approved by BAG, and have some time at BRfA for community input (like the rest of Wikipedia, everything should be collaborative, this helps weed out errors, and provides extra ideas, Rich has been making a number of mistakes in his unapproved tasks, which I believe is largely due to him skipping this stage). - Kingpin¹³ (talk) 15:55, 26 May 2010 (UTC)[reply]

If I may ask, what are examples of some errors that he has made? --Kumioko (talk) 16:01, 26 May 2010 (UTC)[reply]

In this edit, he erroneously removes see also links [15]. That diff is from the last time he ran this task without approval. — Carl (CBM · talk) 16:03, 26 May 2010 (UTC)[reply]

(edit conflict) They are linked to in the previous threads, just in the one thread linked to by CBM, there is this error. Also, in a thread created by Xeno somewhere it was pointed out that Rich was making three edits to each page in a row, where the same job could have been accomplished with one. Just things like these, I expect there are more, and even if there are not, the reason that BRfA is there is to check these things, because of the potential damage a bot completing 10 edits per minute can cause. Even if the bot hasn't made any errors in the past (which is not even true in this case) it stills needs to go through BRfA, since BRfA is a preliminary measure - Kingpin¹³ (talk) 16:10, 26 May 2010 (UTC)[reply]

@Wehwalt: Who knows – the term "general clean" is so vague that nobody can tell what these edits are actually supposed to do. The point here is that large-scale changes (this one is over 500 edits in the last 2 hours or so) need to be clearly described and agreed-upon, which is what the bot approval process ensures.

In the last ANI thread, Rich received feedback such as

• "I'd like to propose that something be done to assure that he can't run bot tasks without approval again. ..." Equazcion 00:52, 2010-5-10 (UTC)
• "Rich has been running bot tasks, both approved and not, sometimes insignificant, from his main account for some time. I've approached him several times with the suggestion that he move it onto an actual flagged bot to no avail. ... " –xenotalk 00:56, 2010-5-10 (UTC)

It's hard to reconcile these with the same task being run, in the same way, a second time. If you were a bot operator, would you view those as supportive? — Carl (CBM · talk) 16:03, 26 May 2010 (UTC)[reply]

If he's allowed to run bots without approval so long as they are improving the project, does that mean everyone can run bots without approval so long as they improve the project? VernoWhitney (talk) 16:31, 26 May 2010 (UTC)[reply]

'fraid not ;), and he's not allowed to. - Kingpin¹³ (talk) 16:38, 26 May 2010 (UTC)[reply]

For the record, despite being aware of this thread for a half hour ([16] [17]), Rich has continued the bot job without responding here in any way. — Carl (CBM · talk) 16:21, 26 May 2010 (UTC)[reply]

He seems to have slowed down a bit, but still manages to make mistakes which then require reverting and redoing, which could be avoided if he examined the edits properly before submitting them. - Kingpin¹³ (talk) 16:38, 26 May 2010 (UTC)[reply]

• Perhaps this should go to WP:RFC/U. I stand ready to certify. –xeno^talk 16:40, 26 May 2010 (UTC)[reply]

A BRfA has also now been created (but not yet transcluded), but he's continuing to make these edits. - Kingpin¹³ (talk) 16:44, 26 May 2010 (UTC)[reply]

Of course that's not nearly good enough, since bot approval requests are to determine whether or not you should be running it at all to begin with. Seems like a token gesture so all the complainers might be appeased. No, though. I'd like to see an RFC/U too. Equazcion ^(talk) 16:49, 26 May 2010 (UTC)[reply]

Sure, maybe RFC/U is the way to go from here. - Kingpin¹³ (talk) 16:51, 26 May 2010 (UTC)[reply]

I'm deeplly, deeply unimpressed to see someone who's supposed to be setting an example behaving in such an un-collegial fashion. I would also like to see an RfC/U. ╟─TreasuryTag►duumvirate─╢ 17:26, 26 May 2010 (UTC)[reply]

• I think the issue here is the assumption that it is a bot doing the edits. It's not a bot, but rather normal user edits with an automation tool; therefore a bot approval is not required. Confusingly Rich Farmbrough does have a bot User:SmackBot, but it may make communication considerably more difficult because if you talk to Richard about bot edits he will naturally be thinking his bot User:SmackBot. Regards, SunCreator ^(talk) 16:59, 26 May 2010 (UTC)[reply]

  Much of the time it seems that he sets it to auto-save (he has apparently modified the AWB code), so it is automatic rather than semi-automatic. A bot by any other name... –xeno^talk 17:03, 26 May 2010 (UTC)[reply]

'it seems', 'apparently modified' ... is that speculation or has Richard actually modified the software? The Contributions are slow, a bot would be much faster imho. Regards, SunCreator ^(talk) 17:10, 26 May 2010 (UTC)[reply]

(edit conflict) As I mentioned above Rich slowed down since this thread was created, previously he was doing around an edit every 6 seconds. Also, if you take a peek at the previous ANI threads about this, you'll see Rich as good as admits it is automated, also the normal AWB would stop when he gets messages on his talk page, which this doesn't seem to do, it's also rather unlikely he really sits in front of AWB for hours at a time just pressing save, as well as this he seems to make a large number of mistakes, which he then reverts manually (similar to the way you'd fix a bot's mistake, whereas if it was semi-automated he'd be able to prevent the edit even happening), - Kingpin¹³ (talk) 17:17, 26 May 2010 (UTC)[reply]

It is entirely possible that he is sitting at his computer for hours-long stretches (including Christmas Day) pressing "save" at times reaching speeds close to 30 edits per minute [18], but I subscribe to occam's razor. He has also never denied modifying AWB, despite the numerous times I have suggested he has done so. –xeno^talk 17:23, 26 May 2010 (UTC)[reply]

Yeah, an explanation of this is required. Regards, SunCreator ^(talk) 17:48, 26 May 2010 (UTC)[reply]

As to bot tasks I just had Wikipedia:Bots/Requests for approval/SmackBot XXIX approved and that is now running, and will be moving on to Wikipedia:Bots/Requests for approval/SmackBot XXX shortly.

Meanwhile there is much other work that need to be done, just for this one fairly small piece - tidying up the {{Portal}} templates which have like most of WP grown organically and need a refactor. Inevitably some percentage of this will be more effectively done with AWB than Firefox - I already have 127 tabs open in one window on one PC, but it would be possible to use FF.

Most of these edits are trivial to check - which doesn't mean error free of course, and it is useful to do as many as possible in order to pick up any unusual situations.

Ok taking the dog for a walk, back soon. Rich Farmbrough, 17:14, 26 May 2010 (UTC).[reply]

Rich, it's not okay for you to "approve" trial of your own tasks (or to run unapproved/prior to approval), neither is it okay for you to run bot tasks on your main account. If these edits are so trivial to check, why do you keep missing mistakes? Making so many edits at that rate actually makes it very difficult to check them properly. Maybe you could clarify as to whether or not you are actually checking these edits properly before they are made? - Kingpin¹³ (talk) 17:28, 26 May 2010 (UTC)[reply]

I think you'll find Rich is not running a bot but doing it manually on his main account with AWB because that is what the Wikipedia:Bots/Requests for approval/SmackBot XXIX says to do. However, I would agree, Rich needs to explain himself considerably. Regards, SunCreator ^(talk) 17:48, 26 May 2010 (UTC)[reply]

I also glanced at the BOT requests above and noticed that there is a mention of other minor edits. This bothers me a bit and I think that those minor edits should be clarified. --Kumioko (talk) 19:02, 26 May 2010 (UTC)[reply]

Thread retitled from "Please block AWB bot - thousands of unnecessary capitalization changes".

Rich Farmbrough (talk · contribs · blocks · protections · deletions · page moves · rights · RfA)

SmackBot (BRFA · contribs · actions log · block log · flag log · user rights)

Recently, the AWB bot has been making totally unnecessary capitalization changes. These were being "discussed" on Rich Farmbrough's page, here and here. He said that he fixed the problem, but a day later, it was back. When brought up again, his response was to blank (archive) the page. Therefore, I request immediate halt to this use of this bot until this issue is addressed. Q Science (talk) 20:54, 28 September 2010 (UTC)[reply]

• The fact that so many have complained to Rich about pointless template capitalization changes and other sundry changes such as == spacing around headers == makes it clear that these are not uncontroversial edits. As such, they represent a violation of WP:AWB#Rules of use #3. I had laid off complaining about R.F. botting from his main account, but only because the edits were by-and-large useful and uncontroversial. This is no longer the case. These types of edits that change articles from how they were intentionally set by other editors to suit one bot-op's personal preference should stop unless they are approved by BAG. –xeno^talk 21:00, 28 September 2010 (UTC)[reply]
  • Would there be any objection if a regular editor simply hit the big red button on SmackBot's user page until an admin deals with the matter? Delta Trine ^{Συζήτηση} 21:03, 28 September 2010 (UTC)[reply]

I don't see that Rf is blockable about this, but we can stop the bot if we feel there is a problem. S.G.^(GH) _ping! 21:06, 28 September 2010 (UTC)[reply]

• Done. Delta Trine ^{Συζήτηση} 21:08, 28 September 2010 (UTC)[reply]
  • (edit conflict × >9000)  Done... about a minute after you did. Never mind. I left an informative message about this thread though. GiftigerWunsch [TALK] 21:12, 28 September 2010 (UTC)[reply]
• (edit conflict) I've blocked him until this can be resolved. This is clearly causing disruption. In addition to this, it has tagged the Main Page as uncategorized. According to the bot policy, automated bots cannot be run on main accounts unless approved by BAG (and AFAIK, this is not). (X! · talk)  · @926  ·  21:14, 28 September 2010 (UTC)[reply]
  • (edit conflict) Done[19]. Communication with the bot owner is going to be exceedlingly hard, it is difficult to have a meaningful conversation with a user undertaking blanking and implementing 1h[20] (one hour) auto-archiving on the talk page designated as the point-of-contact for the bot. There were multiple threads open on the User_talk page on the topic at the time of blanking. —Sladen (talk) 21:18, 28 September 2010 (UTC)[reply]
• RF has a long history of controversial mass-actions and refusing to discuss them or even consider that anybody else might possibly be right. Suggest he simply be banned from running a bot or engaging in any automated edits, or edits that seem to be automated, for one year. At the end of that year, if he has demonstrated that he will actually discuss his edits and not summarily blank discussions, he may apply at BAG to have his bot reinstated. → ROUX ₪ 21:26, 28 September 2010 (UTC)[reply]
• Good, I warned him like 4-5 times about changing {cite foo} to {Cite foo} in the last two days, and he was still making them. In general, it would probably be a good idea to force him to do these AWB runs on a BAG-approved bot rather than on his main account. Headbomb {talk / contribs / physics / books} 21:33, 28 September 2010 (UTC)[reply]

Copied from user's talk

Neither the bot nor I are editing at the moment, nor will we be for some time. I have revised the ruleset on Cite templates, as I said. When people start destroying the structure of the talk page the choice is to revert or archive. I had 35 threads, all pretty much dead, it seems reasonable to archive them - all accessible and new messages can still be left. I have no revised the rulset further and removed the Cite templates completely, restoring the status quo ante. Rich Farmbrough, 21:16, 28 September 2010 (UTC)

I advised him that he really shouldn't be changing the first-letter capitalization for any templates without consensus or approval; if a human editor used {{small case}} then it can and should remain small case. –xeno^talk 21:35, 28 September 2010 (UTC)[reply]

What is the point of doing that anyway? Does it help the server or something? Wknight94 ^talk 21:43, 28 September 2010 (UTC)[reply]

No. I think Rich's belief is that it somehow helps new users identify templates and improves readability [21]. My belief is that it just bloats the diff and makes it hard to see what the actual meat of the edit was, while imposing a personal preference that does not seem to be shared by the majority of editors. –xeno^talk 21:46, 28 September 2010 (UTC)[reply]

No, it is purely Rich's preference on the aesthetics of the templates. Headbomb {talk / contribs / physics / books} 21:48, 28 September 2010 (UTC)[reply]

(edit conflict) I had noticed this being done before, and found it mildly annoying that my templates were being capitalised for no apparent reason, especially as personally I think {{cite news|...}} looks better than when it's capitalised anyway. I figured this had basis in policy somewhere so I didn't protest; the edit summary including a "build number" and being performed by a bot suggested that it had been community-approved. GiftigerWunsch [TALK] 21:57, 28 September 2010 (UTC)[reply]

As far as I know, capitalization of templates hasn't been specifically approved. –xeno^talk 22:11, 28 September 2010 (UTC)[reply]

Yes, I don't appreciate that it is only one user's preference, plus the fact I don't really see any gain from doing this. Truthfully, I am surprised that Rich has been so unresponsive in this matter. He has been helpful in the past, performing Admin duties in a clear and objective manner. So what about this appearance of being community approved? Since it was not community approved, perhaps that was not intentional. ---- Steve Quinn (talk) 22:16, 28 September 2010 (UTC)[reply]

If the edits are uncontroversial there should be no difficulty in forming a Wikipedia-wide consensus, producing a policy, and then specifically authorising a bot to undertake the work. Wikipedia has processes for doing all of these. The large number of threads on just this one topic recently shows that it is controversial and therefore not something that is appropriate automated deployment (whether bot, or automated "manual" edits). —Sladen (talk) 22:25, 28 September 2010 (UTC)[reply]

• The ownership displayed in operating bots against consensus and removing avenues for discussion is deeply concerning conduct. Fifelfoo (talk) 01:50, 29 September 2010 (UTC)[reply]
  • We're not really that short of avenues for discussion. This has been on two noticeboards and one project space talk page, so far. See above. Uncle G (talk) 01:55, 29 September 2010 (UTC)[reply]
    • It isn't that we're short of venues; it is that the user is deliberately closing off the natural venue while making (to me) extremely controversial edits without consensus. I was noting that this is clearly a conduct issue. Fifelfoo (talk) 02:17, 29 September 2010 (UTC)[reply]
• My stalker of these many moons is currently turned off? I'd better sneak some writing in. ☺ In the meantime, I hope that everyone commenting on this is aware of all of the prior discussion, (now) linked to at the top of this section. Uncle G (talk) 01:55, 29 September 2010 (UTC)[reply]
• A bot being misused is as bad as at least ten regular vandals. Please don't tolerate such things. In case of repeated issues, impose a total automation ban (like Betacommand had back in the day) and/or an edit speed limit of 20 edits per hour or thereabouts, and generally urge the editor away from any repetitive editing of any type. 67.122.209.115 (talk) 08:20, 29 September 2010 (UTC)[reply]

I have two questions:

• What happened with the SmackBot/Citation Bot conflict. Did Citation Bot switched to the capitilised Cite web or not?
• Does anyone know how many of the 200k Cite web templates are capitilised and how many aren't?

-- Magioladitis (talk) 11:19, 29 September 2010 (UTC)[reply]

• In order:
  1. Rich Farmbrough told Dispenser that Dispenser should fix Reflinks to conform to SmackBot. See the discussion on Dispenser's talk page linked-to at the top of this section.
  2. Possibly. It's possible to find out, but expensive in terms of traffic for mere mortals without toolserver access.
• Uncle G (talk) 13:33, 29 September 2010 (UTC)[reply]

SmackBot is running again, it seems. I didn't manage to sneak in any writing, alas. ☺ Interestingly, as can be seen from this edit where {{silicate-mineral-stub}} was changed to {{Silicate-mineral-stub}}, it is still capitalizing the names of all templates. Uncle G (talk) 13:52, 29 September 2010 (UTC)[reply]

Why is Farmbrough blocked but this bot isn't? Shouldn't it be the other way around if the bot edits are the ones people dislike? Wknight94 ^talk 14:36, 29 September 2010 (UTC)[reply]

He was blocked for running bot tasks on his main account; the bot itself hasn't been doing much wrong right now (though it does seem to be used for non-bot edits). Ucucha 14:44, 29 September 2010 (UTC)[reply]

Angusmclellan just blocked SmackBot. Ucucha 14:45, 29 September 2010 (UTC)[reply]

Uncle G's diff is from today and includes the sort of pointless case change complained of. Since RF can't now (and before the block, seemingly wouldn't) change this behaviour, there seems to be no reason to leave the bot running and add to the comedy. Angus McLellan (Talk) 14:57, 29 September 2010 (UTC)[reply]

SmackBot is not following its own documented stop process, and I have just drawn Rich's addition to this.[22]. The instructions given at User talk:SmackBot are to place the string "STOP" in that page and a new section link is provided to do this. This "STOP" string continues to be the present, but the bot is making edits[23][24] including the these capitalisation changes under discussion[25]. A bot making edits while apparently stopped is a fairly serious bug as there is then no reliable way to stop the bot without resorting to an administrative block (as has had had to be performed here). —Sladen (talk) 14:59, 29 September 2010 (UTC)[reply]

Unless Rich has reprogrammed AWB, editing the bot's talk page will stop the bot until the orange bar is cleared and it is restarted by the operator. I would guess this is what happened. –xeno^talk 15:01, 29 September 2010 (UTC)[reply]

Same question as above but for Femto Bot (talk · contribs). Wknight94 ^talk 17:45, 29 September 2010 (UTC)[reply]

• Hm, it seems this bot does not have approval. (See also). –xeno^talk 17:50, 29 September 2010 (UTC)[reply]
  • Blocked for now. Not sure if it's worth blocking the rest, I'll have a look through to see if they are editing. - Kingpin¹³ (talk) 17:55, 29 September 2010 (UTC)[reply]
  • None of the other bots are active. So lack of approval won't be concern. As for Rich having access to unblocked account, I don't think that should be a concern here. IF he does start editing with one of them it's not going to do him much good, so not worth blocking the others, imo - Kingpin¹³ (talk) 17:57, 29 September 2010 (UTC)[reply]
    • It probably wasn't worth blocking even that one, to be honest. Part of the complaint here is that 'bot-like edits are being done through the main administrator-privilege account. The irony of blocking Femto Bot is that it was making edits that had heretofore been made through the Rich Farmbrough (talk · contribs) account, apparently entirely uncontroversially, since at least May 2010 (list). It was a 'bot intended to do exactly what people have been asking for.

      I think that we're starting to lose sight of the goal here, as this snowballs into desysopping discussions and the like. The goal is not to stop Rich Farmbrough at every turn. It's to get xem to get SmackBot and other people's 'bots onto the same page when it comes to changing/retaining capitalizations. Uncle G (talk) 18:15, 29 September 2010 (UTC)[reply]

  • I don't see any reason to block any bot account that does good edits and have approval of the community. -- Magioladitis (talk) 18:00, 29 September 2010 (UTC)[reply]
    • I don't see where it has approval? If Rich wants to move some approved tasks from SmackBot to Femto Bot, the appropriate course of action is to ask for a bot flag for the cloned bot at WP:BN. –xeno^talk 18:02, 29 September 2010 (UTC)[reply]
      • Actually, it's moving some regular monthly gnomish and robotic tasks from the administrator-privilege account, where they've been performed for months, to an unprivileged account. This is part of what you want, surely? Uncle G (talk) 18:19, 29 September 2010 (UTC)[reply]
        • Yep, it is ideal for the bot task to moved to a proper bot account, but it needs to be flagged and approved per the WP:BOTPOL. As I said, if the task is already approved (I'm not sure if it is, there are so many SmackBot BRFAs), R.F. can skip directly to BN to just ask for a flag as was done here. –xeno^talk 18:27, 29 September 2010 (UTC)[reply]
          • I was referring to the idea to block all Rich's accounts. For instance, Mirror Bot mustn't be blocked. Moreover, since edits that don't have consensus stopped I don't see any reason to keep the block and prevent Rich from doing other tasks. -- Magioladitis (talk) 18:31, 29 September 2010 (UTC)[reply]
            • I only pointed to the page listing the other bots, I didn't suggest they all need to be blocked. –xeno^talk 18:37, 29 September 2010 (UTC)[reply]

Seems a bit silly to have an administrator in an indefinite block. If he can't be trusted to edit at all, why would he be trusted to be an admin? If he isn't going to respond to the concerns or even respond to having been blocked, it seems the desysop process needs to begin before long. Wknight94 ^talk 15:28, 29 September 2010 (UTC)[reply]

• Well (1) I don't believe that desysopping is within the scope of ANI (RFC / ARBCOM) and (2) as you know, indef doesn't mean infinite. Syrthiss (talk) 15:32, 29 September 2010 (UTC)[reply]
• I don't think anyone is concerned about his administrative actions at this point, merely his bot-like edits. –xeno^talk 15:36, 29 September 2010 (UTC)[reply]
  • Ditto. No doubt he is distracted by something in RL and will take care of this in due course. Or he may be adjusting the programming as we speak. Once he solves the problem and implements it, there is no particular reason to keep him blocked.--Wehwalt (talk) 15:38, 29 September 2010 (UTC)[reply]
• This seems to be a bit overboard. The desired result is for Rich Farmbrough, Dispenser, and others ‎to get their tools singing from the same hymnal — no blocks, no desysoppings, no fuss, no acrimony. I made the point a week and a bit ago that this sort of thing is usually sorted out informally amongst 'bot owners. That's been my experience, as a 'bot owner. I'm rather saddened to see my argument undermined by the fact that this time, it as yet hasn't been. Uncle G (talk) 15:57, 29 September 2010 (UTC)[reply]

In my experience the user in question is unfit to be combining adminship and botting. Had an ordeal with it in Jan 2009 when it was inserting {{Ibid}} into 1000s of articles, which I was forced to revert with mere rollback. Stunningly, in one planned action the user behind it used rollback to revert these reverts and then Smackbot to reverse himself.One Example In general there are too many princessy bot operators who cannot be trusted with their tools. I'm sick and tired of dealing with the problems they cause, though of course bots in general are a net plus. Deacon of Pndapetzim (Talk) 16:07, 29 September 2010 (UTC)[reply]

The only reason I ask is the absurdity of having an admin indefblocked. If he's such a menace that he can't edit, surely he can't be an admin. Otherwise, if we're just waiting for him to return from RL distractions, then unblock him. Shouldn't have one without the other. Wknight94 ^talk 16:40, 29 September 2010 (UTC)[reply]

I would guess the block was placed as a form of 'wake up call'. If R.F. were not an admin, his AWB access could simply be revoked (admins have implicit access). –xeno^talk 16:42, 29 September 2010 (UTC)[reply]

If you unblock him and he continues on without resolving/discussing, then nothing happens. If he unblocks himself and continues on without resolving/discussing then you have cause to ask arbcom for an emergency desysop. (This is about any blocked admin in general, not a judgement on the specific admin involved).--Cube lurker (talk) 16:47, 29 September 2010 (UTC)[reply]

If you unblock him and he continues, then he gets re-blocked. If an editor can't reliably keep himself in an unblocked status, they often get banned. They sure as hell shouldn't be an admin! Wknight94 ^talk 16:53, 29 September 2010 (UTC)[reply]

It's not about what should be, it's about what is. There's no desysop process outside of arbcom. If he needs desysoping you there either needs to be a case filed or he would have to cross one of those bright lines that would pass arbcoms emergency desysop test. Such as unblocking himself.--Cube lurker (talk) 16:59, 29 September 2010 (UTC)[reply]

Well I guess I'm testing the waters for the viability of an ArbCom case. If no one is prepared to take that step, then he should be unblocked. I don't know Farmbrough and I don't care, but you simply can't have an indefblocked admin. Unblock or proceed to step 1 of desysopping. Wknight94 ^talk 17:14, 29 September 2010 (UTC)[reply]

I'm not sure I follow your line of reasoning. Indefinite is typically chosen when a time-limited block would not necessarily have the desired effect. In this case, the user is indef blocked pending a certain outcome (a commitment to cease making edits of the disputed nature until consensus and BAG approval is attained for the same - see comments from blocking administrator). The commitment has not yet been made, so the user remains blocked. The fact that they hold administrative rights is entirely peripheral. –xeno^talk 17:23, 29 September 2010 (UTC)[reply]

So you're okay with leaving someone blocked forever - assuming they never meet your requirements for unblocking - even though they have a sysop bit? Wknight94 ^talk 17:29, 29 September 2010 (UTC)[reply]

You've posited a hypothetical situation that I doubt will come to pass in the present case (I expect Rich will agree to eliminate the disputed changes from his AWB matrix until consensus and BAG approval are obtained for them), but yes - if a user is indefinitely blocked because of their doing X and they refuse to agree to stop doing X, then they will remain blocked indefinitely (+sysop notwithstanding). If this were the case, one would have a case to ask the committee to consider removing the bit, but it's premature at this point. –xeno^talk 17:37, 29 September 2010 (UTC)[reply]

If you - the blocking admin - "expect Rich will agree to eliminate the disputed changes from his AWB matrix until consensus and BAG approval are obtained for them", then you need to unblock him. Wknight94 ^talk 17:41, 29 September 2010 (UTC)[reply]

X! (talk · contribs) was the blocking admin. –xeno^talk 17:42, 29 September 2010 (UTC)[reply]

Eck, y'all and your X names.... Still, if the consensus here is that Farmbrough will break out of this odd trance, then he needs to be unblocked. Like now. For all we know, he is waiting to be unblocked before he'll even discuss. I don't see any comments from him about RL distractions. (Or are they offline?) Wknight94 ^talk 17:49, 29 September 2010 (UTC)[reply]

If you think the user should be unblocked you could ping X! (talk · contribs) for his thoughts. I agree that it may be ideal to have the user conditionally unblocked (conditional upon them not resuming their AWB tasks until the matter is finalized) so they can participate here directly, rather than by proxy. –xeno^talk 17:53, 29 September 2010 (UTC)[reply]

If my reading of the consensus is correct, it would be useful to unblock Rich and allow useful, administrator activity to proceed on the condition that Rich agrees to abide by BAG (that means no automated edits, no AWB, no Smackbot, no Army/*bot). For those worried that unblocking might be premature, perhaps we can agree (and document) that Rich would be blocked again immediately if any automated edits are made. That would allow discussion to continue, and for Rich to apply for suitable bot permission. If WP:BAG is being followed (in spirit and letter) then there is no longer a problem. For the avoidance of doubt, it should be reiterated in the process of unblocking that bot-like activity is not allowed from main accounts and the same for bot accounts that do not have up-to-date approval. The suggestion of <20 edits/hour may be a way to enforce this (although it is a technical solution to a social problem); without automation, the 10 edits per minute speed that I have clocked Rich at previously is unlikely to be attainable.

Above all, demanding punishment is the wrong direction: all that is being requested is simple compliance with Wikipedia policies. —Sladen (talk) 20:01, 29 September 2010 (UTC)[reply]

resp to Wknight94 17:29, 29 September 2010 (UTC) post - as far as I am aware, RF is still able to perform sysop functions (such as, but hopefully not, unblocking himself) but as blocked cannot post on any page other than his talkpage to say what he has done. RF can block, move over redirect, protect, and have access to The Chocolate Biscuit Jar, etc, as any other admin. It is his editing privileges only that are blocked. LessHeard vanU (talk) 20:29, 29 September 2010 (UTC)[reply]

As Deacon knows perfectly well "Rollback" is a blunt instrument, which he was using against policy. Specifically it reverts all consecutive changes by that user. Moreover he simply mass rollbacked a bunch of articles without differentiating by edit summary. Had Deacon used "undo" - even blanket undo it would not have been a problem. As it was he created a situation where potentially very old, very complex, fixes for which the code no longer exists (because they were one-offs - eg importing population figures, or correcting RamBot grammar problems) could have been undone. Since any edit, however trivial, would now prevent the recovery of this information without manual analysis of every single history of however many articles it was, I speedily reverted the hasty patch wherever possible, picked out those articles that could not be fixed for manual analysis, and removed the "ibid" tag, that he found so offensive, cleanly, without damaging the articles in any other way. As I recall I spent a considerable time undoing his mess, whereas if he had simply let me sort it out it would have been minutes. Nice to see that he bears a grudge about it though. Rich Farmbrough, 13:43, 1 October 2010 (UTC).[reply]

So if you mess up c. 1000 articles with your bot and you refuse to reverse your actions, anyone seeking to revert you is supposed to use undo? And you expect people to care about your time being spent? As you should remember, I informed you that I was using rollback and explained, which is enough to comply with rollback policy (not that anyone cares about that these days). If you did it now I would just block you, but I was trying to mencourage you to co-operate of your own free will. At this rate, you are unlikely to retain both your bot and admin access, but if you started being responsive and respecting bot policy and stopped arguing with everyone giving you feedback, you might have a chance. Deacon of Pndapetzim (Talk) 11:41, 3 October 2010 (UTC)[reply]

My response at the time:

Please note that Deacon rollbacked these edits without discussing with me. It is not an issue that he has reverted several hundred recent edits that are not those he is targeting (although they have a different edit summary) - recent stuff can by and large be redone - the problem with rollback is it undoes all the consecutive edits by that user to the article. So for example, edits the bot made in 2006, using code which will no longer run could be reverted. Adn there is no way to know which articles this applies to. Had deacon come to my talk page as clearly requested on the bot's talk page and discussed the matter there, we could have avoided a lot of work for both of us. I have rollbacked as much of Deacons rollbacks as I can, and am re-applying the removal of the template he finds so disquieting. I will be left with probably several hundred articles to go and check the history of manually. Deacon, you really needed to talk to me about this, rather than just apply rollback which is for anti-vandalism purposes only. Rich Farmbrough, 17:43 22 January 2009 (UTC).

I did not mention that you were rollbacking at 60 edits per minute. Hardly "mere rollback". Rich Farmbrough, 19:45, 3 October 2010 (UTC).[reply]

The 'mere rollback' was in reference to its power vis-a-vis bots. Deacon of Pndapetzim (Talk) 19:48, 3 October 2010 (UTC)[reply]

The way you were using it was more powerful than bots, and considerably more of a blunt instrument. And curiosity prompted me to check - in addition to the several hundred unaffected articles which you rolled back, you caused (unintended) damage to another 146, destroying edits going back to April 2007. It's no big deal but nor does it seem to me a shining example. Rich Farmbrough, 16:49, 4 October 2010 (UTC).[reply]

Rich, you are only further illustrating your tendency to avoid taking responsibiity for your own actions while arguing childishly with those trying to give you feedback. Believe it or not, this continued protesting only harms you. Deacon of Pndapetzim (Talk) 21:29, 4 October 2010 (UTC)[reply]

Feedback or rollback? You had a problem with my actions, your response was to violate policy in two different ways and break hundreds of articles. And to report me to ANI. I fixed up all the articles you broke, undid the actions you had objected to responded to your comments, asked you to talk to me about any future problems, and considered the matter closed. 18 months later you bring it up again and call me childish? So who is being responsible for their actions? The editor that takes action to resolve them, and invites discussion, or the one that gets out his admin-tools, and creates havoc? Rich Farmbrough, 15:32, 5 October 2010 (UTC).[reply]

The two complaints you are playing up are 1) rollback was used to revert bot disruption and 2) when your edits were reverted, good edits were reverted at the same time.

1) See WP:IDIDNTHEARTHAT; as was explained to you, this was necessary and complied with policy . 2) If you included good edits with bad edits in a bot run, then that's your mess, not the person reverting you; then as now, if you want your good edits to stick, don't package controversial ones along with it. Not everyone has a bot, and they aren't expected to spend days and days cleaning up the mess of bots when it can be done much faster.

These are poor and unpersuasive ways of deflecting blame. What's childish is not that your disruptive bot runs get remembered, but that you constantly argue with people trying to help you and constantly try to evade responsibility. Because you are very bad at doing this, all people perceive is immaturity and inconsiderate brat-ness ... the community expects people with active bots to be mature, to take responsibility and to deal with people with care. If you look like you have a princess complex, you are unsuitable. Deacon of Pndapetzim (Talk) 16:23, 5 October 2010 (UTC)[reply]

There seems to be a history of poor botop practices on RF's part here. This is not a new problem. This is a problem that has been going on for years. Bot operators are expacted to respond to concerns about their bots, and instead, he has reverted them as "vandalism". This is not appropriate conduct for a bot operator. What more, one should know that running one on your main account is prohibited, and that is also not a new problem. Even if the problems that led to the block are resolved, I would like to see some sort of action taken as a result of this. If nothing happens, this is just bound to happen again. It should go without saying that all of his fully-automated tasks are operated from his bot account and approved by BAG, for each and every task. If he refuses to comply, I think a reblock may be needed. I am reminded of Lightmouse in this situation: good intent, poor execution. (X! · talk)  · @728  ·  16:27, 29 September 2010 (UTC)[reply]

This is indeed not, as I commented above, a new issue. RF should be banned from bot or bot-like edits, period. Same as Betacommand was. → ROUX ₪ 17:54, 29 September 2010 (UTC)[reply]

Adding my view, and a new note too: RF does not do sandboxing. Every experiment is in main space and real time. "One more 30 trials by BAG for BRFA please" - go ahead. "Oops, I the bot something botched it". I don't get why this admin-bot-loner is cared for in our community this way. -DePiep (talk) 21:54, 7 October 2010 (UTC)[reply]

Rich wrote somewhere (I can't be bothered to find right now, I am busy in real life too) that he removed the cite -> Cite from SmackBot's code. Should we move on, unblock, let SmackBot keep doing its main tasks and re-report of there are still complains? -- Magioladitis (talk) 08:06, 30 September 2010 (UTC)[reply]

• Check User_talk:Rich_Farmbrough#Is_it_not_possible.... Rich removed the cite -> Cite and the spacing around heading from his fixes. -- Magioladitis (talk) 09:27, 30 September 2010 (UTC)[reply]
  • There's still a problem with other templates, like the stub one. Also, Rich was also blocked for running unauthorised bots on his main account, I'm yet to see any suggestion that this is going to stop, and it's an on-going issues, which he's messed up repeatedly. I think editing the main page like that (arguably making this an unapproved admin-bot) can not be ignored. Personally, I think that an edit limit of ~20 edits/hour, along with a(nother) stern warning that all automated tasks must be approved by BAG, would be a good way to go here. - Kingpin¹³ (talk) 10:09, 30 September 2010 (UTC)[reply]
• I am opposed to unblocking him yet, per "My View" section above. (X! · talk)  · @491  ·  10:47, 30 September 2010 (UTC)[reply]
  • So what would satisfy your concerns? Let's come up with something concrete and actionable. Here's a starter that you can boldly modify: Uncle G (talk) 11:54, 30 September 2010 (UTC)[reply]
    • for context, "here" is referring to this section. (X! · talk)  · @553  ·  12:16, 30 September 2010 (UTC)[reply]
    • I would only support this is it was made explicitly clear that all automated bot-like tasks be approved by BAG. (X! · talk)  · @553  ·  12:16, 30 September 2010 (UTC)[reply]
      • Be bolder with the section! ☺ It's there to be edited. Uncle G (talk) 12:21, 30 September 2010 (UTC)[reply]
• I don't think unblocking is the way to go until/unless he agrees to some kind of restriction on automated edits. StrPby (talk) 11:23, 30 September 2010 (UTC)[reply]
• I have to agree. This response goes some way to addressing concerns but it does not go far enough. I have suggested an alternative, simpler, set of possible conditions below. I would like to try to minimise any chance of this problem reoccuring before unblocking. — Martin (MSGJ · talk) 12:30, 30 September 2010 (UTC)[reply]

Rich should be unblocked at least to comment in this discussion. I bet nobody believes that Wikipedia is at danger if we unblock him. -- Magioladitis (talk) 18:28, 30 September 2010 (UTC)[reply]

Conditions that would satisfy X!, Kingpin13, MSGJ, and others

[edit]

1. No more changing the cases of the initial letters of any templates. No more changing {{for}} to {{For}}, or changing {{silicate-mineral-stub}} to {{Silicate-mineral-stub}}, or changing {{coord missing}} to {{Coord missing}}, or anything else.
2. No automated editing at all from main account. Specifically:

1. All 'bot-like tasks, like this one, no matter how uncontroversial, to be farmed out to non-administrator accounts like Femto Bot (talk · contribs), and approved via Bots/Requests for approval.

2. Use of a dedicated non-administrator account, in accordance with AutoWikiBrowser rule of use #2, if editing at speeds like 10 edits per minute with AutoWikiBrowser.

3. Clear linkages be provided on the bot pages to the appropriate approvals through Bots/Requests for approval.

4. No altering a bots function outside of the linked approvals without approval of the change.

5. Scope and function(s) of the bot explicitly stated both in the application for approval and on the bot page.

3. A message to any bot's talk page stops the bot;

3.1 the task is not restarted until the issue is resolved.

4. No unblocking one's own bots.

--(end of list)--

Small-ish suggestion re point 2:

Merged with above.

Looking at the preceeding discussion, it should be crystal clear regarding he be fully transparent and accountable in his use of bots.

- J Greb (talk) 15:31, 30 September 2010 (UTC)[reply]

• I support the modified conditions now that the BAG approval is added. (X! · talk)  · @914  ·  20:56, 30 September 2010 (UTC)[reply]
• I have to say, one of the things which I personally find wanting is Rich's attitude. He seems very reluctant to ever admit that he's actually done anything wrong (even after slapping a maintenance template on the main page..), for example, his first unblock request showed a clear lack of remorse, and his comments on his talk page display that he doesn't really seem to appreciate what he was actually blocked for, let alone be prepared to admit that he shouldn't have done the various things which lead up to the block. However, I do agree with the conditions above. Although I'm not completely convinced they would be enough, they're all basically already in policy, so Rich should be doing most of these already.. - Kingpin¹³ (talk) 21:18, 30 September 2010 (UTC)[reply]
• I agree with the above and guess that User:Sladen would too, judging from RF's talk page. It's not really a complaint but I think RF's 'man on a mission, the only one who can possibly solve wikipedia's problems' attitude is starting to look a bit silly. I thought the unblocking was so that he could respond here, not so he could carry on with what he was doing before. Is this guy actually listening to anyone? Can someone suggest he post a short note to us mortals here on his own wp:ani section. Please :) Sf5xeplus (talk) 22:33, 30 September 2010 (UTC)[reply]

1 is redundant to 2.4

2 has nothing to do with some of its sub-conditions

2.1 goes off at a tangent

2.2 is good

2.3 is good

2.4 is good

2.5 is good

3 Is unreasonably onerous. AWB tasks will be stopped by a talk page message. Other tasks you will have to find an admin to block if I am not around - although I am likely to be for non-AWB content tasks.

3.1 Again unreasonable. This gives the other party veto - on Wikipedia you will find someone to oppose the tiniest changes. I will discuss, as I have with everyone (except with one editor who has been gentleman enough not to bring it up - for which my apologies), but we are talking about approved tasks here. Ninety nine times out of a hundred problems are sorted out on talk pages, but it is not reasonable to expect every one to be. A Bag member can be called if the other party thinks there is clearly a problem that a botop is refusing to acknowledge, and they have the power (or so the template documetnatin says - and templates documentation, I am informed, is the ultimate authority on Wikpedia (yes-joke)) to revoke BRFAs. There are 17 "Active" Baggers and 24 "Inactive". Or you can find an admin to block the bot (pretty easy - changing one letter got SmackBot blocked) or maybe even a 'crat who will do it on the basis of two duff diffs? (Yes another joke, but also true.)

4. Seems reasonable, given my arguments at 3.1. As long as I am allowed to remove CBM's blocks. Rich Farmbrough, 01:38, 4 October 2010 (UTC).[reply]

Incidentally I would expect those admins that zoomed to stop/block my bots to have taken the trouble to leave me a note to that effect, especially as this ANI is supposedly about communication? Well maybe they had collective amnesia, but five admins all failed to leave me a note, including the one who left a stinky edit summary in his block. Rich Farmbrough, 01:38, 4 October 2010 (UTC).[reply]

• I (no admin) support. RF's reply here into 3: RF is opposing "no resuming the bot while unresolved", and that says it: "unreasonably onerous": well, RF, this is what this community is about. If you can't stand -let alone cope with- a stopped bot, then you're in the wrong place or in the wrong attitude. -DePiep (talk) 22:52, 7 October 2010 (UTC)[reply]
• Support: as for not being able to restart the bot until it's resolved, I'm not sure I see the problem; "resolved" is a fairly flexible term: I would consider it resolved if a) the user didn't give a reason or gave an uncontroversially frivolous reason (or indeed, were "just playing it safe" and it suggests to do on User:Smackbot's page), b) the user agrees that it is resolved, or c) the community determines that the issue is resolved or the bot should be resumed. As for the redundancy between points 1 and 2.1: it's redundant, so it's a moot point: redundancy isn't always a bad thing, and in this case it serves to make it double clear what the proposed constraints will allow or disallow. GiftigerWunsch [TALK] 23:10, 7 October 2010 (UTC)[reply]

Unblocked

[edit]

Rich Farmbrough has finally agreed[27] to both participate in discussion here, and to cease doing the disruptive and unresponsive editing that got him blocked in the first place. So I've unblocked him. Wknight94 ^talk 20:07, 30 September 2010 (UTC)[reply]

Thanks for the unblocking. - it's hard to know where to start with this one. It is more about human nature than anything else - and text communication. So lets start with Wknight94's message above.

"Rich Farmbrough has finally agreed both participate in discussion here... "

OK so this is minor, maybe, and in good faith, but the implication is that I was reluctant to join the discussion. Obviously that is the impression Wknight94 picked up, probably from something said on my talk page by my unblock request. However I was in the middle of typing a comment here when I was blocked.

• 21:01 notification of ANI
• 21:03 - 21:06 started reply
• 21:09 - blocked.

As my comment (later forwarded by Xeno, for which thanks) said "Neither the bot nor I are editing at the moment, nor will we be for some time. "

-so I wasn't exactly reluctant to "cease doing" .. "that [which] got him blocked in the first place".

Further "disruptive and unresponsive editing" is rather jumping to conclusions, based on what others were saying.

More later as I am being pinged on my talk page (about responding here I think). Rich Farmbrough, 07:24, 1 October 2010 (UTC).[reply]

OK. I'm going to keep this short: I could write a book, but it would be TLDR - I hope the following is both informative and reassuring.

What happened? SB dates maintenance tags as it's most intensive task. It also does various minor cleanup as it goes - as people have said pretty unexceptional.

One of the features of templates - indeed all wikilinks - is that they are not simply literals but a minor grammar in their own right for example:

____ __ _ __ :____ __ _ __ Template____ __ _ __ :____ __ _ __ Citations____ __ _ __ needed____ __ _ __ is a perfectly good link to {{Citations needed}}. Particularly when SB's regexes were hand crafted for each template (back then merely 1000 , now well over 2000 counting redirects) dealing with this complexity meant canonicalisation of template names was the only way to go. (I thought dating a few templates was going to be trivial when I started.) Therefore standard functionality is to replace the clean up template names with a clean version, following redirects. This also has the benefit that the number of different possible clean up templates left after a run is 569 (!) rather than four or five times that number. It also means that the template is capitalised - an "arbitrary but intelligent" decision I made - yes I know algol coders, C coders, perl hackers just love lower case - and I have been all of those things - but for someone who has never coded it seems to me that the capital says "Here is a new thing starting that is somewhat like a sentence." - and it is not a great leap from {{Citation needed|reason=this seems unlikely|date=July 2009}} to "Citation needed, because this seems unlikely, request added July 2009" (Incidentally anyone looking at http://en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Template:Citation_needed&hidetrans=1&hidelinks=1 right now will see six articles that appear to redirect to Citation needed - these are almost certainly articles that have had the redirect placed at the top and the article text left in place - normally I would go and fix them, but I am being "chided" for not writing here as a priority.) Having canonicalised the templates - which - only takes (569 + a few) rules, dating them is simple - provided that they haven't already been dated, don't have an invalid date don't have "date" mis-spelled (SB will pick up "fate" but not "jate" - that is left for some poor human drudge to do - as being a very unlikely mispelling SB is pretty conservative to avoid errors, similarly it will pick up "date=Spetember" and correct it to September but "date= Josh is ghey" will simply get over-written with the current month and year) - so another 569 rules for the basic dating and a few hundred to deal with specials like "As of". Anyway some of the minor cleanups SB picked up were related to templates in wide use that either had oodles of redirects or were moved. Again pretty unexceptional. Foolishly on 6th Spetember (or September if you prefer) I added the Cite templates to this list - this was foolish because cites are an area where "angels fear to tread" much like dates and MoS - I have been foolish enough to contribute to MoS too. Having said that it was foolish, it wasn't mind-numbingly stupid, despite what others may think, I had been pleasantly surprised not to receive negative feedback on other changes, and there are a surprising number of redirects to , for example {{Cite web}} - 21 in fact. That's 21 templates - not 21 pages, the number of pages is 12,118 and the number of actual uses will be higher still. Moreover I knew that removing those four templates would be fairly trivial. So what was the response? Were seven different kinda of hell unleashed upon my talk page? Find out in the next thrilling episode. Rich Farmbrough, 01:51, 2 October 2010 (UTC).[reply]

I get the impression that you think the only reason you were blocked, or that people are concerned, is this recent problem with SmackBot's capitalisation, which I see this more as being the last straw. I think the underlying problems are: You ignoring bot policy, by running unapproved bots; running bots on your own account; not responding to concerns, which you are also expected to do as an administrator, but instead you blank messages, ignore concerns, claim to be too tired (even claim that you're always too tired), you even seem to play word games. These are the problems which need to be addressed, since they are what lead to problems such as the template capitalisation. It's no good just dealing with the result of these problems, as we know (from prior experience with you in regard to bots) that all that happens is problems arise again. This isn't a one-off mistake. That said... Looking at what you say above, it mostly seems to be explaining how the task works, that's nice, but really the question is can you prevent SmackBot from changing the capitalisation of all templates (not just the cite templates or whatever). You could maybe even use a regex find/replace after the other changes are made to effectively "revert" any capitalisation changes made (but before actually saving to the wiki)? - Kingpin¹³ (talk) 05:36, 2 October 2010 (UTC)[reply]

The point of explaining how the task works - which is pretty deadly dull - is to lay the ground work. Understand, for example, that powerful though AWB is, it is an application, not a programmers framework like Pywikipedia. SmackBot's rulebase runs to 750k+ of XML - let me find out how many regexes that is - 5067 rules plus some "advanced" rules. The suggestion you make above might be workable - while I try to keep the rules as simple as possible, there may be an elegant solution, but on the face of it I would have to pull apart the redirect consolidation rules and have a separate one for "Sentence case" and "lower case", and the same would apply to any specific rule - since there are about 2500 redirects and some hundreds of other rules this would mean a massive increase in the rulebase (possibly more than doubling it). I outlined what is easy and what is hard to change, on my talk page, along with the benefits. And I really don't hear a clamour for {{infobox... There are two reasons I find commenting here tiring: one is the fact that every word is hostage to fortune - as shown in your comment. And indeed every edit or lack of an edit: - I don't know whether its funny or sad to have people counting my edits between being unblocked and starting to comment here. The suggestion that it would have been better for the project to leave redlinked categories on a hundreds articles than to keep the ravening hordes of ANI waiting - especially when commenting on the volume of text here, let alone the 50k or so on my talk page was likely, and still is likely, to take some time, may have some merit, but I can't see it. More later. Rich Farmbrough, 13:02, 2 October 2010 (UTC).[reply]

This reads a bit like "Smackbot is too big to be maintainable". If that's the case, break Smackbot up into small pieces running on separate bots that are individually auditable. If the answer is that individual smaller tasks would mean loosing the opportunity to discreetly make whitespace/capitalisation changes otherwise deemed without merit, then that's actually a positive; the minor changes brought your activities to a head—as Kingpin mentions (and I'll reiterate for the explicit avoidance of doubt) there is a wider general problem; which is one of interaction (acting on feedback, not disputing/arguing it; and participating in discussion to a closure). —Sladen (talk) 14:58, 2 October 2010 (UTC)[reply]

Sorry but here (and on my talk page) you are plain wrong about software maintainablity.

• (Citation[ _]+style|Cleanup-references|Cleanup-citation|Ref-cleanup|Citationstyle|Citation-style|Refstyle|Reference[ _]+style|Reference-style|Cleanup-refs|Citestyle|Cleanrefs|Refclean|Refsclean|Source[ _]+Style|Sourced[ _]+wrong|Ref-style|Refcleanup)

is longer than your proposed

• (clean(up)?-?(ref(erences|s)|citation)||(cit(ation|e)|ref(erence)?|source?)-?([Ss]tyle|clean(up)?)|sourced[ _]+wrong)

But it is also more maintainable and more readable

Your version

1. Has errors of coding
2. Has errors of design
3. Is hard to add to
4. Is hard to remove items from
5. Has no discernible performance benefits, and maybe performance costs (although I do agree that this is "in this case" not critical, I'm fed up with people saying "don't worry about performance" as a blanket statement when we have literally hundreds of fantastic servers worth millions of dollars which time out serving pages, yet my little desktop, encumbered as it is with the world's worst operating system, runs most of the software I write (pace infinite loops) before I can blink, or at least IO bound. I was running SmackBot - and everything else on a skip-rescue PC until about 18 months ago.)
6. Is less readable

It is also very very clever - and I am not being sarcastic. In fact I am being a little peacocky, because it is exactly the sort of regex I was using until I simplified and automated. And it caused a number (not necessarily a lot) of problems, picking up incorrect templates.

(See now, this has taken me over half an hour to write, maybe I'm slow, maybe I'm just being careful what I write - and maybe other people spend as long and as much care on what they write, but I certainly see evidence that some of them don't read what I say, and just bash of a few hundred words at top speed to express their feelings. But I have probably already spent about 4 hours on this thread, let alone my talk page. And I am being accused of "not responding" - I know there are subjects here I haven't even broached, and I have made it clear that it will take time to get to them - anyone who can't wait - well I would offer an informal reply on my talk page - but it would only get quoted back out of context here -as has already happened.)

Rich Farmbrough, 08:37, 3 October 2010 (UTC).[reply]

At the risk of putting words into Rich Farmbrough's mouth, I'm going to respond to something Kingpin13 wrote: that RF has claimed "to be too tired (even claim that you're always too tired)" -- only because it sounds true. There seems to be a familiar pattern to the last chapter of the career long-term Wikipedians: increasing lack of patience with others, obsession with details (which may appear to be WikiLawyering), & an increasing weariness with contributing or the discussion which follows contributions. The bastards finally wear the dedicated & selfless volunteers down. Now if this is truly what is happening here, then the only advice I can offer to Rich (I say "only" because I honestly don't have a better solution & wish there was one) is to simply cut back on what you do. If running certain bots on Wikipedia is getting to be more of a pain than it is a joy, then stop doing it. Wikipedia can survive without all of the bots being operated here, believe it or not; & if I'm wrong, it's likely someone else will pick up the slack. If someone doesn't, the resulting carcass will get preserved, & another group will try to resurrect the online encyclopedia with a slightly different set of rules of operation. And I'm writing this because I, too, feel tired with Wikipedia, just like Kingpin13 says RF claims to be. And after I finish the projects on my plate here (i.e., a few groups of articles & upload a few PD images), I'm going to drop my involvement here even more. Or if one of these leads I'm chasing gets me back into the job market, maybe sooner. -- llywrch (talk) 06:45, 3 October 2010 (UTC)[reply]

llywrch is right - but I don't blame the "Aha! I have a diff... " brigade. "I too was once as you." (Yes that's (self-deprecating) humour, not being patronizing.) Rich Farmbrough, 08:37, 3 October 2010 (UTC).[reply]

Well, take time to consider my advice. Maybe if enough experienced Wikipedians say "I'm burned out, so I'm quitting" the PTB may decide that it would be better for the Wikimedia projects to allocate resources to retaining veteran editors than increasing the the pool of Crowdsourcers in places like India. The idea is to create a quality encyclopedia, not to recruit every Tom, Dick & Hari to make questionable edits to Wikipedia. -- llywrch (talk) 22:56, 3 October 2010 (UTC)[reply]

Changes to cite template

[edit]

I have just discovered that, without any discussion, Rich made significant changes to the Template:Cite web. Specifically, while we were complaining about his bot, he changed the template examples from lower case to upper case. Since this was his response to complaints on his page, and since his deleting the comments on his talk page without responding to our comments is what started this whole discussion, I think that these changes need special attention.

• 22:21, 21 September 2010 Start of "Could you not capitalize citation template in the future?"
• 15:49, 26 September 2010 Start of "cite vs Cite"
• 12:47, 27 September 2010 Rich says that the bot is no longer changing "cite" to "Cite"
• 18:41, 27 September 2010 I complain again because the bot is still making the changes
• 18:42, 27 September 2010 Rich changes the case of the first character in the Template:Cite web examples
• -- There are additional comments in both threads
• 20:29, 28 September 2010 Rich blanks the talk page without responding to anyone since
  • 15:19, 26 September 2010 in the 1st thread, and
  • 12:47, 27 September 2010 in the 2nd.
• 20:54, 28 September 2010 This ANI discussion was started by me.

It was very difficult to step back through his contribution log. It appears that on Sept 28, he made well over 5,000 edits. (Perhaps over 100,000. And all with AWB. It is totally unbelievable that the admins allow this. Link to contribution log so no one else will have to search for it.)

As a result of this "new" information, I am requesting others to comment before I simply undo his uh, changes, to the template. I for one do not like them. For another, I think this was an underhanded slap in the face. He didn't even have the courtesy to mention this on his talk page when two groups of people were complaining about the same subject. He also did not mention it in any of the other discussions since. Q Science (talk) 01:08, 3 October 2010 (UTC)[reply]

It's obviously been established above that he has no consensus for the capitalisation changes, so I say change them back. Strange Passerby (talk • c • status) 01:50, 3 October 2010 (UTC)[reply]

Just revert Q. Rich Farmbrough, 08:02, 3 October 2010 (UTC).[reply]

ok reverted [28] Note there's one reason why none of the fields are capitalised, and that is that non-bot editors can enter them without having to press shift key. Clearly the first field could be an exception, and changed by bots later. There are many arguments, it probably didn't need changing - anyway continue that debate on relavent page.Sf5xeplus (talk) 12:31, 3 October 2010 (UTC)[reply]

Just revert Q. RF, above. Why this message at ANI? It was not a question, RF. It was a example of problematic and strange behaviour of the bot operator. You did not see that - q.e.d. -DePiep (talk) 18:47, 7 October 2010 (UTC)[reply]

• I stalk Rich's talk page, but even now, I still fail to understand why some editors object so much to capitalisation. The typical reason seems to be they are immaterial and thus unnecessary. Whilst I am not sure why he changes the capitalisation, it makes not a jot of difference to anything, whether in the smaller or the larger scheme of things. Our servers recognise and resolve both. The important thing I see is that SmackBot is providing an invaluable service with all the detritus it picks up. This business about capitalisation should be allowed to overshadow the huge contributions (whether in terms of load or in types of small changes) by Rich and his bots. --Ohconfucius ^¡digame! 05:08, 4 October 2010 (UTC)[reply]
  • The question is indeed, why does he do it? Imagine a page with all cite templates in lower case. RF comes along, changes them all to uppercase. I add a new cite, using the edit box cite functionality. This by default add cites in lowercase. We now have, thanks to RF's unnecessary edit, an article where some of the cites are in uppercase, some in lowercase. Everything still works, but we get less consistency for no good reason at all... Fram (talk) 07:42, 4 October 2010 (UTC)[reply]
    • The great thing is, being only visible in the read mode makes them totally inconsequential. Not worth busting a blood vessel over it, IMHO. --Ohconfucius ^¡digame! 09:35, 4 October 2010 (UTC)[reply]
      • Totally inconsequential, but a nuisance when you notice an article on your watchlist being updated by an RFbot. You get tons of "differences" which consist only of meaningless spacing or capitalization, and to find what was actually changed (usually, admittedly, improved) takes a lot more work than it should do. Fram (talk) 09:38, 4 October 2010 (UTC)[reply]
        • In their defence, the changes should happen only once - once the page has been spaced and template tagged to the bot's satisfaction it should make no further edits ? Sf5xeplus (talk) 18:02, 6 October 2010 (UTC)[reply]
          • That is not in fact, the case. Go to the top of this discussion and read the past discussions linked there, where you will find the discussion of the pointless back and forth that brought this issue to prominence in the first place. Uncle G (talk) 12:21, 7 October 2010 (UTC)[reply]
            • I was ignoring the "bot wars" which I think you are referring to. Clearly that was an error. Excluding that I believe my statement is still (ie currently correct)Sf5xeplus (talk) 16:31, 7 October 2010 (UTC)[reply]

"This business about capitalisation should be allowed to overshadow the huge contributions" I'm not sure if you're suggesting we should let him continue with these disputed changes (which makes zero sense to me), or if we should "cut him a break" as long as he doesn't do it going forward. You have to keep in mind that this ANI would not have come to be had Rich stopped making these changes after being asked several times. –xeno^talk 12:59, 4 October 2010 (UTC)[reply]

I'm not suggesting that Rich ignore the talkpage complaints; au contraire. Of course he should be transparent with the rationale for the capitalisation changes too. If I understand correctly just what Rich he has programmed his bot to do, I'd say it combines a large number of inconsequential changes which would otherwise never be made with jobs such as tag-dating. One way forward is perhaps he will program his bot with more consequential tasks (such as date format alignment or other style fixes), so that the chances of there being only inconsequential edits is further reduced. Ohconfucius ^¡digame! 14:21, 4 October 2010 (UTC)[reply]

The capitalization changes aren't merely inconsequential but also undesired. –xeno^talk 14:22, 4 October 2010 (UTC)[reply]

FYI re: cleanup template capitalization

[edit]

FYI WP:AWB does cleanup template ucfirst capitalization in-house (see Wikipedia talk:AutoWikiBrowser/Bugs#autotag makes disputed capitalization changes), so Rich is at the mercy of his tools. –xeno^talk 16:31, 4 October 2010 (UTC)[reply]

Even if AWB does it automatically, operators are still fully accountable for actions the script takes. The warning is on top of Huggle, Igloo, AWB, etc. (X! · talk)  · @131  ·  02:08, 5 October 2010 (UTC)[reply]

FYI: Wikipedia:Bots/Requests for approval/SmackBot 35. Fram (talk) 08:28, 8 October 2010 (UTC)[reply]