Gymel

Thanks very much for your comments on the authority control proposal on the Village Pump. We've refined it and worked out some more details after the discussion, and there is now a community Request for Comment to approve it being implemented. Any comments gratefully received! Andrew Gray (talk) 09:59, 29 June 2012 (UTC)Reply

Hours ago I posted a note to Andrew Gray including this paragraph:

By experience I know there are many disintegrities at GND (many at least re English-language writers and illustrators who may not be notable) but they seem to be flagged "not differentiated" which I take to mean a known problem, not an error to report. (quoting myself at User talk: Andrew Gray#Authority data integrity)

This hour I have read one of your explanations in our English Wikipedia talk space and I have skimmed your user page. It isn't clear to me what may be useful to report regarding GND data.

For what it's worth, I have noted probably dozens (24?) of en.wikipedia biographies where VIAF does not match the LCCN ID with the one unique or with any one of multiple GND ID. That is, LCCN (which I have always confirmed) and GND (sometimes confirmed) both have records for the biography subject but VIAF does not match them; there are at least two different VIAF for the subject. I have noted these instances by a comment in the biography source code as for Emily Gravett (External links code). --P64 (talk) 21:32, 4 January 2013 (UTC)Reply

Hi P64,

"not diffentiated" is the prevalent flavor of GND authority records for persons and the like. It is not a problem in itself but manifests a conceptual difference. As it turned out over the last 20 years there is no other authority file with a comparable concept and this kind of record is of extremly limited use for todays demands. But these records exist and their use is still gerater zero (they are assigned an authority number you can use in links). What interests here are the problems which arise when "not differentiated" and differentiated ("individualized") records both exist and come into conflict with each other.

This kind of problem has aggravated in the last months as kind of side-effect to efforts of improving the GND: DNB is introducing individualized records for persons at a higher pace (good) at the price of only checking their own holdings (very understandable: they simply cannot gain complete knowledge about all present and past usages of a given record in all institutions). This means that the "non differentiated" record is usually not obliterated or upgraded any more but resides in the authority file and for all other participationg institutions nothing changes (with respect to data they produced up to that point). Now since VIAF is basing its clustering chiefly on inference (identifying bibliographical records and concluding that author links to different authority files must be considered "matches") *and* the DNB holdings are just one of several huge collections serving for mining GND links this implies that the improvements in the GND have almost no impact on VIAF: The insufficient "non differentiated" record is still connected with significant holdings known to VIAF (and maybe there never was a DNB record involved - keep in mind they never have items by foreign authors in foreign languages) and recent VIAF improvements can be interpreted as giving older records some kind of gravity. Thus the cluster does not change and the more suitable newer and individualized GND record stays in an orbit of its own. There are 2.75M suitable GND records for persons opposing 4.5M "non differentiated" ones and the overlap (= number of records where action is needed) must be enormous.

Fortunately enough people at DNB are aware of the problem and its impact on their regular business: Creating new individualized records is necessary, but the real benefit is only yielded if they can get rid of the non differntiated ones in the authority files, i.e. it has to be ascertained that nobody is using them any more - a hard task especially if you don't want to multiply the intellectual efforts of rechecking everything everywhere. They are considering to implement a VIAF-like matching between the large GND-using datasets: When bibliographical records can be identified and one set is linked to an differentiated authority record and the other with an undifferentiated, then they can either infer identity or - in the case several differentiated authority records are involved - can at least automatically switch the linking to the better suited record. After this operation they hope to be able to mass-destruct a seven-figure number of non differentiated records. When these are out of the way VIAF in turn will be able to construct much more adequate clusterings. And maybe in the interim VIAF will devise more clever dealings, for instance allowing two GND records in the same cluster if they are of different kind (but that ist pure speculation).

As for en:WP the decision was made to connect articles with VIAF where no direct means of coping with duplicates exist or even can be thought of. This is not comparable with the approach of de:WP to connect to a real Authority File where also a workflow of cooperation could be established. But in de:WP we additionally note LCCN and VIAF numbers and also have to cope with all kinds of discrepancies - those by duplicates/inconsistencies within GND and the additional ones evoced by the fact that VIAF is not perfect. Therefore also at de:WP we always have to keep in mind that we are not a documentation centre for authority file peculiarities (we'll never be alerted if somewhere someone fixes a problem which we had noted, except in the case the article is included in the authority data workflow supported by maintenance categories which guarantee some kind of follow-up editing). But on the other hand authority numbers in wikipedia articles are an utility to link into and out of wikipedia and therefore one has to settle for "the best" number available (at the moment of editing). Thus one has to assess the VIAF clusters in question by several conflicting criteria: Size of the cluster, which cluster includes the LCCN, is the LC record without birth and death dates, are there LCCN duplicates, which cluster holds the most records with full birth and death dates (this might turn out to be the most stable over time and "attract" the others) and so on.

Personally, I would record anomalies as comments in articles very sparingly, perhaps only if no "main" VIAF cluster can be identified, e.g. there are two or more of roughly comparable size and quality and both of them involve more than a single record. As you see with your example of Emily Gravett, since June VIAF has been able to identify the clusters in the meantime, thus http://viaf.org/viaf/232644200 is now technically a redirect to http://viaf.org/viaf/70060109 . And of course VIAF did not send a "trace" notice to en:WP.

I'm quite certain that this didn't answer your question. But it's all like Conway's Game of Life: There are so many bodies involved, all of them plotting their next actions based on their own needs and at best loose coordination with other interested parties. There is no central plan, not all pitfalls are known and nobody knows best. And in six or twelve months every one will be more experienced but the situation will be even more complex because libraries et c. are even less able to mass-revert tens of thousands of individual edits if hindsight demands an adjustment of course. -- Gymel (talk) 00:13, 5 January 2013 (UTC)Reply

Hi, Thanks for your time. I do see now that VIAF has identified the Emily Gravett clusters since I added {{Authority control}} and my comment to her biography 2012-06-29. That was one of my earliest instances, the one that I used for copy-and-modification in the footers of other instances. Of course I should have checked Gravett at VIAF before using that instance for illustration 2012-12-19 and again yesterday!

--P64 (talk) 21:55, 5 January 2013 (UTC)Reply

Hi, Gymel. Are we ready to use template AC on redirect pages?

Anyway, if you do continue with work such as 12 April 2015

• Jerome Ackerman ‎ (+AC, +PD)
• 07:05, 12 April 2015 (diff | hist) . . (-185)‎ . . Jerome and Evelyn Ackerman ‎ (→‎External links: moved AC to Jerome Ackerman and Evelyn Ackerman) (current)
• 07:05, 12 April 2015 (diff | hist) . . (+348)‎ . . Evelyn Ackerman ‎ (+AC, +PD)

where the target is a joint biography such as Jerome and Evelyn Ackerman, please tag the redirect(s) with {{Redr|to joint biography}}.

Today I revised about 50 redirects from 18:45 to 20:00 with edit summaries "{Redr|to joint biography}}" or "{Redr|to joint biography}} {{DEFAULTSORT:" [1].

--P64 (talk) 21:45, 21 April 2015 (UTC)Reply

P.S.

1. I don't understand where we should use _ _STATICREDIRECT_ _ so I neither add nor delete it.
2. I have numerous AC templates and also ISFDB identifiers at User:P64/FSF/Children's/VIAF/Joint ready to use, primarily compiled 2012/2013 at another page in my user space.

--P64 (talk) 21:56, 21 April 2015 (UTC)Reply

Hi P64, we seem to have been ready all the time since, see how Wikipedia:Authority control#Non-1:1 and non-exact matches instructs us to create redirects for that purpose when necessary...

I will use {{R to joint biography}} (same effect as {{Redr}}?) next time in that situation, but I don't think there will be coming much: My recent activity here had to to with sifting through d:Wikidata:Database reports/Constraint violations/P214#"Unique value" violations where among others the following typical case occured: en.wikipedia provides a VIAF number in the (i.e. one or illegaly several) AC template of a joint biography article. When this is imported to wikidata usually a conflict occurs because the same VIAF number is also listed at the wikidata entry of the real person. It may be that some more cases are buried in the complementary report d:Wikidata:Database reports/Constraint violations/P214#"Single value" violations when VIAF numbers for more than one real person have migrated into one wikidata item and there are no items for the real persons (at least with VIAF numbers) to trigger the conflict at a place easier to spot. -- Gymel (talk) 22:58, 21 April 2015 (UTC)Reply

Thanks.

1. I have generally switched to using {{Redr|} --created or managed by Paine Ellsworth-- on line three following blank line two, because it takes multiple redirect classifications as parameters, such as {{Redr|to joint biography|to section|from move}}. Secondary parameters "toggle" the messages; I haven't learned them but see e.g. Template talk:This is a redirect#Language codes and WT:WikiProject Redirect#Categories for "double" redirects. {Redr} hides the template messages, which costs one extra "click" and costs display space for one short message (as for {{R to joint biography}} alone), saves space for several messages or one long one.

2. As I understand you, we should add template AC to Redirects from people --and to some redirects from alt names-- but leave to Wikidata the creation of new items and import of authority identifiers from our redirects. As WD incorporates wikipedia redirects by such uploads, then its d:Wikidata:Database reports/Constraint violations will reveal buried errors in its identifiers data. (This makes an argument, too, for adding to WD those people who are at VIAF but have no wikipedia pages anywhere.)

3. Anyway, Category:Redirects to joint biographies increased from about 100 to about 150 pages by my follow-up to your recent work, mainly. Many were entirely unclassified redirects.

Thanks for your prompt reply. --P64 (talk) 17:43, 22 April 2015 (UTC)Reply

ad 2) Unfortunately Wikidata can't import anything "from our redirects" since redirects are not represented in wikidata at all. Thus the primary motiviation really is to do something better than deleting them with the AC templates not appropriate in joint biographies: Moving them to a redirect page is simply a means of saving them. Empty AC templates in redirects however are probably of not much use, at least at the moment.

On the other hand we can use authority control: Having one VIAF number twice here (in an article and a redirect to a joint biography) tell us that it is not yet recognized that the article and the redirect pertain to the same person. And having the same VIAF number in a redirect here and a proper item in Wikidata (albeit not linked to this wikipedia) will come handy if/when wikidata will have a strategy how to link to redirect pages: Those here with AC numbers then can easily integrated into existing items. -- Gymel (talk) 21:37, 22 April 2015 (UTC)Reply

Thanks. Last year I skimmed the 2013 discussion D:Wikidata:Requests for comment/A need for a resolution regarding article moves and redirects and concluded that I should store authority identifiers for joint biographies in my user space --or article talk space, when I started Talk:Ingri and Edgar Parin d'Aulaire#Authority control data. Now I will begin to add/move it to personal redirects.

1. Should existing links from WD items to Wikipedia redirects (as at D:Q5372432) be deleted? Only last hour I reported at D:Talk:Q7767892 "Emily Windsnap and her Wikipedia redirects" concerning Emily Windsnap.

P.S. At D:Q7767892 I tried to add enwiki:The Tail of Emily Windsnap but I don't know how --nor yet how to wikilink my discussion there. Then I revisited D:Wikidata:Notability "a sitelink cannot point to a redirect.[4]" and the RFC linked above, and then returned here. --P64 (talk) 23:25, 22 April 2015 (UTC)Reply

2. Do you know, is there a general policy concerning hard inter-language links? Example at ES.wiki [2] (which I added yesterday); contrast one at DE.wiki [3] (not mine), where comment disables the links.

Good night, gute nacht, --P64 (talk) 23:14, 22 April 2015 (UTC)Reply

--P64 (talk) 23:14, 22 April 2015 (UTC)Reply

Hi P64. ad 1) AFIK for editors it is not possible to add "language links" to redirect pages in wikidata. For robotical bulk creation of items in the early stages of wikidata or in case of redirects which where automatically created by moving wikipedia articles the situation is not uncommon. If they would really harm someone would take the work of detecting and removing / delinking them from wikipedia in an automated manner. But the main point of the long Bonnie-and-Clyde diskussion you brought back into my mind is not that these kind of links are harmful but rather they cannot provide the service wikidata was created for, namely 2) traditional interlanguage linking had become such a mess that it became unmanageable. During the initial imports existing interlanguage links were exploited to assign the articles of a given wikipedia to already created wikidata items and the interlanguage links were removed. However if assignment was not possible because another article of the wikipedia in question was already present at the item the interlanguage link was kept in place. Instances of interlanguage links <!-- commented out --> IMHO are traces of an older practice: In times before wikidata there were many bots patrolling for the dreaded "interwiki conflicts": I.e. when article A here linked to X in some other wikipedia the bots tried to insert there a link back to our A. But when X already had a link to english wikipedia but to a different article B they solved the situation by commenting out the link to X in A (in the assumption this was the more recent change). It's well possible that also human editors used the comment syntax to note candidates for interwiki links they did not or could not choose for different reaseons, the viability and admissibility of interwiki linking to redirect pages was a sore spot already in these times past. -- Gymel (talk) 13:13, 23 April 2015 (UTC)Reply

Hi,
You appear to be eligible to vote in the current Arbitration Committee election. The Arbitration Committee is the panel of editors responsible for conducting the Wikipedia arbitration process. It has the authority to enact binding solutions for disputes between editors, primarily related to serious behavioural issues that the community has been unable to resolve. This includes the ability to impose site bans, topic bans, editing restrictions, and other measures needed to maintain our editing environment. The arbitration policy describes the Committee's roles and responsibilities in greater detail. If you wish to participate, you are welcome to review the candidates' statements and submit your choices on the voting page. For the Election committee, MediaWiki message delivery (talk) 14:22, 24 November 2015 (UTC)Reply