User talk:Lockal/Archive/2014/10

EOL

Hi Lockal,

I see you are adding links to EOL. In itself this is fine, but it is necessary to check every time if there is meaningful content at EOL. I see several cases of links you have made where EOL has no content, and where the link thus is wasteful (something like 80+% of the entries at EOL is empty). - Brya (talk) 17:25, 16 October 2014 (UTC)

EOL overview pages might look empty, however full pages are not empty. EOL always provides taxonomy, nomenclature and taxonomic hierarchy. What is special about EOL is that it always provides aggregated references to content providers and literature references. So if we add links to EOL, we could later use them to cross-validate almost any wikidata-statements for species (including references for other databases). --Lockal (talk) 07:54, 20 October 2014 (UTC)

Nope, most EOL-pages do not merely look empty, they are empty. There is some standard junk posted to make it look like there is content, such as an ITIS taxonomy, but that does not count. ITIS is mediocre at best, and for that matter it is already here (it is too much work to get rid of it, although Wikidata would be better of without it. EOL does not have nomenclature, it just has an entry for every name, or variant of a name, found somewhere on the www; an immense amount of junk. Everybody in search of information should stay away from it, if possibly they can.

        There are good EOL pages, that are worth linking to, but they must be individually checked.

        Theoretically, we could indeed add links to other sites that have no meaningful content, and which consist only of links to sites that have no meaningful content either. But all that will give us is a humongous amount of links to empty sites, that exist only for the reason to provide links, in the hope that somebody someday will add content. One big project to cheat the reader, and to clog up the wwww. - Brya (talk) 10:52, 20 October 2014 (UTC)

There is no such thing as "good", "bad" or "mediocre" statements in Wikidata. Statements can be correct or wrong. Mass-adding EOL identifiers leads to a number of constraint violations, which reveals number of incorrect statements, imported from Wikipedia. Of course, I can create a list of violations without adding EOL identifiers to Wikidata, however this process is time consuming and I am not going to redo everything each time some Wikipedia bot creates duplicate articles about synonyms of species. Treat this like Authority control links: you can but not bound to use these links, as long as you have better sources, but if you have none, database links are better than nothing. --Lockal (talk) 11:46, 20 October 2014 (UTC)

You can play with words any way you like, but there is a lot of junk in EOL (empty pages, phantom pages) and if you are adding links to EOL pages without checking, you are actively degrading Wikidata. - Brya (talk) 16:21, 20 October 2014 (UTC)

Can you give me a link to any "phantom" page in EOL I added recently here? I do a number of checks before adding statements to Wikidata, all statements are easily verifiable and I can't even imagine what could be wrong in my edits.

If you think that EOL provides unreliable content and inappropriate for Wikidata, you should request property deletion instead of arguing with anyone, who adds links to EOL. --Lockal (talk) 20:34, 20 October 2014 (UTC)

EOL does provide unreliable content and inappropriate for Wikidata, but in addition it also provides reliable content that is appropriate for Wikidata. It needs checking, page by page. If EOL had only uniformly good pages the proper strategy would be to use a bot, as with so many other databases.

        I did not find any phantom pages you added, but this was unlikely in the first place as I did not check all that many pages. A lot of those I checked were empty and useless. The marginal case would be something like Platymantis which has no content, but does offer good pictures. This could be argued either way, although we certainly could get along perfectly well without it. - Brya (talk) 06:01, 21 October 2014 (UTC)

Encyclopedia of Life ID (P830)

Hi Lockal, what's the source of your massive addition? --Succu (talk) 21:31, 16 October 2014 (UTC) PS: The result. --Succu (talk) 08:51, 18 October 2014 (UTC)

The source is EOL itself (so it passes Help:Sources/Items not needing sources), by checking ITIS TSN (P815) and taxon name (P225) (if any), while skipping pages which violate "single" constraint for aforementioned properties. I expected such kinds of violations; they come either from incorrect data import from Wikipedia or wrong interwikis. What I can do is to remove such statements (and refill properties with correct values), however I am not too accustomed to do it with Widar. However it does not solve problems with wrong interwikis, so probably it should be solved manually. --Lockal (talk) 08:26, 20 October 2014 (UTC)

Your explanation is not sufficient to explain for example this addition. Monera is the name of a kingdom that is not in use anymore, but it certainly is not the same as Archaea as the EOL-link suggests. Lymantria (talk) 16:40, 22 October 2014 (UTC)

The information comes from http://eol.org/pages/7920/names. I sent a message to EOL editors to sort things out. Thank you for noticing! --Lockal (talk) 08:36, 24 October 2014 (UTC)

taxon name (P225) = Pseudicius marshi (Q2127321): TSN indicates Pseudicius marshi, but you added the objective synonym Hakka marshi. So your explanation wents a little bit short.--Succu (talk) 20:57, 24 October 2014 (UTC)

I used this API-query for checking, if EOL page belongs to specific taxon. Both "Hakka marshi" and "Pseudicius marshi" are present there as canonical form from different sources. Similar pages in web UI: [1]. --Lockal (talk) 15:51, 25 October 2014 (UTC)

Pomadasys multimaculatus (Q6430525): -um or -us? --Succu (talk) 22:33, 26 October 2014 (UTC)

Wikidata weekly summary #114

Here's your quick overview of what has been happening around Wikidata over the last week.

• Other Noteworthy Stuff
  • The Wikidata Game now has a 'Commons Categories' game
• Did you know?
  • Newest properties: At the Circulating Library ID, MacTutor id (biographies), allmovie identifier, number of survivors, given name version for other gender, name in native language, tempo marking, manifestation of, zbMATH author ID, Executive Order number
• Development
  • Made significant performance improvements (to be rolled out next week)
  • Worked on usability improvements to the editing of sitelinks
  • When you link to an image on Wikimedia Commons in a statement it will now show up in “global usage” there.
  • Made the phpunit tests for Wikibase much faster
  • Wikibase phpunit tests on travis pass with hhvm now
  • Fixed label/description uniqueness constraint checks
  • Work on entity usage tracking
  • Year formatter now shows the year instead of the unformatted ISO string in case of a precision mismatch
  • Diff and old revision pages don’t run the JavaScript UI any more, should be much faster now
  • Introduced a Changers concept to the JavaScript frontend, wrapping the API and entityStore functionality
  • Identified issue causing content in the old serialization format to be included in XML dumps
• Monthly Tasks
  • Hack on one of these.
  • Help fix these items which have been flagged using Wikidata - The Game.
  • Help develop the next summary here!
  • Contribute to a Showcase item

Read the full report · Unsubscribe · John F. Lewis 20:28, 25 October 2014 (UTC)