Wikispecies talk:Project Wikidata

From Wikispecies
Jump to navigation Jump to search

A new project page, like we did with Project cleanup: What we can establish, is a type of mirror page to Wikidata:Wikispecies, where discussions on Wikidata with relevans to Wikispecies can be reported, and how we can as community collaborate our opinions, point of views and consensus versus the Wikidata page. The most important part may then be to take an active part in the collaboration of the integration of Wikispecies and Wikidata. Dan Koehl (talk) 13:55, 15 April 2015 (UTC)

Why Wikidata didn't include us until now?[edit]

It is interesting to know why WD didn't include us when they added all the other wikis. Is anyone there we could ask this question? Perhaps if we knew why, we could address the problems which prevented our inclusion. Mariusm (talk) 14:41, 15 April 2015 (UTC)

I agree that's a relevant question, which may be addressed at Wikidata. @Mariusm: I suggest you ask that question at Wikidata_talk:Wikispecies. Dan Koehl (talk) 15:31, 15 April 2015 (UTC)
When I read this discussion, there was the idea to include Wikispecies to an extent, that after inclusion it would be superfluous. Probably, such an idea was not favorable for cooperation. Anyway, editing Wikidata objects, in my opionion is a cumbersome issue, when you don't understand the architecture behind. As far as I see, organisation of the stuff in Wikidata is not at all suitable for presentation directly to a broader public. Even if all Wikispecies data would be incorporated into Wikidata, some kind of interface for visitors will be necessary. If not, we would have a burying ground for taxonomic data: Every information is present in some data objects/properties/statements, but virtually inaccessible for casual visitors.
Yes, of course, administration of interwiki links also for Wikispecies is desirable, but the mission of Wikidata goes far beyond that, at least in concepts. If these concepts are realised, it seems possible to me, that finally every bit of information has to be added by editing Wikidata objects, and Wikispecies then will only be a visitors interface. This would require a much higher extent of standardisation. Moreover, editing will be more difficult for people, who only have a good background of taxonomic knowledge, but who do not want to become some kind of database nerds. Regards --Franz Xaver (talk) 15:50, 15 April 2015 (UTC)
Having first edited Wikispecies, and then more recently Wikidata, I find the latter easier. True, it is something new to learn, but that is not difficult. Once certainly doesn't need to be a "database nerd", and it's possible to make tools to do the editing via other interfaces (see "the Wikidata game" for several examples) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:27, 15 April 2015 (UTC)
There is no special reason (see Access for remaining sister projects). The sister projects are added step by step. --Succu (talk) 16:41, 15 April 2015 (UTC)

I would very much like to see the generally sound information invested in WikiSpeciesused as a foundation for the representation of biological matters on the various language versions of Wikipedia. Otherwise there will always be differences between the taxonomic validity of (more populist) Wikipedia articles and the (probably more scientific) WikiSpecies information. However, taxonomy is a bit like an iceberg. 90% of the data is historic and fixed, but 10% floats in the modern era and needs continual updating. More strangely, the top of one iceberg occasionally has to be transferred to a different underwater base. In database terms, this may present something of a challenge. If new evidence means that a species name has to be transferred to a different genus, of if a sub-species name has to be raised to species status or even transferred to a different species name, then changes have to be made and the transparency of the interface will be paramount in ensuring that changes are made correctly. Its easy to create a WikiData item stating that the river Thames runs through London in England, as this is not an issue liable to change. However, it has to be possible in a taxonomic system to move species coridon from genus Papilio to genus Lycaena and then to genus Lysandra as new taxonomic evidence requires. Any taxonomic data construction in WikiData must be able to cope with this kind of change and the method for making such changes must be transparent to Wiki editors who do not necessarily have a deep understanding about Database construction. I therefore agree with Dan's suggestion that we need to investigate an initially non-public system of migration or cross-linking from WikiSpecies to WikiData to see if a single underlying taxonomic authority can be generated. But the system has to be flexible enough to allow changes when new scientific evidence is published and open enough to allow such changes to be made by editors who are not database experts. Accassidy (talk) 18:45, 15 April 2015 (UTC)

Changing a species name, or genus or taxon's rank, in Wikdiata cana lready be done, and is trivially easy. No knowledge of "Database construction" is needed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:27, 15 April 2015 (UTC)
Accassidy "Any taxonomic data construction in WikiData"? means what? Highjacking the newest name combination? --Succu (talk) 21:39, 15 April 2015 (UTC)
@Succu: It means exactly what it says: The manner in which taxonomic data is structured in WikiData must be changeable and self-evident. Is this statement easier for you to understand? Accassidy (talk) 09:06, 16 April 2015 (UTC)
WikiProject Taxonomy gives an overview. --Succu (talk) 13:12, 16 April 2015 (UTC)
@Succu: What Accassidy really means is that we have at WS a pretty high rate of page-name-changes/page-movements compared to other wiki projects. Once all links to WS are established in WD, would it be possible to run an automated software agent, say every 3 months to update all the pages which are linked to WS to reflect the changes done at WS? Mariusm (talk) 14:30, 16 April 2015 (UTC)
Possible: yes; sensible? Not at all. Why woudl we store data twice, with only occasional synchronisation? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:42, 16 April 2015 (UTC)
How do you organize it today? Manually? --Succu (talk) 16:35, 16 April 2015 (UTC)
@Succu: @Pigsonthewing: Have a look at the WikiData and WikiSpecies pages for the same taxon, for example WP + WD, and make a comparison of the information available under these two systems. The WikiSpecies page shows higher taxonomy, subordinate (child) taxa, name and type information, synonymy and references, and includes two images. If you can transfer all this information into the WikiData page for this species, so that they have equivalent contents, and also show that this underlying data has been referenced and included on the Wikipedia page(s) for Jamides bochus, we will have some idea about how to integrate the two systems. Accassidy (talk) 21:27, 16 April 2015 (UTC)
Please see https://tools.wmflabs.org/reasonator/?q=Q6146500 for a demonstrator of how Wikidata's taxonomic data can be rendered in the manner you describe - there is no need for the data storage to be changed. It is also already possible to transclude Wikidata data into infoboxes (aka taxoboxes); and Wikidata also stores references. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:10, 16 April 2015 (UTC)
I added some information to Jamides to show you how it works. You can add types to species as well. We have a taxobox to illustrate higher taxonomy. You can make use of my bot, if you want to add references to a lot of item properies (eg. Squirrels of the World). If you miss something we can propose a new property. It works, but the information in your page has to be transferred to Wikidata of course. --Succu (talk) 09:30, 17 April 2015 (UTC)
@Succu: @Pigsonthewing: Can you explain the derivation of the "Q_numbers" such as Q1762313 for Jamides and Q6146500 for Jamides bochus? Are they just added sequentially when new pages are created, or is there some sort of system? Do you have a way of showing "Child Taxa", such as species on a genus page or subspecies on a species page? Do you have an idea of the work and time involved in migrating all the current detailed taxonomic information into the necessary categories for completion of equally full WikiData entries? IT looks like a lot of work to me and there seems to be much more detailed data in WikiSpecies as a whole. Accassidy (talk) 20:48, 17 April 2015 (UTC)
Q-Numbers are like database IDs. One way to get "Child Taxa" is WDQ. Sure I have an idea what time it takes to migrate WS references into WD references and to sort out phantom references. BTW: Who moved Papilio bochus to Jamides bochus --Succu (talk) 21:44, 17 April 2015 (UTC)
@Accassidy: As I said above: "Please see https://tools.wmflabs.org/reasonator/?q=Q6146500 for a demonstrator of how Wikidata's taxonomic data can be rendered in the manner you describe". Q numbers are issued sequentially; and carry no semantic meaning. A FAQ may eb found at d:Help:FAQ. The amount of time to migrate data is not an issue; it will happen anyway, and the bulk of it will be done by bots. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:29, 25 April 2015 (UTC)
@Succu: @Pigsonthewing: If you look at Jamides bochus on Wikispecies you will see that Papilio bochus was reassigned to Lampides by Fruhstorfer and then to Jamides by Hubner. How does the entry in WikiData for Jamides bochus reflect these earlier combination? The WikiSpecies page also lists 23 chid taxa (subspecies) of Jamides bochus yet the WikiData page carries no information on such child taxa. If WikiData is to replace WikiSpecies as a reliable taxonomic data storage facility then it must make these earlier combinations, any current synonyms and child taxa all explicit when accessed from another Wiki. I have looked at WDQ and I am no clearer as to how this enables access to child taxa. The system is far too opaque to be understood readily by users perhaps a little less familiar than yourselves. I still need more information on how WikiData/reasonator might cope with these basic taxonomic issues. Accassidy (talk) 17:18, 25 April 2015 (UTC)
I have only looked at a few examples, in taxa I am very familiar with so I did not need to look anything up (I am a professional turtle taxonomist so that's what I looked at), however I have to concur with @Accassidy: the struct you are using seems far to basic for taxonomic data, particularly in older groups. Every name ever used needs to be accessible, whether they have been replaced, split or lumped or any of another dozen possibilities. The ones I looked at the parent taxa did not seem to be correct and not all sisters where included, your imported data has come from incomplete wikipedias (frWP for what I looked at) and based on outdated reviews (Fritz and Havas 2007). The maintenance of taxonomic databases requires a reading of the complete literature on a species, I accept that even Wikispecies has not achieved this, it has to not only stay up to date but also go back to the source (as an example for the genus Chelus you need papers going from 1735 to 2015 covering some 12 species just to analyse 3 species. You have another issue in that your structure must obey the rules of nomenclature, including its ruled exceptions. The closest anyone has come to a database that fully does this is Zoobank, but even that is essentially manually updated. I am happy to help you with these issues, I can even show you the struct for the Zoobank database since I have it. However to go anywhere with using this as a taxonomic database it will need to have the basic taxonomy set up in the hierarchy that is accepted under the rules of the ICZN and the ICBN. Cheers Faendalimas talk 18:34, 25 April 2015 (UTC)
The structure used on Wikidata is being built by knowledgeable taxonomists; you are of course welcome to contribute; and to fix the issues you say you found; just as people fix issues on Wikispecies. However, while you say you "concur with Accassidy the struct you are using seems far to basic for taxonomic data", I don't see him saying that; and it certainly isn't the case. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:00, 30 April 2015 (UTC)
I realise I am sounding harsh, its not meant that way I actually applaud what your trying to do, but when when building a database of the necessary complexity of a nomenclatural one, the more problems predicted early and dealt with the better. So I looked at it in a deliberate effort to break it, but not with cynasism, to try to see what can be improved so someone else does not. I am happy to assist and can certainly provide the data on the turtles with good references. Cheers, Faendalimas talk 22:15, 30 April 2015 (UTC)
@Accassidy: You can see for yourself what the entry in WikiData for Jamides bochus looks like; and that it deals with the Papilio name through its link to d:Q19810423. Wikidata also has capacity for information on child taxa, and I've just created an entry for Jamides bochus argentina so you can see how that works. There is also a Reasonator page for the subspecies; but note that I gave Reasonator as an example of how Wikidata material can be rendered in different ways, not as the definitive way it will be rendered for taxa. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:52, 30 April 2015 (UTC)

Puzzle[edit]

How can I find out which articles describe an taxon, an author or a reference? --Succu (talk) 20:50, 17 April 2015 (UTC)

As far as I understand, your question is about some characteristics that allow automatic recognition? --Franz Xaver (talk) 21:05, 17 April 2015 (UTC)
Yes, it is. --Succu (talk) 21:46, 17 April 2015 (UTC)
A Taxon has allways "== Taxonavigation ==" on the first line (or in the case there's a taxon image - on the second line).
A reference is always a template, and has no regular page of its own. Mariusm (talk) 06:58, 18 April 2015 (UTC)
Unfortunately there are some thousand pages without "== Taxonavigation ==" eg. Axiocerses tjoane, Rhizopoda, Gymnodiniphycidae, Pelagophyceae, Octopicolidae etc. --Succu (talk) 08:45, 18 April 2015 (UTC)
It seems there have been some alternative formating of this, something that needs to be weeded out. However in the absence of "== Taxonavigation ==" it will usually have "== Name ==" further down for a taxon name, your example of Axiocerses tjoane has this. Cheers, Faendalimas talk 01:37, 20 April 2015 (UTC)
The reason some pages do not have the ==Taxonavigation== title is because it is a duplication of the same title in the collapsible Taxonavigation box. A lot of my pages had this simplification in the past. If the title is really needed as well as the taxonav box, then they will have to be reinstituted. Accassidy (talk) 17:19, 20 April 2015 (UTC)

A request (not sure if right place to ask)[edit]

I'm not intricately familiar with Wikispecies nor Wikidata, but once they are integrated, I assume it would allow a WikiSpecies taxon entry to display all the Wikipedia articles on said taxon. Currently, it must be done manually, e.g. adding en:Latrodectus hesperus and nl:Latrodectus hesperus within brackets to Latrodectus hesperus, a species with articles in a dozen languages. Prior to full integration (whatever that may entail) I'm curious as to whether a bot command could be executed that would scan Wikidata and automatically add all available language links to Wikispecies taxon pages. Animalparty (talk) 00:35, 20 July 2015 (UTC)

Wikidata obviously has other problems than continuing the integration of sister projects :-( --Murma174 (talk) 21:31, 16 August 2015 (UTC)

3 month later[edit]

Is there anyone who tried to got a little bit more experience with our data model? --Succu (talk) 20:36, 16 August 2015 (UTC)

I am not sure, how to proceed in cases like Monnina wrightii. In Wikidata, there is d:Q15577519 (Monnina wrightii) and d:Q17468651 (Pteromonnina wrightii). Should these be merged to one? --Franz Xaver (talk) 09:59, 17 August 2015 (UTC)
The same with Monnina tristaniana, Monnina richardiana, Monnina resedoides, Monnina dictyocarpa, Monnina cardiocarpa. --Franz Xaver (talk) 10:06, 17 August 2015 (UTC)
Sorry for the delay. We use Property:P566 (=basionym) to connect them. --Succu (talk) 08:48, 5 September 2015 (UTC)
As far as I understand, this does not produce an interwiki link from e.g. sv:Pteromonnina richardiana to vi:Monnina richardiana. As it is all about the same species, in my opinion, there should be interwiki links. --Franz Xaver (talk) 07:38, 10 September 2015 (UTC)
Ideally the software should virtually create sitelinks based on certain properties. But this is not implemented. So you have to assume some POV and move the sitelinks to one of the items. Sometimes this would make someone unhappy. --Succu (talk) 15:40, 14 September 2015 (UTC)

Sitelinks[edit]

Sitelink integration is anounced for 20.10.2015. --Succu (talk) 16:32, 9 September 2015 (UTC)

Symbol thumbs up.svgTommy Kronkvist (talk), 20:24, 17 September 2015 (UTC).

Spelling of taxon names[edit]

Some time ago I've made some simple spell checks and found a lot of problems, mostly related to gender. Maybe someone is interested helping to resolve these problems. It would make the integration of sitelinks to wikispecies a little bit easier. --Succu (talk) 20:01, 17 September 2015 (UTC)

There is one general problem concerning epithets ending with -cola. These are not adjectival epithets, but nouns in apposition. E.g. agricola (farmer), monticola (inhabitant of mountains) etc. So, they always will be spelled -cola, irrespective of gender of the genus. Many authors have treated these like adjectives, but IPNI usually gives the correct form. Correction follows ICN Art. 23.5. However, I do not know, whether there are rules in the zoological code that apply here. I have to add one additional spelling error: Pteromonnina exaltata (d:Q17468661) has to be corrected to Pteromonnina exalata (d:Q15584427) - see Monnina exalata.
Im not sure, how we can help. E.g. Acer martini (d:Q9578046) will require to move the article in pt.wiki to the correct name Acer martinii. Maybe, I will be able move the page, but if someone is opposing, I will have a hard time to explain the nomenclature code in Portoguese. Is it about giving hints which of two alternative spellings is the correct one, according to the respective code? Or is it about executing corrections in a wide scope of language versions? --Franz Xaver (talk) 22:41, 17 September 2015 (UTC)
Meanwhile I made an attempt to solve the problem with Barleria kitchingi, but I received some error message when merging d:Q15346468 into d:Q17676543. --Franz Xaver (talk) 23:22, 17 September 2015 (UTC)
Dear Franz Xaver, as I said, the spell checks are simple. If it helps I can create a separate page for epithets ending with -cola. Redirecting is one possibility. I prefer to mark them as duplicated page (an example is d:Q10398559). One reason for this solution is d:User:Succu/zh article duplicates. Regards --Succu (talk) 18:14, 18 September 2015 (UTC)

Wikidata links enabled[edit]

Work has begun on Wikidata integration; stage one enables "interwiki" links. Here's the announcement. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:47, 21 October 2015 (UTC)