Wikidata talk:WikiProject Taxonomy/Archive/2015/07

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Which taxon

Whoch one of these are the 'real' taxon? Lacmellea (Q20640783) vs Lacmellea (Q9019776)? If someone merges them, please fix Lacmellea lactescens (Q20640777) and c:Category:Lacmella lactescens as well. (t) Josve05a (c) 23:51, 3 July 2015 (UTC)

I don't really inderstand the question, but the correct spelling is Lacmellea. - Brya (talk) 03:49, 4 July 2015 (UTC)

dingo (Q38584) and Canis lupus dingo (Q5032447)

According to en:Dingo (disambiguation) Q38584 is common name and Q5032447 is taxonomic name. I think instance of taxon has to be removed for Q38584 and also some of the language links has to be moved to Q5032447. --Termininja (talk) 07:34, 16 July 2015 (UTC)

These things tend to be 'loaded'. There are a lot of pages involved. A page like nv:Nahatʼeʼiitsoh Bikéyahdę́ę́ʼ łééchąąʼí ałchiní appears to be strictly taxonomy, but most of the other pages will have mixed content. The viwiki pages are clearly separated along the lines you suggest, but the other way round. A quick fix would be to switch the enwiki pages, and then remove "instance of taxon" for Q5032447. However, to be safe it would be better to consult all the pages involved and inventory all the taxonomic concepts. - Brya (talk) 10:59, 16 July 2015 (UTC)

I just checked by Google translator the viwiki Chó cỏ (from Q5032447) and it not looks like article about taxonomy, I think it is not even about the Dingo..., I think viwiki link is probably wrong, because the article is in Category:Vietnam animal breeds (the template is also for dog breeds). This link probably needs a new item in Wikidata. And the other link for viwiki in Q38584 looks like mixed article (Dingo + taxon). --Termininja (talk) 13:00, 16 July 2015 (UTC)

For viwiki related things you can try to ask @Cheers!:. --Succu (talk) 13:10, 16 July 2015 (UTC)

Also I think the same taxon name (P225) is wrong to be used in more than one item. So, why is not forbidden? --Termininja (talk) 12:46, 16 July 2015 (UTC)

We try to avoid this. --Succu (talk) 13:10, 16 July 2015 (UTC)

Yes, like I said Chó cỏ is not about taxonomy. Looking at Google translator, the topic may be "dog breeds in Vietnam" (one of which is Canis lupus dingo). Like I said, these things tend to be 'loaded'.

        The most sacred principle of Wikipedia is NPoV, that is, to accommodate multiple points of view. There is no reason why there couldn't be three or four items with the same taxon name (P225). In practice it is fairly unusual (but see Sylviidae (Q187014), Sylviidae according to Sibley (Q688840), Sylviidae sensu lato (Q6880139)), as the users in Wikipedias are not very diligent in following policy. Theoretically a lot could be done with the right references, but the latter also tend to be few. - Brya (talk) 16:35, 16 July 2015 (UTC)

I am looking for reliable sources, if it is the same, we can merge.--Cheers! (talk) 10:05, 19 July 2015 (UTC)

Oh, merging is not an option. - Brya (talk) 16:05, 19 July 2015 (UTC)

"name that may not be used"

What means this description for example in Eunomia viridissima (Q1873181)? It is with accepted status name for taxon, what is wrong with it? I saw this description on many pages. --Termininja (talk) 20:39, 17 July 2015 (UTC)

Good of you to catch this. It is supposed to look like this, but a few "well-intentioned users" had altered it. This is an awkward format, but the best that can be done (as far as I am concerned it really needs an additional property).

        This indicates that there is no taxon that can be correctly indicated by the name Eunomia viridissima, as there is no genus in the family PHYLLODOCIDAE that can be correctly indicated by the name Eunomia. The name Eunomia can be used only for one genus/subgenus (for animals), and it was already in use, so the later name does not count. What does exist are names, but for all practical purposes these are just names, not taxa. For all practical purposes, these are fictitious taxa.

        That WoRMS is making this mistake is somewhat unexpected (it is a well-known case, and has been known for at least seventy-five years). However, sources like GBIF and EoL are not reliable (CoL has a rate of error that is off the scale, and GBIF and EoL have copied CoL). - Brya (talk) 03:44, 18 July 2015 (UTC)

Does this means that Eunomia Risso (1826) non Hübner (1818) (Q20672177) (genus of annelids) has to be deleted, by the way I found and one genus of plants - Eunomia (Q20656046).. and I totally lost myself :D --Termininja (talk) 06:35, 18 July 2015 (UTC)

1. Don't alter other user's comments (this is not what I said))
2. And, yes Eunomia Risso (1826) non Hübner (1818) (Q20672177) can be emptied and redirected to the earlier item. And, yes, a name for a genus with the same spelling may be used once each for animals, plants, and prokaryotes. That makes three (in some cases also for a higher-level taxon). - Brya (talk) 06:55, 18 July 2015 (UTC)

I never change other user's comments, in this case I just added item link to the name (to be more clear), because you wrote "for animals" and I thought you mean the same item, sorry. How to make a redirect for Q20672177, with merge? --Termininja (talk) 07:11, 18 July 2015 (UTC)

For starters, putting in a template may mean that a user who has set a different language may be reading something else. For another thing, who knows what will happen in the future, and in an archive this may read quite differently. And, above all, it is not what I said.

        And yes, a redirect may be made with a merge. - Brya (talk) 09:39, 18 July 2015 (UTC)

You are right, I forgot that here these things are different for different languages, but I knew it that you will see it and will correct me if I'm wrong;) It's already clear. --Termininja (talk) 10:19, 18 July 2015 (UTC)

Redirected, redirected. - Brya (talk) 04:36, 19 July 2015 (UTC)

I see that WoRMS had noticed that this wasn't right. They just failed to implement the consequences at the species level. - Brya (talk) 04:51, 19 July 2015 (UTC)

common (vernacular) name as a class

Although, strictly saying, it is not a scope of this project let's talk about the subject. I propose to mark somehow about which taxa common name is. This is also applicable to polyfiletic taxa. I suppose that following model is useful: Spotted cleaner shrimp (Q3933821) subclass of (P279) cleaner shrimp (Q3002557), or Plasmodiophorida (Q20675800) subclass of (P279) slime mold (Q949817) and slime mold (Q949817) subclass of (P279) protist (Q10892). Another variant is to use part of (P361) in higher taxon itself: see properties in cleaner shrimp (Q3002557). Any objections? --Infovarius (talk) 21:04, 19 July 2015 (UTC)

Use taxon common name (P1843) with a reference. --Succu (talk) 21:06, 19 July 2015 (UTC)

Yes, there are relationships between common names, but these are of many kinds, and do not (necessarily) match taxonomic relationships. Trying to express these in relationships between items will lead to much confusion. - Brya (talk) 03:43, 20 July 2015 (UTC)

@Infovarius: This is language dependant and as such is more of the wikitionary scope. But Aliases ar exactly meant to be used like that. author  TomT0m / talk page 12:56, 20 July 2015 (UTC)

List of empty items for taxonomy

My bot found ~20,000 empty (lost) items (it scanned the links for only 7 wikies), each one with link(s) to wiki page about taxonomy (examp: Q16527983). I think to add instance of (P31) = taxon, but only this maybe will not be enough. I'll try also to find already filled taxon item with the same taxon name (P225) as the interlanguage link, but this not guarantee the match will be correct, so I'll not add other statements. Do you have some suggestions how is better to proceed... ? --Termininja (talk) 11:18, 20 July 2015 (UTC)

Adding instance of (P31) = taxon would bring them at least to the attention of wikidata-todo. --Succu (talk) 12:55, 20 July 2015 (UTC)

Fossil range

Hello. If I want to add fossil range to some item, how is right to do it? By using two kinds of pair properties - temporal range start (P523) and temporal range end (P524) with the names of the geological periods for values and start time (P580) and end time (P582) with the year as value (for example as in Tyrannosaurus (Q14332)), or by using P523 with qualifier P580 and P524 with qualifier P582, how is written in the description of the properties. --Termininja (talk) 08:39, 22 June 2015 (UTC)

@Tobias1984:, what do you think? --Succu (talk) 15:05, 23 June 2015 (UTC)

@Termininja: I think it is better to make separate statements. A qualifier should only be used to further describe a value. For the absolute age, that could be the radiometric dating system for example. I will fix the examples. --Tobias1984 (talk) 17:23, 23 June 2015 (UTC)

I think it solely depends on the source you have: if it says T. rex lived 1234 B.C. than this is a statement for its own. However, most scientific publications might say "we found this T. rex in a geological layer from the Holocene, which at this location means about 12345 B.C." Then, the absolute value further describes the geological period and should be added as a qualifier.  — Felix Reimann (talk) 05:21, 24 June 2015 (UTC)

• I understand, but I'll use the data from http://eol.org/pages/4433634/data, so maybe I have to use start time and end time as in Tyrannosaurus (Q14332) plus earliest date and latest date, all them in separate statements. My idea is this information to can be used in en:Template:Geological range. What do you think, are these the correct four properties (Example for Daspletosaurus torosus). --Termininja (talk) 12:22, 24 June 2015 (UTC)

It would be better to use the original source summarized as („base of the Maastrichtian (Q278141) to the top of the Lancian (Q6483765) or 72.10000 to 66.00000 Ma“). Fossilworks (Q796451) gives 70.6 to 66.043 Ma. --Succu (talk) 18:03, 24 June 2015 (UTC)

Yes, the same values as in eol, I see only the earliest date is updated in paleobiodb to 83.6. But there is one problem - it is not possible to use decimal value for start time (P580). --Termininja (talk) 13:11, 26 June 2015 (UTC)

If you want to add the value 69.1 Ma you have to type in the value −69100000 an set the precision to 100,000 years ([1]). --Succu (talk) 14:47, 26 June 2015 (UTC)

Oh, thank you, I'll use it. --Termininja (talk) 15:26, 26 June 2015 (UTC)

I see these. These look pretty weird. I doubt that EoL is a good source for anything, but this is sure to be wrong. - Brya (talk) 12:34, 27 June 2015 (UTC)

I don't know this, but I saw the property Encyclopedia of Life ID (P830) which has to means that it is enough good source. I see also that the information for Geraniaceae is the same in Paleobiology Database (Q17073815). Which source is good? --Termininja (talk) 14:06, 27 June 2015 (UTC)

Oh no, we have properties for Eol, CoL, GBIF, ITIS, as they are well-known. This does not mean they are reliable sources. EoL has some good stuff (relatively very little), but mostly is a framework, that is empty, or has other dubious databases copied into it. - Brya (talk) 14:21, 27 June 2015 (UTC)

Ok, I'll continue only with the EoL ID. Is it need to delete this data about the fossil periods? --Termininja (talk) 14:29, 27 June 2015 (UTC)

That is probably a good idea, but perhaps Succu or @Tobias1984: has a good idea about a reliable source? - Brya (talk) 16:14, 27 June 2015 (UTC)

I think both, PaleoBioDB and Fossilworks, are reliable sources. But we can't simply copy their data (not CC0 licensed). --Succu (talk) 16:43, 27 June 2015 (UTC)

I thought that PaleoBioDB is pulling all its data from Fossilworks? We do have written permission by John Alroy to copy the identifiers. Not sure about the other data. --Tobias1984 (talk) 16:57, 27 June 2015 (UTC)

It's vice versa. --Succu (talk) 16:28, 22 July 2015 (UTC)

I looked a little at ages in Fossilworks, and I get the distinct impression that these are not recorded but are computed. That is, these are not figures from the literature, but Fossilworks looks at the locations where the taxon has been found ("Distribution"), and then computes the joint age range of those. If few find-places have been registered, the the age is way off. In other words, Fossilworks is not a reliable source for ages. It may become one in the future, but not in the near future. - Brya (talk) 04:43, 28 June 2015 (UTC)

@Brya: I am not totally sure what you are saying. - But I agree that using only Fossilworks as a source can be problematic in certain cases. The absolute-age-range does not give any information how these ages were derived, and that would be important for the qualifiers. Two of the common possibilities are: a fossil is dated directly or it is constrained between two layers that can be dated (dating the minerals of a volcanic-ash layer). It would somehow be nice to give that background information along with a few other radiometric-dating specifics. --Tobias1984 (talk) 21:32, 28 June 2015 (UTC)

Well, if you look in Fossilworks at Geraniaceae, you will see it is recorded as having been found at one place: Huntington Reservoir mastodons. The Geraniaceae page provides an age which matches the age of the "Huntington Reservoir mastodons". That means that Fossilworks only will give reliable ages when it has recorded the full range of geological sites where it was found. This will vary from taxon to taxon. Incidentally, this same applies to component taxa: if we are to believe Fossilworks, Geraniaceae contains only one taxon: Genarium [sic: the focus on particular taxa appears to apply also to spelling]. - Brya (talk) 05:07, 29 June 2015 (UTC)

Extinct taxon

What is the property to mark some species as extinct when IUCN status is not available for it? --Termininja (talk) 20:54, 17 July 2015 (UTC)

There is no property yet. --Succu (talk) 16:28, 22 July 2015 (UTC)

For those which are fossil (the bulk), I put "(fossil)" in the description. A property would be nice. - Brya (talk) 16:51, 22 July 2015 (UTC)

@Termininja: There was no big discussion when we created temporal range end (P524), but the original plan was that setting p524 should imply that the species is extinct (See: Wikidata:Property_proposal/Archive/7#P524). Let me know if that seems reasonable in your use-case. --Tobias1984 (talk) 16:52, 22 July 2015 (UTC)

Yes, but for most fossil taxa there won't be a temporal range end (P524) available. - Brya (talk) 16:55, 22 July 2015 (UTC)

@Brya: Not sure if anyone else likes the idea, but I thought that we could use p524 = Phanerozoic (Q101313) in those cases. Then we don't need a new property and querying extinction would have only one interface. --Tobias1984 (talk) 17:05, 22 July 2015 (UTC)

Looks rather user-unfriendly. Isn't there some other item that could be put in, looking more intuitive? - Brya (talk) 17:39, 22 July 2015 (UTC)

@Brya: How about just creating a new item like p524 = "Unknown geologic time" or "Unknown time"? --Tobias1984 (talk) 17:47, 22 July 2015 (UTC)

Well, there already is the standard option of "unknown value", so it is only worthwhile to create a new item if it is more evocatively named. - Brya (talk) 18:23, 22 July 2015 (UTC)

Hybrids

I'm looking for good example for hybrid item..., maybe Cistus banaresii (Q15330290)? But then coydog (Q3622157) has to be wrong. --Termininja (talk) 21:20, 21 July 2015 (UTC)

Magnolia × soulangeana (Q731443) is a classic case. - 04:53, 22 July 2015 (UTC)

I see various spellings on this ×, even in Magnolia × soulangeana (Q731443) we have Magnolia ×soulangeana (in label and taxon name) and Magnolia × soulangeana (in commons category and image). Is it the same? Maybe this with space on the both sides is right. --Termininja (talk) 05:10, 22 July 2015 (UTC)

The exact placement of the "×" in Magnolia ×soulangeana is a matter of typography. In as far as there is official guidance, it is not to have a space after the "×". - Brya (talk) 05:50, 22 July 2015 (UTC)

I added "parent of hybrids" in liger (Q182573). Is it ok to use here "sex or gender" as qualifier? --Termininja (talk) 05:38, 22 July 2015 (UTC)

Yes. Maybe it would be better to have a separate approach for hybrids of animals, with a hybrid formula instead of a name, and not treat them as taxa. - Brya (talk) 05:50, 22 July 2015 (UTC)

So you mean, after Liger is not taxon, we don't have to use also "parent taxon" and "taxon rank"...? --Termininja (talk) 06:22, 22 July 2015 (UTC)

That would be the idea. As they have neither a taxon name, nor can form a viable population, why treat them as taxa? - Brya (talk) 10:37, 22 July 2015 (UTC)

Just a comment: the clade definition of taxons, individuals classified in a taxon includes all descendants of a the individuals in a taxon. In that view an hybrid belongs to the taxa of his parents ... I guess if he has no descendants it would be ... a clade by itself (a singleton (Q1165112)    one :) and an instance of both the clades of his parents. So indeed, the set of all this kind of hybrid is not a clade. It's obviously a class (wikidata sense), we could define as individuals who are both instances of "species 1" and "species 2". @Brya: Note that if you don't want to treat them as taxons, the more generic classification sytem as class/instance still works and you will be able to mark

⟨ a famous hybrid example ⟩ instance of (P31) ⟨ hybrid of X and Y ⟩

. As I advocate for a long time, the generic classification system can handle taxonomy. Including its edge cases. author  TomT0m / talk page 11:25, 22 July 2015 (UTC)

The concept of "clade" only applies to taxa above the rank of species. - Brya (talk) 16:18, 22 July 2015 (UTC)

@Brya: Why not a little thinking outside of the box ? :) author  TomT0m / talk page 14:08, 28 July 2015 (UTC)

Just a comment: Your „outside of the box“ is based on set constructions? So all „yellow flowering plants“ (flower (Q506)) should be an instance of yellow flowering plant? --Succu (talk) 21:08, 28 July 2015 (UTC)

@Succu: Nope, this is totally unrelated. author  TomT0m / talk page 07:34, 29 July 2015 (UTC)

Fossilworks now in Mix'n'Match

With the help of User:Magnus Manske Fossilworks is now part of Mix'n'Match (https://tools.wmflabs.org/mix-n-match/?). I will also alert the en-wiki Paleontology people so we can find matches for the currently 265945 unmatched taxons. I already noticed some issues with spelling and duplicates. I suggest we gather these issues at Property_talk:P842, so the maintainer of Fossilworks only has to check one page on Wikidata (that is not going to be archived as frequently as this talk page). --Tobias1984 (talk) 12:27, 21 July 2015 (UTC)

That's bad news. Especially the option to create a new item. --Succu (talk) 13:05, 21 July 2015 (UTC)

@Succu: Bad news? From what I have seen on Wikidata the community has always been disciplined with mass creating items and only matching those that we have (That is what I do). And for our existing items (especially fossils) Fossilworks is a vital and sometimes only authority control. --Tobias1984 (talk) 13:14, 21 July 2015 (UTC)

I was expecting this, Tobias1984. Magnus could you please deactivate the dataset until we have a correct on?! --Succu (talk) 07:47, 28 July 2015 (UTC)

I have hidden the dataset from the main view (had to write new code and change the database to do that...). It might still crop up in searches etc. --Magnus Manske (talk) 15:06, 28 July 2015 (UTC)

Magnus, something is wrong with your dataset. Mix'n'Match gives me 8 different IDs for common box turtle (Q1044378), but the correct ID (fossilworks query) is present since 2014. --Succu (talk) 18:45, 21 July 2015 (UTC)

@Tobias1984:? --Magnus Manske (talk) 20:00, 21 July 2015 (UTC)

@Succu: I scraped the IDs directly from the website, so this is probably an artifact from the database that Fossilworks uses. I will inquire with John Alroy what we can do about this situation. - As I mentioned before I think Wikidata will also have to work with, what are essentially upstream-databases, to improve the webs data-landscape. We have a lot of great tools to find discrepancies and we need to communicate them. - Here are examples of multiple IDs with the same taxon:

• http://fossilworks.org/?a=taxonInfo&taxon_no=53094
• http://fossilworks.org/?a=taxonInfo&taxon_no=90482
• http://fossilworks.org/?a=taxonInfo&taxon_no=96386 (+ 5 more) --Tobias1984 (talk) 20:59, 21 July 2015 (UTC)

Tobias1984: Why not simply download the latest version?

„to work with, what are essentially upstream-databases, to improve the webs data-landscape“ - Do you really think we don't do this? --Succu (talk) 21:26, 21 July 2015 (UTC)

@Succu: I don't know who "we" is in this case. I try to comment in a way that also new users to Wikidata can follow along with the discussions. So I don't understand what you mean with your question. --Tobias1984 (talk) 21:38, 21 July 2015 (UTC)

@Succu: Not so sure about the download form. If the URLs are exposed to the internet, then a search engine could already pick up an outdated ID for a taxon. Let me get back to you when I've spoken to John Alroy about this. --Tobias1984 (talk) 21:46, 21 July 2015 (UTC)

Fossilworks redirects taxon names regarded as synonyms to the taxon name which is accepted by them e.g.:

• Accepted name Terrapene carolina (53094), synonyms:
• 90482 is Terrapene innoxia
• 96386 is Terrapene eurypygia
• 100159 is Cistudo eurypygia
• ...

The dataset should be corrected. --Succu (talk) 15:21, 22 July 2015 (UTC)

PS: You can get a list via paleobiodb api: https://paleobiodb.org/data1.1/taxa/list.txt?rel=all_taxa&limit=300000. (limit=all did not work) At the moment there are around 250,000 taxa not more than 325,000 as claimed here. --Succu (talk) 16:36, 22 July 2015 (UTC)

@Tobias1984:: Any progress? The first wrong matches arrived at wikidata. --Succu (talk) 16:20, 24 July 2015 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @Succu, Brya: I've now spoken to John, and the IDs and synonym IDs are now exposed to the user, therefore also making it much more clearer that synonyms are redirected to the valid name. We now have two choices: (1) Only give each taxon the valid ID. (2) Collect multiple IDs and give them qualifiers for the synonyms. Both options have pros and cons in my opinion. Once we have a decision I will send a new dataset to Magnus. --Tobias1984 (talk) 11:00, 25 July 2015 (UTC)

WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. --Tobias1984 (talk) 11:00, 25 July 2015 (UTC)

I don't understand the question. Could you expand a little? - Brya (talk) 16:43, 25 July 2015 (UTC)

A fossilworls ID should only be added to item which has the corresponding taxon name (P225) here not to the item with an accepted taxon name. That is

• 53094 goes to common box turtle (Q1044378) (accepted name by fossilworks)
• 90482 goes to to an item with Terrapene innoxia (synonym)
• 96386 goes to to an item with Terrapene eurypygia (synonym)
• 100159 goes to to an item with Cistudo eurypygia (synonym)

and so on. We can use taxon synonym (P1420) to add the synonyms to common box turtle (Q1044378) with a reference to fossilworls. This is fits to our data model and keeps the relation Fossilworks taxon ID (P842) ↔ taxon name (P225) unique. --Succu (talk) 06:26, 26 July 2015 (UTC)

It would, except that it is unusual to make links to redirects? - Brya (talk) 06:51, 26 July 2015 (UTC)

Yes this could be confusing. So I think we should start with a dataset that contains only accepted taxon names. --Succu (talk) 07:56, 28 July 2015 (UTC)

That looks like a good idea. - Brya (talk) 17:30, 28 July 2015 (UTC)

Tobias? --Succu (talk) 19:43, 28 July 2015 (UTC)

@Succu: I agree with the both of the above suggestions. For Mix'n'Match too, it is probably better to only upload the list of valid names (excluding synonyms), so we have a little more control over what goes on. John will mail me an up-to-date list any time now. --Tobias1984 (talk) 19:51, 28 July 2015 (UTC)

Corrected dataset is now in Mix'n'Match. --Tobias1984 (talk) 20:30, 31 July 2015 (UTC)