Wikidata:Property proposal/Atlas of Living Australia ID
Jump to navigation
Jump to search
Atlas of Living Australia ID
[edit]Originally proposed at Wikidata:Property proposal/Authority control
Motivation
[edit]A very important database for Australian taxa. Interestingly, for plant taxa it instead pipes APNI ID (P5984) a URL constructor, so this property will not apply to plants. Plants can be found using APNI ID (P5984) and a third-party formatter URL (P3303) which I have added to that property. But this proposal should apply to any non-plant taxa. -- 99of9 (talk) 07:56, 3 July 2019 (UTC)
- Following feedback below, I have dug into the variety of different identifier formats they use (see the large file download at https://bie.ala.org.au/search?q=&fq=idxtype:"TAXON"). So I'm now proposing to generalise this property to accept all of their taxon identifiers. --99of9 (talk) 03:01, 5 July 2019 (UTC)
Notified participants of WikiProject Australia
WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.
Discussion
[edit]
Comment This will be extremely useful! Ambrosia10 (talk) 09:16, 3 July 2019 (UTC) In light of what has been explained below I regret to say I am withdrawing my unconditional support. This seems a shame, as I find Atlas of Living Australia extremely helpful and would very much like to add their external identifiers to taxon items so that others might benefit from the links. If only their identifiers were persistent. Ambrosia10 (talk) 10:57, 3 July 2019 (UTC)- Comment My concern is that these identifiers are not persistent, the Atlas of Living Australia can change them at anytime. If you look at the web page for Example 2 you can see a series of identifiers labelled “Previous AFD LSID”. Each of these would have been the ALA identifier at some point in the past (I encountered this problem of identifiers changing while building Ozymandias: a biodiversity knowledge graph (Q63687022)). I don’t know what Wikidata’s policy is on identifiers that can change, so maybe this isn’t an issue, but it will complicate the lives of anyone relying on these identifiers being persistent. --Rdmpage (talk) 10:30, 3 July 2019 (UTC)
- Wow, I hadn't spotted that, thanks. I'll investigate whether using the alternative format: https://bie.ala.org.au/species/urn:lsid:biodiversity.org.au:afd.name:320977 works better. That appears not to shift when the URNs change. The number 320977 seems to come from the Australian Faunal Directory, but I can't see where they are listed, so I'd like to better understand. Any clues would be gratefully accepted. --99of9 (talk) 12:09, 3 July 2019 (UTC)
- But the afd.name ids are for taxonomic names NOT taxa so they aren’t the same thing (although they are often conflated). Wikidata itself conflates the two, but I suspect it may be too late to undo this. I presume the reason ALA identifiers change is that some researchers argue that if our notion of a taxon changes then that taxon “concept” should have a new identifier. There is logic to that approach, but it makes linking to taxa harder. Note that I’m not arguing against having ALA identifiers, just trying to clarify their properties. I don’t know whether there are other examples of Wikidata identifiers that change overtime, and whether there are mechanisms to track those changes. --Rdmpage (talk) 12:41, 3 July 2019 (UTC)
- My understanding of this URL format is "get the ALA taxon associated with the afd.name with the index 320977", which ends up with a taxon. Certainly if you click the link in my previous comment, there is no clue that the page you get to is not about a taxon. --99of9 (talk) 12:46, 3 July 2019 (UTC)
- Regarding identifiers that change with time. An example where it works ok is ComLaw ID (P2461) which changes every time an act of parliament is amended by a later act. So for example Australian Security Intelligence Organisation Act 1979 (Q4824660) has a long list of values, with start and end dates. However, this works best when URLs with old IDs still give useful info (e.g. here they link to the pre-amendment act). For ALA, I think I'd rather not maintain a list of unpredictably broken links. --99of9 (talk) 12:52, 3 July 2019 (UTC)
Support David (talk) 06:41, 4 July 2019 (UTC)- Comment @Rdmpage, Ambrosia10, ديفيد عادل وهبة خليل 2: Thanks very much for your feedback. I've dug through this and made a major overhaul of the proposal. It turns out they have quite a few identifier formats for their taxon pages, because their data comes from a range of sources. The only thing in common is that they follow the URL formatter: https://bie.ala.org.au/species/$1. So I'm now proposing that we treat everything after that as the ID. This has the advantage that it will allow this property to be used on plants and NZ species, etc. So although less clean as an ID, it is more powerful and provides the stability associated with accepting the afd.name format discussed above. What do you think? --99of9 (talk) 03:01, 5 July 2019 (UTC)
- Support Thanks for all the work you are doing on this proposal 99of9. I'm going to let Rdmpage be my guide on this. I was wanting to ensure that any work I did adding this identifier wasn't going to be made redundant as a result of broken links. I'm happy to support now that you have a solution to this worry. However it may be that Rdmpage has more technical concerns. I'll read with interest what he has to say as I would very much like to add this identifier to my species wikidata workflow.--Ambrosia10 (talk) 03:33, 5 July 2019 (UTC)
- Support I’m still a bit concerned about the stability of the identifiers, but I lack data on how many identifiers change and how often (other than my anecdotal experience that sometimes they do). Given how useful this identifier will be to people interested in the Australian biota I’m happy to support it. --Rdmpage (talk) 07:54, 5 July 2019 (UTC)
- Support revised proposal, per Rdmpage. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:19, 8 July 2019 (UTC)
- Support --Lymantria (talk) 09:32, 12 July 2019 (UTC)
- Comment I'm a little surprised this is proposed as eventually complete under Expected Completeness, that would suggest that there will never be a new species described, identified, found or even accidentally imported to Australia. Surely all species related identifiers, not just this one should be defined as Always Incomplete ? CanadianCodhead (talk) 20:48, 18 July 2019 (UTC)
- Fair enough, I've switched it. Maybe it depends how long you're willing to wait for "eventually" :) --99of9 (talk) 23:27, 18 July 2019 (UTC)
@ديفيد عادل وهبة خليل 2, CanadianCodhead, Ambrosia10, Lymantria, Pigsonthewing, 99of9: @Rdmpage: 
Done: Atlas of Living Australia ID (P7051). − Pintoch (talk) 15:12, 23 July 2019 (UTC)