Template talk:Taxonbar/Archive 3
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 | → | Archive 8 |
Plants of the World Online
Plants of the World Online is already a useful database for plants; it's more up-to-date than The Plant List, which is getting a bit out-of-date now. At present, it appears to use the IPNI identifier, so Wikidata isn't happy to add this database to Wikidata items – they are interested in identifiers not databases.
I value {{Taxonbar}} because of the links it provides to taxonomic databases. I see no reason why we here should be interested in identifiers purely as identifiers rather than as link providers. So, regretfully, because I'd rather it be done at Wikidata, I've added |powo=
(the abbreviation they use) to {{Taxonbar}}. Where there is an entry, you use the same identifier as IPNI. See Nanorrhinum scoparium for an example. Peter coxhead (talk) 09:56, 3 February 2018 (UTC)
- I gather from your comment that there are those on Wikidata who have rejected adding a statement for POWO. That is unfortunate as Wikidata is the place where such information should be stored. Its a strange interpretation of Wikidata's purpose that is restricting what data it contains. If POWO is using the IPNI identifier, an option might be to use |powo= as a flag. If there is no number, the IPNI number could be used automatically in the taxonbar. Jts1882 | talk 15:40, 3 February 2018 (UTC)
- See Wikidata:Wikidata:Property proposal/Plants of the World online; it doesn't look as though it will be added. It would be useful if an empty
|powo=
could pick up the IPNI identifier – is this possible? Peter coxhead (talk) 15:57, 3 February 2018 (UTC)- Yes, I had found that discussion and have just added my support. Jts1882 | talk 16:27, 3 February 2018 (UTC)
- See Wikidata:Wikidata:Property proposal/Plants of the World online; it doesn't look as though it will be added. It would be useful if an empty
- Taxonbar can be made to display a link to the PotW site, using values from Wikidata property P961. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:37, 3 February 2018 (UTC)
- @Pigsonthewing: Andy, yes, but the Wikidata item needs to show that the POWO (the abbreviation they use) entry exists. It can't just be added automatically by {{Taxonbar}}, since not all IPNI/P961 values are present in POWO. So, as far as I can see, every language wiki that wanted to use something like {{taxonbar}} would need an editor manually to check and add the POWO flag. This is what Wikidata is meant to avoid. How Wikidata items record the existence of a POWO entry is not relevant to us here; that they do record it, is. What it needs is something equivalent to "POWO entry = true" in the Wikidata item. Peter coxhead (talk) 16:46, 3 February 2018 (UTC)
- See below. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:39, 3 February 2018 (UTC)
- As Peter Coxhead says, how does Taxonbar tell if there is a corresponding POWO item? As you pointed out, POWO uses a subset of IPNI entries, which means most IPNI IDs don't have a corresponding POWO entry. If Taxonbar uses P961, many Taxonbars would point to non-existent POWO items. A Wikidata identifier for POWO woould provide two pieces of information, the fact that POWO contains a corresponding item and the ID to address it. Property P961 only provides the latter. Jts1882 | talk 16:51, 3 February 2018 (UTC)
- I've already addressed that question - and demonstrated a solution that works with existing properties - on Wikidata. It's really not helpful having the discussion split over multiple venues like this. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:39, 3 February 2018 (UTC)
- @Pigsonthewing: Andy, I've responded to your demonstration on Wikidata, so won't here. There are two issues though: (1) what can be done at Wikidata, if anything, which I agree is best discussed there, although the rancour and lack of civility of some of the discussants doesn't make that very attractive; (2) how we deal here with what is done or not done at Wikidata, which is obviously to be discussed here. Peter coxhead (talk) 09:21, 4 February 2018 (UTC)
- I've already addressed that question - and demonstrated a solution that works with existing properties - on Wikidata. It's really not helpful having the discussion split over multiple venues like this. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:39, 3 February 2018 (UTC)
- @Pigsonthewing: Andy, yes, but the Wikidata item needs to show that the POWO (the abbreviation they use) entry exists. It can't just be added automatically by {{Taxonbar}}, since not all IPNI/P961 values are present in POWO. So, as far as I can see, every language wiki that wanted to use something like {{taxonbar}} would need an editor manually to check and add the POWO flag. This is what Wikidata is meant to avoid. How Wikidata items record the existence of a POWO entry is not relevant to us here; that they do record it, is. What it needs is something equivalent to "POWO entry = true" in the Wikidata item. Peter coxhead (talk) 16:46, 3 February 2018 (UTC)
Is there a list of all or most available POTWO IDs? I could go through and query IPNI for each of them to see if there are any mismatches. ~ Tom.Reding (talk ⋅dgaf) 08:19, 5 February 2018 (UTC)
- Tom, the position appears to be that all the entries in POWO that anyone has looked at use IPNI identifiers. So you don't need to do this, I think. Andy demonstrated a way of marking the Wikidata entry to show that the IPNI identifier also worked in POWO, which we could have used in taxonbar, but his edit was immediately reverted. Peter coxhead (talk) 12:56, 5 February 2018 (UTC)
- I'm sure that both ways will work, but, out of principal, and so as not to set an undesirable precedent, I prefer the use of a separate property. All of the references I see under Q157343#Identifiers are related real-world entities of their parent, and all of them share a name or expand the acronym of their parent. Putting POTWO under 'IPNI plant ID' breaks this logic and structure, for no good reason. A good/less-bad reason I see would be if POTWO was a static DB, i.e. they lose funding and stop growing, or they are subsumed by the IPNI, for example, and all POTWO links become, and will forever be, a subset of IPNI or some other ID; then I'd be more neutral in how it's identified in WD. ~ Tom.Reding (talk ⋅dgaf) 18:05, 7 February 2018 (UTC)
- I agree with you that a separate property would be better, but it doesn't seem that it's going to happen, nor even that [User:Pigsonthewing|Andy]]'s solution is acceptable at Wikidata, so I think we are stuck with adding
|powo=
to Taxonbar. Peter coxhead (talk) 14:32, 8 February 2018 (UTC)- *fingers crossed* I have faith in the closing admins to weigh all arguments. We'll see if it's not misplaced! ~ Tom.Reding (talk ⋅dgaf) 14:52, 8 February 2018 (UTC)
- I agree with you that a separate property would be better, but it doesn't seem that it's going to happen, nor even that [User:Pigsonthewing|Andy]]'s solution is acceptable at Wikidata, so I think we are stuck with adding
- I'm sure that both ways will work, but, out of principal, and so as not to set an undesirable precedent, I prefer the use of a separate property. All of the references I see under Q157343#Identifiers are related real-world entities of their parent, and all of them share a name or expand the acronym of their parent. Putting POTWO under 'IPNI plant ID' breaks this logic and structure, for no good reason. A good/less-bad reason I see would be if POTWO was a static DB, i.e. they lose funding and stop growing, or they are subsumed by the IPNI, for example, and all POTWO links become, and will forever be, a subset of IPNI or some other ID; then I'd be more neutral in how it's identified in WD. ~ Tom.Reding (talk ⋅dgaf) 18:05, 7 February 2018 (UTC)
@Peter coxhead, Plantdrew, Tom.Reding, Jts1882, and Pigsonthewing: What do you think of worldfloraonline.org? I personally think the entries of World Flora Online contain more data than POWO, though I'm not sure about how current the taxonomic approach is. I certainly think at least one or both of these global plant databases need to be linked to but I want to hear thoughts on World Flora before I choose sides here.~ Mellis (talk) 17:00, 8 February 2018 (UTC)
- @Mellis: World Flora Online should certainly be incorporated at some point (fortunately, it uses its own IDs, so shouldn't be a problem getting a property for it at Wikidata). WFO is the successor to The Plant List which will not be updated further. WFO will include richer data than TPL (distribution, descriptions, etc.). Currently their nomenclatural data seems to be still largely based on TPL. I've been checking into WFO every few months for the last couple of years to get an idea of their progress. Back in July 2017 they were missing huge swathes of species. Their coverage is far better today, and they may have incorporated everything that was in TPL (although I really should spend some more time exploring what they have). The only question for me about including WFO in Taxonbars/Wikidata is whether it is sufficiently mature as a resource yet. I think it may be at this time. The "deadline" for WFO to be "finished" is in 2020, so it should be improving further.
- I'm not really sure what the story is behind POWO. Kew is one of 41 institutions partnering to produce the WFO. I'm guessing Kew had a lot of data ready to go, and just decide to roll it out independently as POWO, rather than waiting for WFO to be ready to accept it. I expect WFO is likely to eventually include the data in POWO. Plantdrew (talk) 21:57, 8 February 2018 (UTC)
- Yes, it's difficult to understand exactly what the relationship is between WCSP, POWO and WFO. I have reservations about both POWO and WFO at present. I've pointed out some obvious errors in POWO and never had a reply, whereas Rafaël Govaerts responds rapidly, and if appropriate updates WCSP equally rapidly (including at the weekend), but of course WCSP has limited coverage. WFO seems to have imported most of its information from TPL 1.1, but that means that it continues to have all the well known problems with records TPL got from Tropicos, in particularly incorrectly saying that some taxa are "accepted" when Tropicos says no such thing, and importing multiple synonyms from Tropicos as independent taxa. WCSP and POWO have distribution data for all taxa, WFO does not (yet). So I think that all three are currently important, and it remains to be seen what the final outcome will be. Peter coxhead (talk) 23:10, 8 February 2018 (UTC)
Another example of the value of multiple from parameters
Since there seems to be some opposition at Wikidata to the idea of multiple links to it from {{Taxonbar}}, I thought that, for the record and for use in any future discussion should there be one, I'd give another example of the value of the multiple from
parameters. It concerns Syrmatium haydonii.
- Obviously the link to the Wikidata item at this name is useful. The name is, as of now, accepted by GBIF, POWO & TPL.
- The basionym is Hosackia haydonii. Although this name doesn't seem to be in use at present, being able to look up the basionym in various taxonomic databases is useful, e.g. for links to the protologue.
- The synonym Acmispon haydonii is, as of now, accepted by ITIS, GRIN & TPL. (TPL is of course in error in accepting both this name and Syrmatium haydonii.)
- The synonym Lotus haydonii is, as of now, accepted by Tropicos.
So with no clear accepted name, all the taxonomic database links available for all four names are important in understanding and researching this taxon. Tropicos is, I think, a bit out of date in using Lotus, but either of Syrmatium and Acmispon are defensible. I chose Syrmatium based on POWO simply to get a self-consistent list of former Lotus species. Peter coxhead (talk) 17:23, 7 February 2018 (UTC)
- I take it any manually specified IDs (i.e. POWO) will be associated with the first item, and it is not possible to specify multiple manual IDs from the same source? Plantdrew (talk) 17:39, 7 February 2018 (UTC)
Yes, that's what happens now, but it's not desirable, since there can well be 'extra' taxonomic databases associated with different synonyms. For now, the 'solution' in such cases seems to be use a separate {{taxonbar}} template for each synonym. Exactly how best to specify in a single taxonbar that e.g.See below. Peter coxhead (talk) 14:36, 8 February 2018 (UTC)|powo=
applies to|from3=
not|from1=
isn't clear to me.- Use the number as in
|powo3=
? Jts1882 | talk 14:46, 8 February 2018 (UTC)Well, this seems the right approach if the Lua code at Module:Taxonbar is fixed to deal with numbers following any parameter added to Module:Taxonbar/conf. Can you fix it?Peter coxhead (talk) 15:05, 8 February 2018 (UTC)- @Peter coxhead: The code to allow multiple rows was originally written so that it recognizes numbers at the end of any parameter (this is documented at Template:Taxonbar#Multiple_Wikidata_entries). See Template:Taxonbar/testcases#Q24072911_as_#3 for an example using
zoobank
andzoobank2
. What exactly need to be fixed? --Ahecht (TALK
PAGE) 15:37, 8 February 2018 (UTC)- Nothing. Clearly I should have read the documentation. Excellent anticipation! Peter coxhead (talk) 16:04, 8 February 2018 (UTC)
- @Peter coxhead: The code to allow multiple rows was originally written so that it recognizes numbers at the end of any parameter (this is documented at Template:Taxonbar#Multiple_Wikidata_entries). See Template:Taxonbar/testcases#Q24072911_as_#3 for an example using
- Use the number as in
@Peter coxhead:Where is the 'opposition at Wikidata to the idea of multiple links to it from {{Taxonbar}}
'? A link would be very helpful, I have not seen this discussion.--~ Mellis (talk) 06:28, 9 February 2018 (UTC)
- @Mellis: sorry, I've just searched for the comment and can't find it, so it may have been withdrawn. (I've lost patience with the Wikidata discussions, which seem to be going nowhere, and haven't been following them.) Somewhere I pointed to the value of multiple synonyms being shown in a taxonbar and Brya (if no-one else) criticized this approach. Peter coxhead (talk) 07:12, 9 February 2018 (UTC)
Article correction needed
The Help:Taxon identifiers article needs to be corrected. It's clear that the identifiers are not, in almost all cases, "taxon" identifiers. They are "taxon name" identifiers. If we take homotypic synonyms based on different genus placements, for example, IPNI's 60448538-2, 144005-2 and 124200-2 are the same taxon but different names. If we include heterotypic synonyms, 1034844-2, 143974-2 and 124211-2 are also the same taxon. Tropicos also has six identifiers for these six names; ITIS and GBIF have at least three. Yet there is only one species, one taxon. Peter coxhead (talk) 08:07, 9 February 2018 (UTC)
- @Peter coxhead:Taxonbar includes identifiers of the various names and the taxon itself. Because of many taxonomic viewpoints, there will of course be multiple names to refer to a single taxon. It is not clear what changes you are suggesting. Of course IPNI refers to taxon names. Are you also arguing ITIS and GBIF are databases of names rather than taxa? Please do not alter the article to refer to all entries as taxon names, I feel concensus would be needed in that case. ~ Mellis (talk) 21:11, 9 February 2018 (UTC)
- Those databases that present an "accepted" name for a taxon, like ITIS and GBIF, nevertheless contain identifiers for synonyms. Thus GBIF, for example, has 5350114 for Lotus procumbens, 2947069 for Syrmatium procumbens, 2947068 for Hosackia procumbens and 2947064 for what it regards as the accepted name for the taxon, Syrmatium sericeum. So to say that these are "taxon identifiers" is clearly wrong; they are "taxon name" identifiers. A "taxon identifier" would uniquely identify a taxon; these don't. It's simply wrong to say that most of the identifiers are "taxon identifiers" and the help page should be accurate. Peter coxhead (talk) 08:17, 10 February 2018 (UTC)
- While the taxon identifiers clearly don't identify a unique taxon, which should be their logical function, they don't purely identify a taxon name. The entries also include the relationships to other names and the actual taxon being dealt with. I've taken to thinking of them as taxonomic opinions, which isn't totally satisfactory as different people have different taxonomic opinions using the same name and don't get separate entries. On the other hand, perhaps we should ask what a taxon is? Is it the group of organisms themselves or a taxonomic viewpoint on the organisms. Given species boundaries are not clearly defined and that taxonomy is in flux, one could argue the latter. Jts1882 | talk 08:34, 10 February 2018 (UTC)
- I can't see that there's a real difference between "the group of organisms" and "a taxonomic viewpoint on the organisms". In the ICN, it's clear that a taxon (= taxonomic group) is a group of organisms treated as a unit by taxonomists (see e.g. the Preamble). Taxonomic groups don't exist in nature; they are always human constructs. So taxon names always represent opinions – although not complete opinions, because so long as two groups include only one name-bearing type (to use ICZN terminology) they will have the same name but can have different circumscriptions. (The only way to distinguish these is to add "sensu X", "sensu Y", etc.) I can only repeat that in practice almost all the identifiers in a taxonbar identify names; yes, the entries have additional information attached, but it's attached via the name.
- On a slight side issue, what is unquestionable is that in some databases, particularly the all important IPNI, there are multiple identifiers per name, although the Wikidata item allows only one to be attached. I've changed Help:Taxon identifiers to make this clear, since it's simply a matter of fact. Peter coxhead (talk) 11:03, 10 February 2018 (UTC)
- While the taxon identifiers clearly don't identify a unique taxon, which should be their logical function, they don't purely identify a taxon name. The entries also include the relationships to other names and the actual taxon being dealt with. I've taken to thinking of them as taxonomic opinions, which isn't totally satisfactory as different people have different taxonomic opinions using the same name and don't get separate entries. On the other hand, perhaps we should ask what a taxon is? Is it the group of organisms themselves or a taxonomic viewpoint on the organisms. Given species boundaries are not clearly defined and that taxonomy is in flux, one could argue the latter. Jts1882 | talk 08:34, 10 February 2018 (UTC)
- Those databases that present an "accepted" name for a taxon, like ITIS and GBIF, nevertheless contain identifiers for synonyms. Thus GBIF, for example, has 5350114 for Lotus procumbens, 2947069 for Syrmatium procumbens, 2947068 for Hosackia procumbens and 2947064 for what it regards as the accepted name for the taxon, Syrmatium sericeum. So to say that these are "taxon identifiers" is clearly wrong; they are "taxon name" identifiers. A "taxon identifier" would uniquely identify a taxon; these don't. It's simply wrong to say that most of the identifiers are "taxon identifiers" and the help page should be accurate. Peter coxhead (talk) 08:17, 10 February 2018 (UTC)
Italicization of taxon names
I've just implemented a change to ensure that in a taxonbar like
the connecting term "var." is not italicized – it looks very wrong to me when it is, as it would have been before the update. This was an easy fix to make as it re-used code already written for other purposes, e.g. {{Species list}}.
However in a taxon bar like
the family name is italicized, because the taxonbar code doesn't look at the rank of the taxon. It appears possible to avoid italicizing ranks above genus, although somewhat tedious to implement. I'm wondering whether this is worthwhile. Comments please. Peter coxhead (talk) 11:10, 13 February 2018 (UTC)
- It doesn't look as grating as the italised var.. You do occassionally see families italised for emphasis in higher taxonomy listings. I'd fix it if there is a clear and simple method, but wouldn't consider it high priority. Jts1882 | talk 13:22, 13 February 2018 (UTC)
- I would certainly like to see "var." unitalicized. The first time I saw this on a page made it seem to me that it was part of the name...
- @Tom.Reding: this is now implemented; all the botanical connecting terms should now be formatted correctly. Peter coxhead (talk) 22:53, 18 February 2018 (UTC)
- I think Ahecht made a kludge for a rank-lookup function. Did that work? ~ Tom.Reding (talk ⋅dgaf) 22:10, 18 February 2018 (UTC)
- Unfortunately, it's not so simple – I explained at Ahecht's talk page, but I guess it should be repeated here as this is the place to discuss taxonbar issues.
- If all the names were botanical, all that's needed is a list of ranks from genus downwards; any rank not in this list isn't italicized.
- However, it's not so simple when the different codes are taken into account. For example:
- It's necessary to distinguish botanical and zoological uses of the rank of section. The automated taxobox system does this by using two different 'ranks', sectio and zoosectio. This is because botanical sections are below genus rank and so should be italicized (other than "sect.") and zoological sections are above genus rank, so should not be italicized. As far as I can see, Wikidata doesn't distinguish between the two kinds of section, so this information would have to be acquired in some other way.
- The bacterial code has some differences, e.g. it uses the prefix "Candidatus" (unitalicized). However, Wikidata doesn't seem to deal with this – see Candidatus Carsonella ruddii and Carsonella ruddii (Q141647).
- The virus code has quite different rules for the use of italics, which are supposed to be followed in the English Wikipedia (see Wikipedia:WikiProject Viruses/Guidelines#Italics and capitalization) although in practice they aren't always. All ranks, except the top level Groups, are italicized for viruses. So it's necessary to detect whether the taxon name is of a virus or not.
- At present I favour leaving it at the fixes I've made – automating it completely would be a major task (in the almost eight year lifetime of automated taxoboxes, there's never been one that worked properly for viruses). I'm pretty sure that in special cases there would have to be some manual input. I doubt that it's really worthwhile, but if anyone wants to try, do feel free! Peter coxhead (talk) 22:53, 18 February 2018 (UTC)
- Unfortunately, it's not so simple – I explained at Ahecht's talk page, but I guess it should be repeated here as this is the place to discuss taxonbar issues.
Links to Wikimedia Commons and Wikispecies
This is a question I've asked before elsewhere, and needs wider discussion, but I'm interested to know what people here think. An article that is linked to the relevant Wikidata item now has links to Wikimedia Commons and Wikispecies automatically placed in the left margin. So do we need the two templates added to the article as well? Peter coxhead (talk) 10:12, 18 February 2018 (UTC)
- I use the left margin much more frequently from userspace than from mainspace. I often ignore it in mainspace unless I'm trying to find the 'Wikidata item' link, which I usually resort to CTRL+F for anyway, so it goes largely unnoticed. It has a lot of (to me) unnecessary links that I wish I could toggle off. I've often look for, and sometimes follow, the commons-box, so I do think it's useful. Also, the left margin is static, so having a 2nd commons link at the bottom of long articles is convenient. I assume this applies to frequent users of Wikispecies as well. ~ Tom.Reding (talk ⋅dgaf) 19:47, 19 February 2018 (UTC)
What’s up at Darapsa choerilus?
The template has been added to Darapsa choerilus but is generating an error. Nothing obviously wrong with the article or Wikidata (though I would not know what to look for there).--JohnBlackburnewordsdeeds 20:46, 21 February 2018 (UTC)
- The Butterflies and Moths of North America ID was written in Wikidata with
%28
and%29
instead of(
and)
(which Wikidata "helpfully" further URL encoded into%2528
and%2529
). Replacing those with parentheses at Wikidata seems to have fixed the error. --Ahecht (TALK
PAGE) 21:06, 21 February 2018 (UTC)- Oops, forgot to ping JohnBlackburne. --Ahecht (TALK
PAGE) 21:07, 21 February 2018 (UTC)
- Oops, forgot to ping JohnBlackburne. --Ahecht (TALK
Wikidata property for multiple from#s?
It seems to me that it would be useful, if it doesn't already exist somewhere, to have a property that lists all related QIDs. Ideally, wouldn't it be useful to have a property that we can enumerate via module that contains all of the associated QIDs, instead of listing them as multiple |from#=
s? Keep adding multiple |from#=
s whenever necessary, by all means. If something like this has support, we can migrate them to WD in the future. ~ Tom.Reding (talk ⋅dgaf) 22:17, 18 February 2018 (UTC)
This functionality can ultimately be used not only by {{Taxonbar}}, but also by {{Taxobox}} & {{Speciesbox}}. ~ Tom.Reding (talk ⋅dgaf) 22:21, 18 February 2018 (UTC)
- This is a problem for Wikidata to address (some of us have tried, unsuccessfully, to discuss it there). The underlying problem is that the data modelling is incomplete (or even incorrect), as I have explained on too many occasions now. Consider Retama raetam (Q2910823). The Wikidata item says that Q2910823 is an "instance of taxon". If you look at taxon (Q16521), it says that a taxon is a "group of one or more organism(s), which a taxonomist adjudges to be a unit". Actually, Retama raetam (Q2910823) is an instance of a species name; the group of organisms adjudged to be a unit, here a species, has at least three names, Retama raetam (Q2910823), Genista raetam (Q39157123) and Lygos raetam (Q15531393). These are not three taxa – the taxonomists who place(d) them in different genera are not judging them to be different groups of organisms, but rather judging the same group of organisms to belong to different genera. One way of representing the data would be to have items which are really instances of taxa; items that are instances of taxon names can then be declared to be a "name of" items that are instances of taxa. (A more detailed data analysis would distinguish homotypic and heterotypic synonyms.) Peter coxhead (talk) 00:18, 22 February 2018 (UTC)
- Peter coxhead (and anyone else), can you link to those previous discussions? ~ Tom.Reding (talk ⋅dgaf) 00:26, 22 February 2018 (UTC)
Taxonbar desirability brought up at ANI
See here. ~ Tom.Reding (talk ⋅dgaf) 15:43, 26 February 2018 (UTC)
Default QID not showing up
See this version of Ancistrocheirus, where the wikibase item = Q18519508, but only |from=Q2015219
was specified (someone changed WD after the from was placed, of course). Shouldn't the 'master' wikibase item (whichever is linked to the page) display on the top line, then all the other valid |from#=
s below it? ~ Tom.Reding (talk ⋅dgaf) 14:15, 27 February 2018 (UTC)
- @Tom.Reding: The lua code can only generate the number of rows that it has data for. If only from/from1 is specified, the code will never start a second row. --Ahecht (TALK
PAGE) 14:51, 27 February 2018 (UTC)- @Ahecht: - but {{Taxonbar}}'s able to automatically generate a row without any
|froms=
. I thought that behavior was carried over? ~ Tom.Reding (talk ⋅dgaf) 15:01, 27 February 2018 (UTC) - I thought that's what was desired - i.e.
|from=
doesn't "lock-in" the displayed taxon. ~ Tom.Reding (talk ⋅dgaf) 15:05, 27 February 2018 (UTC)- The code will fill in from/from1 from the current page, but not subsequent rows (otherwise you would end up with multiple rows being automatically filled with Wikidata, and Wikidata lookups are resource-expensive functions). If you want a manually input value to be persistant despite page moves, it would need to be put in from2 or higher. --Ahecht (TALK
PAGE) 15:11, 27 February 2018 (UTC)- Ok, so it's a good thing we have the tracking categories to take care these cases, and to do so more efficiently than 'double-pulling' from WD (as would be the case 99% of the time). ~ Tom.Reding (talk ⋅dgaf) 15:25, 27 February 2018 (UTC)
- The code will fill in from/from1 from the current page, but not subsequent rows (otherwise you would end up with multiple rows being automatically filled with Wikidata, and Wikidata lookups are resource-expensive functions). If you want a manually input value to be persistant despite page moves, it would need to be put in from2 or higher. --Ahecht (TALK
- @Ahecht: - but {{Taxonbar}}'s able to automatically generate a row without any
Specifying the from/from1 parameter
I spend a good deal of time monitoring the relevant error-tracking categories and fixing broken taxoboxes (as do some other editors, like Plantdrew). One significant cause of breakage is that the taxobox code was designed to pick up the taxon name from the article title if it wasn't explicitly given as a parameter. This makes life a little easier for editors, since a very minimal taxobox often works. However, there are two consequences I see regularly: novice editors don't understand why taxoboxes don't work if the article is at the English name; and fixes are needed if an article is moved to another title (perhaps from a scientific name to an English name or to yet another new scientific name), but many editors don't understand how to make these fixes. The same problems will arise with taxonbars if editors are encouraged to omit |from=
, so personally I would not encourage it. (Who's going to monitor and fix broken taxonbars?) Peter coxhead (talk) 21:25, 5 January 2018 (UTC)
- Taxonbar will not consider the article title or pull any data from it, it only pulls from the Wikidata page associated with the article, or from the
|from=
parameters. This causes problems and inconsistencies if the article is linked with a Wikidata entry that does not correspond to the name within Taxobox. However these are the limitations we have to work with. ~ Mellis (talk) 00:40, 6 January 2018 (UTC)- Ok, so consider the following. There's an article in the English Wikipedia at "X a" and a link to it at Wikidata item QN. A synonym of X a has a Wikidata item at QM. So the taxonbar template doesn't specify
|from=
, picking up QN from the article link, but does specify|from2=QM
. Now someone moves the Wikidata English article link from QN to QM, as the name there has become more clearly the accepted name. What will our taxonbar display? The only way of preventing such problems is for allfrom
parameters to be specified. Peter coxhead (talk) 11:30, 6 January 2018 (UTC)
- Ok, so consider the following. There's an article in the English Wikipedia at "X a" and a link to it at Wikidata item QN. A synonym of X a has a Wikidata item at QM. So the taxonbar template doesn't specify
- Maybe they should be and it should be encouraged, but there is a legacy where most taxonbars don't specify as they were intended to be automatic. In your example, the taxonbar would show QM twice after the change. The module should be able to check and avoid such duplication and then it would automatically pick up the change from QN to QM at Wikidata. If from is specified then the person changing the wikidata pointer to the English wikipedia would also have to change the taxonbar. Jts1882 | talk 13:52, 6 January 2018 (UTC)
- Peter coxhead, good example & makes sense. I can help add
|from1=
s as well (when the time comes). - Jts1882, (if I understand correctly) how would the module
automatically pick up the change from QN to QM at Wikidata
- is there a way for it to tell the difference b/w a moved page, and an editor who "duplicated"/hardcoded an article's WD ID (as we might be about to do)? The module could, however, automatically put the|from#=
which matches the page's WD ID first on the list of taxons, or is that too much automation/not enough user control?- My understanding is that the module uses Wikibase to get the Wikidata item. So when someone changes the English language Wikipedia item in Wikidata, this gets picked up by the taxonbar module if there is no "from" parameter. This is a desirable effect, which would be lost if from is made mandatory. Jts1882 | talk 14:27, 6 January 2018 (UTC)
- The desired state & behavior is:
- for all synonyms be listed in the bar,
- that the top-most item uses the page's WD ID,
- be resilient to WP page moves that are among available synonyms, and
- be resilient to WD moves that are among available synonyms,
- did I miss anything?
- So as long as the module checks each
|from1=
,|from2=
, etc., looking for one that matches the page's WD ID, and uses that as the top-most taxon (or, to the same effect, the module can force the display of the page's WD ID at the top, then display any other|from#=
s which don't match the page's WD ID below it; which is just more of a coding preference), that should solve both your & Peter coxhead's examples? ~ Tom.Reding (talk ⋅dgaf) 17:28, 6 January 2018 (UTC)
- The desired state & behavior is:
- My understanding is that the module uses Wikibase to get the Wikidata item. So when someone changes the English language Wikipedia item in Wikidata, this gets picked up by the taxonbar module if there is no "from" parameter. This is a desirable effect, which would be lost if from is made mandatory. Jts1882 | talk 14:27, 6 January 2018 (UTC)
- Another question is, should the module
- Force/require a
|from1=
(either by not displaying the taxonbar at all, or displaying a message), and/or - Put those pages without a
|from1=
into a maintenance cat, to make sure nothing falls through the cracks.
- Force/require a
- I like #2 at the very least. ~ Tom.Reding (talk ⋅dgaf) 14:21, 6 January 2018 (UTC)
- @Peter coxhead, Tom.Reding, and Jts1882:
- Taxonbar uses Wikibase to retrieve the associated Wikidata item, this is retrieved with each purge of the page, so it is regularly updated.
- Not specifying a
|from=
or|from1=
in most cases is desirable, as specifying the parameter locks Taxonbar into a specific Taxonomic view that could easy go out of date and not be checked by someone who may make changes to Taxobox, the article title, or the Wikidata entry. In some cases Taxonbar will show a more up to date taxonomic view than Taxobox. I am urging you all not to specify a|from=
unless there is a problematic discrepancy, between Taxonbar and the name used in Taxobox. - Tracking categories would be useful in instances where
|from=
and/or|from2=
are specified. An especially helpful tracking category could be established by Taxobox that actually tracked discrepancies between the associated Wikidata link via Wikibase and the specified taxon name inputted into Taxobox. - I agree with Jts1882 that Taxonbar should check to prevent any duplicate QIDs of
|from1=
,|from2=
, and|from3=
to prevent redundancy in the case of Wikibase link moving.
- ~ Mellis (talk) 15:52, 6 January 2018 (UTC)
- Well, the really important thing to do is to ensure that if any of the
fromN
parameters turn out to be the same as each other or the unspecifiedfrom
parameter, the article is put in an error-tracking category. - Re (3), I'm not sure how this would work in practice, given the variety of templates that create both manual and automated taxoboxes. Taxoboxes are designed to look similar, but they are created in quite different ways. Manual taxoboxes don't actually 'know' what the target taxon is – they just output successive lines for all the parameters specified. I think I'd want to re-code {{Taxobox/core}} in Lua before trying to work out what was the lowest rank parameter given a value in order to check it against Wikidata. For the automated taxoboxes, {{Speciesbox}} works differently from the others, since it only uses the genus to find the taxonomic hierarchy; in this case the check would have to be coded inside {{Speciesbox}} or the species epithet passed upwards. For the other automated taxoboxes, I think the check could be coded in {{Taxobox/taxonomy}} or the Lua module it calls, but I'd have to think about that. It's certainly not straightforward and I'm not sure how worthwhile it would be. Peter coxhead (talk) 16:44, 6 January 2018 (UTC)
- Yes, it would help if {{Taxobox/core}} was written in Lua. Basically what I'm proposing for #3 wouldn't run the Wikibase check based on {{Taxobox/taxonomy}}, because that only holds data down to the genus level in most cases. I'm talking about running the comparison directly from the inputs of
|genus=
,|species=
, and/or|binomial=
parameters specified in the article code, and comparing them directly to taxon name (P225) on Wikidata, and simply tracking which articles have discrepancy so that the taxonomy and Wikibase links of these pages can in theory be checked more regularly. ~ Mellis (talk) 17:25, 6 January 2018 (UTC)- If you are only proposing running the check on species, it is more doable, certainly. Manual taxoboxes could check on the value of
|binomial=
, and {{Speciesbox}} on the value of|taxon=
or|genus=
+|species=
. Peter coxhead (talk) 17:31, 6 January 2018 (UTC)Agree - As you say, in the case of {{Speciesbox}} or {{Automatic Taxobox}} the check would be on
|taxon=
, or for {{Speciesbox}} specifically,|genus=
+|species=
, or|binomial=
. ~ Mellis (talk) 17:44, 6 January 2018 (UTC)- On reflection, it also has to be done for {{Subspeciesbox}} (see Dog), and I assume for {{Infraspeciesbox}}, although I haven't yet seen any taxonbars for botanical infraspecies. Peter coxhead (talk) 17:49, 6 January 2018 (UTC)
- Things get very tricky with trinomial names, lots of different treatments depending on the 'Authority'. Things could get much messier if we track discrepancies that far down.--~ Mellis (talk) 17:58, 6 January 2018 (UTC)
- On reflection, it also has to be done for {{Subspeciesbox}} (see Dog), and I assume for {{Infraspeciesbox}}, although I haven't yet seen any taxonbars for botanical infraspecies. Peter coxhead (talk) 17:49, 6 January 2018 (UTC)
- If you are only proposing running the check on species, it is more doable, certainly. Manual taxoboxes could check on the value of
- Yes, it would help if {{Taxobox/core}} was written in Lua. Basically what I'm proposing for #3 wouldn't run the Wikibase check based on {{Taxobox/taxonomy}}, because that only holds data down to the genus level in most cases. I'm talking about running the comparison directly from the inputs of
- Well, the really important thing to do is to ensure that if any of the
- @Peter coxhead, Tom.Reding, and Jts1882:
- Peter coxhead, good example & makes sense. I can help add
But if taxonbars are added to articles with trinomial names, whether zoological or botanical, any checking system has to deal with them. At higher levels, the problem certainly exists. Thus our article Scilloideae doesn't link to the French article fr:Hyacinthaceae, although it should. A taxonbar at Scilloideae needs to pick up the entries at Hyacinthaceae (Q13833438). It's quite likely, in my view, that the specialists who prefer Hyacinthaceae to Scilloideae will win out and the article will need to be moved. All of this complexity demonstrates that the problem is at Wikidata. They need to properly link alternative names for taxa. It's not our problem. Peter coxhead (talk) 19:20, 6 January 2018 (UTC)
- @Peter coxhead: I spent many hours dreaming up possible solutions to the Scilloideae and interlanguage link problems today. Ultimately I began hitting many roadblocks involving other Wikipedias. For example, nl:Wikipedia has an article at nl:Scilloideae and nl:Hyacinthaceae, even if Wikipedia/Wikidata worked the way you wished, one of those Dutch articles would be lost and unlinked. In my view, the best solution for Scilloideae in terms of Taxonbar would be to show two bars for the two different interpretations. I do wish the interlanguage link problem could be solved, but ultimately not the fault of Taxonbar. ~ Mellis (talk) 00:37, 7 January 2018 (UTC)
- @Mellis: As you say, it's not a problem for Taxonbar. I don't see why one of the Dutch articles need be lost if Wikidata were set up properly. This relates to another major problem of Wikidata: its incorrect modelling of Wikipedia articles, which insists on 1:1 relationships, and the consequent listing of a single article in another language in the left hand margin. This is completely incorrect, not just for taxa, since topics that deserve only one article in one language/culture are often of sufficient importance in another language/culture to be worth forking into several articles. As I've pointed out many times before, Berry and Berry (botany) suffer from this problem. For the taxonbar, there's no reason why the API could not pick up taxon/name identifiers from all interlinked synonyms, and for interlanguage links, it could return more than one link to the same wikipedia. It's all a Wikidata problem: it's a database, and like all databases, if the data analysis is wrong, the database is wrong (I used to teach this stuff).
- My conclusion is that there's no point in each wikipedia introducing considerable extra complexity to multiple templates to try to fix problems with Wikidata. I think we've done as much as is sensible here. Peter coxhead (talk) 07:19, 7 January 2018 (UTC)
So...back to from1
The discussion veered off into {{Taxobox}} & {{Speciesbox}}, which is a much broader/another topic. I'd like to see something concrete/constructive/possibly actionable come out of this thread.
It sounds like objections to |from1=
are quelled as long as the primary WD QID for the page/taxonbar is displayed on the taxonbar - i.e. it's displayed whether or not the right QID exists in any of the |from#=
parameters, and not duplicated on the taxonbar, and tracking-categorized to have someone add the QID to the |from#=
list IF it doesn't already exist in the |from#=
list (to be robust to page moves Wikidata entity/QID changes) (I can add the missing |from=
s, whether or not tracking cat exists). Is that pretty much correct? ~ Tom.Reding (talk ⋅dgaf) 12:45, 14 January 2018 (UTC)
- Looks like the module is behaving as intended regarding this (except the usage of a tracking cat; might be a good time for me to learn lua...). Will start adding
|from1=
s next week if no further issues. ~ Tom.Reding (talk ⋅dgaf) 13:12, 19 January 2018 (UTC)- @Tom.Reding: I mentioned this earlier:
Not specifying a |from= or |from1= in most cases is desirable, as specifying the parameter locks Taxonbar into a specific Taxonomic view that could easy go out of date and not be checked by someone who may make changes to Taxobox, the article title, or the Wikidata entry. In some cases Taxonbar will show a more up to date taxonomic view than Taxobox. I am urging you all not to specify a |from= unless there is a problematic discrepancy, between Taxonbar and the name used in Taxobox.
- So let me know if you have an example article in which you would explicitly need to specify |from1=, other than something like Watermelon, in which the article describes a natural product of the taxon. ~ Mellis (talk) 20:06, 19 January 2018 (UTC)
- @Mellis: There are 2 sources of change: WP & WD. If someone changes the {{Taxobox}} name, and doesn't check WD nor the {{Taxonbar}} for discrepancies, then nothing is gained nor lost by having
|from1=
- the template appears just as it would if|from1=
wasn't filled out. If someone moves the WP page, nothing is gained nor lost by having|from1=
, since the WD QID follows it. However, if someone changes the WD QID pointing to the page, we have no way of knowing/checking that on any meaningful scale, aside from a fleeting watchlist entry, or someone happening onto that particular WD item's history. What|from1=
does is provide a reference point for the QID 'history' of the WP page that is easily accessible on the WP page. This can be easily checked by bot, or possibly by the module itself, to look for pages which have had QIDs changed. Is that not desirable for some reason? ~ Tom.Reding (talk ⋅dgaf) 15:19, 24 January 2018 (UTC)- Just to note that I think Tom has put the argument very clearly, and I agree entirely. Peter coxhead (talk) 11:24, 25 January 2018 (UTC)
- @Tom.Reding and Peter coxhead: It would be desirable to have a tracking category or a bot preform some kind of agreed-upon monitoring. I have limited time to invest and little skill in Lua coding and no experience working with bots. Where I have hesitation is creating policy that requires specifying
|from1=
and suddenly having all kinds of mistakes made by many less experienced editors. Inconsistencies between {{Taxobox}}, {{Taxonbar}}, Wikidata, and Wikipedia titles getting out of hand is a real concern, as noted many times in this discussion. ~ Mellis (talk) 18:35, 25 January 2018 (UTC)- I agree, and I see
23 possibly useful tracking categories (names and which one(s) to use are up for debate):- Category:Taxonbar templates missing from parameter or Category:Taxonbar templates missing from1 parameter
- Category:Taxonbar templates desynced from Wikidata
- Category:Taxonbar templates using multiple Wikidata items
- Category:Taxonbar templates with invalid from parameters (added ~ Tom.Reding (talk ⋅dgaf) 15:04, 1 February 2018 (UTC))
- Ahecht, would you be able to add tracking category functionality to the module, when we get a better idea of what we want? ~ Tom.Reding (talk ⋅dgaf) 18:50, 25 January 2018 (UTC)
- I don't see any technical reason why those tracking categories couldn't be added. --Ahecht (TALK
PAGE) 19:36, 25 January 2018 (UTC)- Litter of cats created. Ahecht, you're up! I think I was specific enough in their description to unambiguously write the associated code, but let me know if not. ~ Tom.Reding (talk ⋅dgaf) 19:10, 31 January 2018 (UTC)
- @Tom.Reding: The template is currently set up so that any row that matches the current page's wikidata item is moved to the top. With this in mind, if from2, from3, etc. matches the current page, do you still want it flagged with the first two categories? --Ahecht (TALK
PAGE) 19:35, 31 January 2018 (UTC)- Short answer: as long as one of the froms match it's good.
- Longer answer: Hmm, it depends how frequent this sort of WD reshuffling is/could be, and how strict we want to be. I think, for now, as long as one of the
|from=
s match, it's ok. Maybe down the road we'll want to get this sort of strictness in place, but I think starting with the general case for now is best. Incidentally, moving the n'th-from to|from1=
may not change the layout of the page (it might if there are 3 or more from's, but definitely not if there are only 2 froms), so those sorts of fixes would fall either at or below the level of WP:COSMETICBOT (unless WT:TREE decides otherwise, but since this isn't a real issue I don't see that we would). ~ Tom.Reding (talk ⋅dgaf) 19:52, 31 January 2018 (UTC) - Category text updated to reflect this. ~ Tom.Reding (talk ⋅dgaf) 20:23, 31 January 2018 (UTC)
- @Tom.Reding: The template is currently set up so that any row that matches the current page's wikidata item is moved to the top. With this in mind, if from2, from3, etc. matches the current page, do you still want it flagged with the first two categories? --Ahecht (TALK
- Litter of cats created. Ahecht, you're up! I think I was specific enough in their description to unambiguously write the associated code, but let me know if not. ~ Tom.Reding (talk ⋅dgaf) 19:10, 31 January 2018 (UTC)
- I don't see any technical reason why those tracking categories couldn't be added. --Ahecht (TALK
- I agree, and I see
- @Tom.Reding and Peter coxhead: It would be desirable to have a tracking category or a bot preform some kind of agreed-upon monitoring. I have limited time to invest and little skill in Lua coding and no experience working with bots. Where I have hesitation is creating policy that requires specifying
- Just to note that I think Tom has put the argument very clearly, and I agree entirely. Peter coxhead (talk) 11:24, 25 January 2018 (UTC)
- @Mellis: There are 2 sources of change: WP & WD. If someone changes the {{Taxobox}} name, and doesn't check WD nor the {{Taxonbar}} for discrepancies, then nothing is gained nor lost by having
@Tom.Reding: I updated the sandbox version. It's coded so that it will display the categories as text (but not categorize the page) when |demo=yes
or when used on a testcases page, so try it out and let me know if the logic works as you would expect. --Ahecht (TALK
PAGE) 21:06, 31 January 2018 (UTC)
- @Ahecht: it looks like the "desynced" cat gets applied to pages without any from's, but it should only be applied when 1 or more (non-null) from's are passed and when none of them equal the current WD QID.
The "desynced" cat should, but doesn't, also get applied when an invalid (non existent) QID is passed viaThe red error msg ("[...] is unknown to the system. Please use a valid entity ID.") does show up on the page, though, so that part works, and might help with finding the relevant spot in the code. ~ Tom.Reding (talk ⋅dgaf) 23:46, 31 January 2018 (UTC)|from1=
(and only|from=
/|from1=
).- @Ahecht: I simplified this by creating Category:Taxonbar templates with invalid from parameters, which is useful regardless.
- Now, any invalid/non-existent Wikidata entities should be added to this new cat (regardless of
|from#=
). - Now, the "desynced" cat should only contain pages where none of the
|from#=
s match the page's QID (I've struck the relevant part of my comment above for consistency). ~ Tom.Reding (talk ⋅dgaf) 15:04, 1 February 2018 (UTC)- @Tom.Reding: I don't think there is a way to "catch" errors caused by the wikidata lookup (see phab:T143970), so I can't add pages to that fourth category. The other category logic, however, should be fixed now. --Ahecht (TALK
PAGE) 18:28, 1 February 2018 (UTC)- @Ahecht: great! I'll check it later today. Amazing that something like that hasn't been implemented yet... FWIW, I'll ask around to see if someone knows a work-around. ~ Tom.Reding (talk ⋅dgaf) 18:41, 1 February 2018 (UTC)
- @Ahecht: looks to me like they're working and good to go live!
- I thought about asking at Help talk:CS1 since that's where all the cool coders hang out, but I didn't want to post something too off-topic, so I posted at VPT instead. Will probably x-post at WD, especially if no adequate responses here. ~ Tom.Reding (talk ⋅dgaf) 03:07, 2 February 2018 (UTC)
- Ahecht, I just looked through the Template:Taxonbar/testcases for the 4 tracking cats, and they appear to all be working (with the ever-present caveat "No plan ever survives first contact with reality."). Could you do the honors? :) ~ Tom.Reding (talk ⋅dgaf) 15:59, 8 February 2018 (UTC)
- Tom.Reding I don't feel we have all come to a consensus yet. Once these tracking categories are added to Taxonbar, Are you suggesting mandating the from1 parameter? I'm not sure I'm happy with what is implied by using the term 'missing' to describe an optional parameter. How about:
- I now see great value in using the from1 parameter in a case like Aeonium aureum where multiple Wikidata entries are linked to, but why was this so important to specify in Lupinus diffusus, Leucospermum cordifolium, Lestes rectangularis, Duckweed firetail, Enallagma pictum, Coenagrion interrogatum? I do not mean to be critical, but it seems to me you are suggesting ~98,000 articles need to be edited to specify a from tag on Taxonbar? The advantage to not specifying the from tag is that someone on Wikidata could update the taxonomy entry of Wikipedia articles and could clue in the reader and editors to a more recent taxonomic view that may not have been implemented on Wikipedia yet. The code is also cleaner and easier for most to use if it is simply
{{Taxonbar}}
on most articles. I'm still trying to understand why specifying the from tag in an article is important in cases where we are not actively linking to multiple wikidata entries.~ Mellis (talk) 18:43, 8 February 2018 (UTC)- I posted the preliminary categories 2 weeks ago. I then left them red for a week waiting for any input, but received none. In the interim, I did give it some thought. These are going to be tracking categories, meaning they are hidden by default from most users. While I don't want to make
|from1=
a required param, since I don't think this is something the average/casual editor would know how to get, I do want to nudge the editor to add it. This nudging will happen both passively, via editors seeing this|from1=
on other pages, and actively, via experienced WP:TREEers going through the various tracking categories to resolve discrepancies. Some moderately experienced editors with hidden cats turned on will also see it, who perhaps aren't privy to these discussions, and will be encouraged to make the improvement via the directions given in the cat. From a project perspective, I believe we do have consensus on the use of|from1=
(the strength with which we want to word that in the /doc is a separate matter though). Whensomeone on Wikidata could update the taxonomy
is exactly what specifying|from1=
uncovers (explained several times above). Without it, we have no idea what's changed at Wikidata. And the sooner we implement the tracking cats, the sooner we'll see the size of the issue. Therefore, I think "missing" is appropriate for a tracking cat (plus it's less verbose and uses fewer characters), open to consensus of course. - Yes, any page with a {{Taxobox}}/{{Speciesbox}} and {{Taxonbar}} will eventually have a
|from=
/|from1=
. This doesn't include all of the 98k transclusions, but certainly most of them. ~ Tom.Reding (talk ⋅dgaf) 19:43, 8 February 2018 (UTC)- My delay in responding was not due to lack of interest or stake in this topic, but were due to the limitations of time I can devote to The Free Encyclopedia.
- I fully support adding the hidden tracking categories, and agree with the reasoning for using them. However I continue to have great uncertainties regarding the impact of specifying and locking in
|from=
parameters across thousands of articles that could easily be forgotten about and ill-maintained for years. We do not have an easy, automated way to compare nomenclature specified in {{Taxobox}}, with what is directly connected via Wikidata, and what is specified in the|from=
parameter of Taxonbar. Until Taxobox is re-written to be more communicative between Wikidata and other templates, I feel like Taxonbar is walking a very awkward tightrope between Wikipedia, Wikidata, and a complicated array of various Taxonomic viewpoints. - Again, the word 'missing' in my view implies too much certainty about the need of the parameter. I do not share this certainty, but I understand your arguments for specifying the parameter. The simplicity of
{{Taxonbar}}
allows anyone to use the template without understanding Wikidata or parameters. This simplicity is largely why the voluntary rollout of Taxonbar has been so successful. - Eventually I see the Taxonbar
|from=
parameter being thrown out, in favor of {{Taxobox}} organizing the Wikidata entities. I see {{Taxobox}} be re-writen in Lua, and the editor would specify the QID in the {{Taxobox}} call rather than the Taxonbar call. Hopefully {{Taxobox}} would pass this QID onto Taxonbar. Because of this uncertainty in the future regarding how Taxonomy templates will all work together in relation to Wikidata, I do not want to take a solid stand or take on the responsibility of organizing taxonomic entities in the form of Wikidata QIDs.~ Mellis (talk) 05:33, 9 February 2018 (UTC)
- I posted the preliminary categories 2 weeks ago. I then left them red for a week waiting for any input, but received none. In the interim, I did give it some thought. These are going to be tracking categories, meaning they are hidden by default from most users. While I don't want to make
- Ahecht, I just looked through the Template:Taxonbar/testcases for the 4 tracking cats, and they appear to all be working (with the ever-present caveat "No plan ever survives first contact with reality."). Could you do the honors? :) ~ Tom.Reding (talk ⋅dgaf) 15:59, 8 February 2018 (UTC)
- @Ahecht: great! I'll check it later today. Amazing that something like that hasn't been implemented yet... FWIW, I'll ask around to see if someone knows a work-around. ~ Tom.Reding (talk ⋅dgaf) 18:41, 1 February 2018 (UTC)
- @Tom.Reding: I don't think there is a way to "catch" errors caused by the wikidata lookup (see phab:T143970), so I can't add pages to that fourth category. The other category logic, however, should be fixed now. --Ahecht (TALK
- Re
thousands of articles that could easily be forgotten about and ill-maintained for years
- tracking categories bring the problem articles to the fore and facilitate maintenance, not hinder it. - Re
[no] automated way to compare nomenclature specified in {{Taxobox}} [... to ...] {{Taxonbar}}
- correct, but it has been suggested to me by Peter coxhead, and I think it's doable (for later discussion). - Re
Hopefully {{Taxobox}} would pass this QID onto Taxonbar
- unfortunately that cannot happen since separate (non-nested) templates cannot communicate with each other. - Re
Because of this uncertainty in the future regarding how Taxonomy templates will all work together in relation to Wikidata, I do not want to take a solid stand or take on the responsibility of organizing taxonomic entities in the form of Wikidata QIDs.
- I think the quote "don't let perfection stand in the way of progress", by I-don't-know-who, is relevant here. This (and Wikipedia) is and will continue be an iterative process, and there are many capable people here able to take it on! Perhaps there could be a property created (or added if something like this already exists) to WD entities that lists all of their WD taxon name identifiers, which can then be enumerated by either {{Taxobox}} or {{Taxonbar}}? Either way, this doesn't preclude the need for|from1=
. ~ Tom.Reding (talk ⋅dgaf) 13:25, 9 February 2018 (UTC)
- Re
Mellis has proposed this change. The arguments for both versions are above, starting at the (now red) Category:Taxonbar templates not using from parameter, so they need not be (I hope) rehashed here! ~ Tom.Reding (talk ⋅dgaf) 13:30, 9 February 2018 (UTC)
- @Tom.Reding: does Category:Taxonbar templates without from parameter look better than my original proposal? I could go with that instead.~ Mellis (talk) 20:54, 9 February 2018 (UTC)
- *hand waving* That's an ok compromise; if there're no other comments here in a few days, sure. ~ Tom.Reding (talk ⋅dgaf) 21:04, 9 February 2018 (UTC)
Done. ~ Tom.Reding (talk ⋅dgaf) 23:08, 12 February 2018 (UTC)
- *hand waving* That's an ok compromise; if there're no other comments here in a few days, sure. ~ Tom.Reding (talk ⋅dgaf) 21:04, 9 February 2018 (UTC)
Benefits of from
It's clear from recent discussions that keeping some examples around of how |from=
is useful will be useful. Please feel free to add additional examples here. ~ Tom.Reding (talk ⋅dgaf) 17:30, 26 February 2018 (UTC)
- Acanthepeira cherokee — This is a time-condensed example of how
|from=
is useful. If the WD item changed weeks, months, or years after its original assignment, there is no persistent way for WP editors to know nor react to that, unless they happen to be online and watching their watchlist that day.|from=
provides this service. ~ Tom.Reding (talk ⋅dgaf) 18:12, 26 February 2018 (UTC) - Ancistrocheirus — Via tracking categories, it provides a means by which editors can examine a short-list of pages to address improper, or incomplete taxonomy. ~ Tom.Reding (talk ⋅dgaf) 18:12, 26 February 2018 (UTC)
- There used to be at least 600 pages in Category:Taxonbar templates desynced from Wikidata (which requires
|from=
to function) for various reasons. These issues were resolved quickly by WT:TREE editors. ~ Tom.Reding (talk ⋅dgaf) 16:17, 1 March 2018 (UTC)
Lua error: invalid capture index %5 in replacement string
This error at Phalonia yuccatana is caused by the presence of square brackets []
in the title of the external URL (to Butterflies and Moths of North America), which URL-encode to %5B
& %5D
, respectively. These get URL-encoded again to get the %
character, %25
; thus, %255B
and %255D
in the desired URL https://www.butterfliesandmoths.org/species/%255BNycthia%255D-yuccatana
. The once-decoded URL of species/%5BNycthia%5D-yuccatana
works at WD (Q19598561), but the {{Taxonbar}} module isn't dealing with this properly.
Using %255B
on WD wasn't an adequate solution because it broke the external link at WD (and didn't appear to fix the WP taxonbar, but perhaps lag was involved). Using []
on WD didn't break the WD link, but it wasn't applied correctly on Taxonbar. Is there a workaround or a code fix that can be done such that the link works both from WD and the Taxonbar? ~ Tom.Reding (talk ⋅dgaf) 14:59, 20 March 2018 (UTC)
- @Tom.Reding: fixed in the module. --Ahecht (TALK
PAGE) 15:59, 20 March 2018 (UTC)
- Glad you're around! ~ Tom.Reding (talk ⋅dgaf) 16:11, 20 March 2018 (UTC)
- Almost everything is broken now. The BAMONA link works on P. yuccatana, but as far as I can tell, not anywhere else. GRIN links and the link to the Wikidata item still work, presumably because these aren't stored as individual properties like all the other values. Plantdrew (talk) 18:28, 20 March 2018 (UTC)
- @Plantdrew: I was missing a set of parentheses, so the links should work now (I checked the testcases for link formatting, but not link destination -- my bad!). Are there other BAMONA links with brackets that still aren't working? --Ahecht (TALK
PAGE) 22:12, 20 March 2018 (UTC)- I don't know; inclusion of brackets is going to be a very rare case and I'm not sure how to search for other instances. Plantdrew (talk) 03:46, 21 March 2018 (UTC)
- @Plantdrew: I was missing a set of parentheses, so the links should work now (I checked the testcases for link formatting, but not link destination -- my bad!). Are there other BAMONA links with brackets that still aren't working? --Ahecht (TALK
- Almost everything is broken now. The BAMONA link works on P. yuccatana, but as far as I can tell, not anywhere else. GRIN links and the link to the Wikidata item still work, presumably because these aren't stored as individual properties like all the other values. Plantdrew (talk) 18:28, 20 March 2018 (UTC)
Several new potential tracking categories
From observations of various error/failure modes, and with side conversations with other prominent WP:TREE-huggers, I've gathered that the following 5 tracking categories would be useful (the 1st one was already suggested on the /doc, I just fleshed it out):
Category:Taxonbar template pages requiring a Wikidata item(implemented)Category:Taxonbar templates with duplicate from parameters(implemented)Category:Taxonbar templates using manual IDs (not an error)(implemented)Category:Taxonbar template pages without Wikidata taxon IDs(implemented)Category:Taxonbar templates with manual IDs differing from Wikidata (may or may not be an error)(implemented)Category:Taxonbar templates with manual IDs identical to Wikidata(added later; implemented)
Please have a look; suggestions welcome. ~ Tom.Reding (talk ⋅dgaf) 14:31, 24 February 2018 (UTC)
- Ahecht, if/when you're free, would you be able to enact some, possibly all, of these? ~ Tom.Reding (talk ⋅dgaf) 16:05, 5 March 2018 (UTC)
- 3/5 now implemented. ~ Tom.Reding (talk ⋅dgaf) 02:40, 3 April 2018 (UTC)
All done, and #6 added in the process.Not bad for my first real go at Lua, if I may say so... ~ Tom.Reding (talk ⋅dgaf) 21:03, 3 April 2018 (UTC)
Mobile view
As has come up recently in a discussion elsewhere, the taxonbar doesn't show in mobile view, which is a pity. (See Wikipedia:Enable mobile version to switch.) Peter coxhead (talk) 16:46, 9 April 2018 (UTC)
- Some previous discussion regarding taxonbar in mobile view took place in this thread: Template_talk:Taxonbar/Archive_1#Proposal:_Switch_Taxonbar_template_to_use_Module:Taxonbar. I do think taxonbar should display in mobile view. Plantdrew (talk) 19:55, 9 April 2018 (UTC)
- Ah, thanks Plantdrew; I hadn't seen that thread because initially I thought that Taxonbar would be like Authority control, which doesn't interest me, so wasn't following the discussion. Now that it can show all the synonyms, the basionym, etc., the situation is different. Purely as a source of taxon ids, it's clearly of limited interest. But now it's in fact a mixture of Bibliography and External links – it's certainly not a navigation aid – and there's no reason to hide the taxonbar from mobile view readers. So I agree that it should display, and we should re-open this discussion. Peter coxhead (talk) 09:16, 10 April 2018 (UTC)