Wikidata:Property proposal/Directory of Open Access Journals ID

Directory of Open Access Journals ID edit

Originally proposed at Wikidata:Property proposal/Authority control

Done: Directory of Open Access Journals ID (P5115) (Talk and documentation)

Description	ISSN identifier of a journal in the Directory of Open Access Journals
Represents	Directory of Open Access Journals (Q1227538)
Data type	External identifier
Domain	Academic journals
Allowed values	\d+-\d+
Allowed units	numbers with a hyphen in the middle e.g 0718-8358
Example	International Journal of Molecular Sciences (Q3153277) → 1422-0067
Format and edit filter validation	\d+-\d+
Source	https://doaj.org
Planned use	I'm importing a list of journals which appear in the Directory of Open Access Journals
Number of IDs in source	11270
Expected completeness	eventually complete (Q21873974) Proposer: I assume this will always be incomplete as journals will keep being added
Formatter URL	https://doaj.org/toc/$1
See also	ISSN (P236)

Notified participants of WikiProject Periodicals

Motivation

I have started to import the list of journals that appear in the Directory of Open Access Journals in the Data Import Hub John Cummings (talk) 15:58, 11 March 2018 (UTC)[reply]

Discussion

Comment I don't know why the formatter URL isn't working, I used $1 and it still has an error......can someone please fix? --John Cummings (talk) 16:21, 11 March 2018 (UTC)[reply]
- Fixed. The cause was a hidden character. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:04, 12 March 2018 (UTC)[reply]
Comment @Pintoch:, @ArthurPSmith:, @Jura1:, @Pigsonthewing: and @Vladimir Alexiev: who have been involved in the discussion on the Wikidata Import Hub. --John Cummings (talk) 16:05, 11 March 2018 (UTC)[reply]
Support David (talk) 16:03, 11 March 2018 (UTC)[reply]
Comment The ID is the ISSN, already created. --Gerwoman (talk) 16:20, 11 March 2018 (UTC)[reply]

@Gerwoman: thanks, yes, Directory of Open Access Journals uses the ISSN or eISSN as their ID numbering system (I think other journal directories also do this), this property proposal is to index Wikidata to DOAJ. --John Cummings (talk) 16:26, 11 March 2018 (UTC)[reply]

Oppose @John Cummings: (I cleaned up the proposal a bit. Don't worry about $1, maybe the template is broken right now.)

But @Gerwoman: is right: if DOAJ uses ISSN (P236) then we should not create a second external-id that's just a subset of ISSN. There is third-party formatter URL (P3303) to connect to the DOAJ site, and if you look at https://www.wikidata.org/wiki/Property:P236#P3303, you'll see someone has already linked ISSN to DOAJ.

Which doesn't mean that your work to import DOAJ is unimportant or unappreciated! Please go on, as you wrote DOAJ has a lot if important info that will enrich WD. You can still use MnM to match DOAJ journals to WD, just use "ISSN" as the linking prop.

Note that ISSN Discussion says "Current uses: 53,306". So your first task is to check which of the 11270 DOAJ journals are already on WD, and to exclude them from MnM: matching by ISSN should be easier than matching by name.

The easiest way to get all ISSN out of WD is https://tools.wmflabs.org/wikidata-todo/beacon.php

Of course, you can and should add extra DATA about these journals. But you can spare the matching step in MnM.

And please move Wikidata:Data_Import_Hub#Directory_of_Open_Access_Journals to a separate page, that will make collaboration so much easier. Thanks! Vladimir Alexiev (talk) 18:50, 11 March 2018 (UTC)[reply]

Oppose per Vladimir - let me know if you need help. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:04, 12 March 2018 (UTC)[reply]
Comment Thanks @Pigsonthewing: and @Vladimir Alexiev:, @NavinoEvans: did the automatching by ISSN number and we uploaded the rest to Mix n' Match for manual matching here. New page coming soon. --John Cummings (talk) 10:35, 13 March 2018 (UTC)[reply]
- I've marked the proposal as withdrawn. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:43, 13 March 2018 (UTC)[reply]
Support --Succu (talk) 05:57, 30 March 2018 (UTC)[reply]
strong Support - much needed, both for direct linking to DOAJ, but also for knowing which publications are in DOAJ or not. The fact that the chosen ID is ISSN is irrelevant in that matter... A lot of databases use the same ID as other DBs, which did not prevent to create a property for each (ex : IdRef ID (P269) which is for authors and Theses.fr person ID (P4285) which is for thesis authors or jury members (in France), and do not link to the same database at all... --Hsarrazin (talk) 07:38, 30 March 2018 (UTC)[reply]
Support Mahdimoqri (talk) 13:50, 30 March 2018 (UTC)[reply]
Comment I wonder if we should create different properties for worldcat, crossref, opac, biostor, scimagojr, sciencedirect, erihplus... All of them use ISSN as DB ID. --Gerwoman (talk) 14:30, 30 March 2018 (UTC)[reply]
- There is more than one identifier for journals at BioStor. --Succu (talk) 14:48, 30 March 2018 (UTC)[reply]
- Gerwoman that's poor argument : these db have a field for ISSN, like all libraries, because it's the international standard id, but it is not the main ID used to connect to records... Worldcat uses OCLC control number (P243), Crossref uses DOI (P356), etc. DOAJ only uses ISSN to create links. Thus, you cannot create a link without using it. --Hsarrazin (talk) 15:10, 30 March 2018 (UTC)[reply]
  - And Erihplus uses ERIH PLUS ID (P3434). strakhov (talk) 23:30, 30 March 2018 (UTC)[reply]
  - I would support an additional property for Crossref journals as well since that is the only id used for a journal in Crossref
- I'm just wondering, Hsarrazin. I didn't "voted". If the planned use of this property is to "import the list of journals", perhaps we can use the ISSN which is a unique id. On the other hand, how many properties are necessary to identify a journal? One for each database where it's listed? Only a few of them? Just asking. --Gerwoman (talk) 17:22, 30 March 2018 (UTC)[reply]
Support since it is focusing on a subset of journals (that have open access) and considering the fact that it is also an external identifier. The internal working of identification of DOAJ (e.g., use of ISSN or eISSN) doesn't seem very relevant to me. John Samuel 09:23, 31 March 2018 (UTC)[reply]
Comment I think we already have ISSN. Somehow the proposal omitted that this is what's it for.
--- Jura 09:25, 3 April 2018 (UTC)[reply]
Support; this seems to be the most practical way of actually linking out to DOAJ, and for that matter for telling whether or not a journal is listed in DOAJ. URL formatters on the ISSN property don't help here, as there's no practical way to tell if a journal will be listed there or not. We could try and work around this with some kind of "indexed in" property but that seems a complicated workaround. Andrew Gray (talk) 16:43, 3 April 2018 (UTC)[reply]
Comment How to avoid items with this property but without ISSN? Or how to avoid duplicated items, one with ISSN and another one with this property? --Gerwoman (talk) 18:19, 3 April 2018 (UTC)[reply]
- A simple SPARQL query helps. --Succu (talk) 18:23, 3 April 2018 (UTC)[reply]
  - How often do you propose to run the query? --Gerwoman (talk) 18:33, 3 April 2018 (UTC)[reply]
    - If needed, as part of the constraint violations runs, I hope you'll help to fix them, if there is anything to do. --Succu (talk) 18:42, 3 April 2018 (UTC)[reply]
    - But you can start with Wikidata:Database reports/Constraint violations/P236 right now. --Succu (talk) 18:46, 3 April 2018 (UTC)[reply]
    - That's the point. ISSN is one of the few real unique identifiers. Nevertheless we have duplicates in Wikidata. Now we are proposing a new identifier that, in fact, is the same, but only for a subset of journals. I don't think that this new property proposal will help to reduce the problem... But, as you wish. --Gerwoman (talk) 18:56, 3 April 2018 (UTC)[reply]
- Questions It's true that it does add another potential source of differences. Is it easier to maintain this in one property or with a second new property? Or should we just ignore that aspect? e.g. If we consider that there aren't that many problems on the report for the existing property.
  --- Jura 10:01, 5 April 2018 (UTC)[reply]
  - Is it possible to have a constraint that says : should be equal to ISSN (P236)? this would certainly help catch those... --Hsarrazin (talk) 10:26, 5 April 2018 (UTC)[reply]
    - I think the question is mainly what happens if there is a problem with P236. How will this impact this?
      --- Jura 10:39, 5 April 2018 (UTC)[reply]
    - @Hsarrazin: I have created a phabricator ticket for that: https://phabricator.wikimedia.org/T191963 − Pintoch (talk) 14:10, 11 April 2018 (UTC)[reply]
Support As per User:Andrew Gray.Battleofalma (talk) 09:50, 5 April 2018 (UTC)[reply]
Comment We could use DOAJ Journal ID instead of ISSN with the same formatter URL - e.g. International Journal of Concrete Structures and Materials (Q50668214) → 40a0d2383cfe43418de63f932b03a66a. DOAJ Journal IDs can be retrieved using the api. Article level IDs are also available and could potentially be useful as external identifiers. Simon Cobb (Sic19 ; talk page) 22:04, 10 April 2018 (UTC)[reply]
- I think that if we start digging for internal identifiers in websites that conveniently use well established, shared unique identifiers that already exist in Wikidata, the situation is pretty serious. IMHO creating a property for that internal identifier would be worse than creating a new property using ISSNs. There is duplication, but at least it is a duplication that is easy to maintain. − Pintoch (talk) 14:08, 11 April 2018 (UTC)[reply]
Support per John Samuel. NMaia (talk) 01:48, 1 May 2018 (UTC)[reply]
Oppose I don't see any reasonable plan to tackle the problem mentioned.
--- Jura 09:21, 1 May 2018 (UTC)[reply]
- @Jura1: I am not sure I understand - you personally oppose but you have marked this as ready anyway? − Pintoch (talk) 12:25, 1 May 2018 (UTC)[reply]
- I don't think the two need to be linked. One is about my view of the proposal, the other about my view of the proposal discussion. I think the other day I even nominated a property for deletion that was created more or less in line with my comments, but not with the discussion.
  --- Jura 12:39, 1 May 2018 (UTC)[reply]
@Jura1:, thanks for marking it ready, do you have the magic powers to press the button and create the property? --John Cummings (talk) 08:57, 2 May 2018 (UTC)[reply]
- Sure, but there 88 other property creators and none seems to have opposed. If you can't find one of them, maybe I would.
  --- Jura 17:22, 2 May 2018 (UTC)[reply]
question An alternate approach could be to just create one or two items and add all identifiers there.
--- Jura 17:22, 2 May 2018 (UTC)[reply]

@jura1:, sorry I don't understand what you mean, can you explain further? For information there are 11,000+ journals registered in DOAJ. Thanks, --John Cummings (talk) 18:30, 2 May 2018 (UTC)[reply]

You could add all of them with P236 to 2 items.
--- Jura 19:30, 2 May 2018 (UTC)[reply]

@Jura1:, I'm sorry, I really don't understand, can you use an example? The each ISSN applies to only one item, DOAJ's identifier scheme uses ISSN numbers as their ID numbering system, however they have individual information pages for each journal which is helpful to link to, hence the reason for this proposal. --John Cummings (talk) 19:39, 2 May 2018 (UTC)[reply]

@John Cummings, Pigsonthewing, ديفيد عادل وهبة خليل 2, Gerwoman, Vladimir Alexiev:@Succu, Hsarrazin, Mahdimoqri, Strakhov, Jsamwrites:@Jura1, Pintoch, Battleofalma, Sic19: This is a similar case to what we had with P:P5037, so I don't see any reason why this property shouldn't be created. Done as Directory of Open Access Journals ID (P5115).--Micru (talk) 21:27, 2 May 2018 (UTC)[reply]

@Micru: thank you very much indeed. --John Cummings (talk) 21:29, 2 May 2018 (UTC)[reply]

@Micru: thanks, indeed :) --Hsarrazin (talk) 08:49, 3 May 2018 (UTC)[reply]

question@John Cummings, Pigsonthewing, ديفيد عادل وهبة خليل 2, Micru, Vladimir Alexiev:@Succu, Hsarrazin, Mahdimoqri, Strakhov, Jsamwrites:@Jura1, Pintoch, Battleofalma, Sic19: Please have a look at the Mix'n'match 1070 catalog. Take some random items (Q50812261 Q50811063 Q50432013 Q50808092 Q50557942 Q50552657). All of them have P:P236 ISSN P:P854 https://doaj.org/toc/... (two times) P:P248 Q1227538. What do you suggest to do before adding P:P5115 ? --Gerwoman (talk) 10:57, 5 May 2018 (UTC)[reply]

@Gerwoman:, we are going to add this property to these items very soon from our spreadsheet automatically. Just finishing off some final Mix n' Match. --John Cummings (talk) 11:03, 5 May 2018 (UTC)[reply]

Oppose I know the voting's over, but I want to add something, since I think we're making dangerous precedent here. ISSN **is** the one and only id used by DOAJ, so "it is not relevant what id is used by DOAJ" is unconvincing rhetoric. So we're describing a subset of all journals having ISSN by making another unrelated property (DOAJ ID). Thus in reality, DOAJ ID is a sub-property of ISSN, but WD doesn't (yet) have sub-properties: the proponents of this prop must argue with the WD maintainers to introduce a sub-property mechanism (or some complicated constraint mechanism). I think it's better to describe the subset with an "indicator" property, eg "listed in catalog: DOAJ" or something similar.

We're creating a maintenance problem here: the initial population of the sub-property is not enough. Eg if a journal changes ISSN (that does happen as publishers modify or merge imprints), someone needs to remember to change two values. --Vladimir Alexiev (talk) 14:10, 8 May 2018 (UTC)[reply]