Template talk:Cite doi

RfC: Should Template:cite doi cease creating a separate subpage for each DOI?

Following this discussion, Template:cite doi currently operates by searching the doi string against an individual template subpage within Category:Cite doi templates (each of which is simply a hardcoded {{cite journal}} citation). There are currently over 50k doi template subpages (more than 10% out of all templatespace) out of approximately 67 million doi in existence. Each citation is of very low (or no) usage. Should every subtemplate under Template:cite doi be substituted into each page as a cite journal citation? -- Ricky81682 (talk) 05:00, 9 July 2014 (UTC)

Support

1. Support The current system goes against WP:T3, namely each template is a hard-coded instance of a cite journal reference. Along with the massive effort required to keep watch of all these templates from vandalism, the additional complexity for new users creates does not make up for the minor savings created because people don't have to copy and paste a fairly small string. -- Ricky81682 (talk) 05:00, 9 July 2014 (UTC)
2. Support One of the very oldest guidelines for the Template namespace has been "Templates should not be used to store article content." Citations support article content, and are themselves article content. Splitting any substantive article text out of the main namespace makes it harder for the content to be reused elsewhere (transwiki'd, translated, etc.). Page watching is a major concern as well, as someone could change the cite doi child template, but that edit would go unnoticed and disconnected from the article itself. The other major reason for the policy is that templates are inherently hard for new users to understand. --Netoholic @ 06:14, 9 July 2014 (UTC)
3. Support unless they can be adapted to overcome the problems of the system as it stands at present. One of the main problems that I see with the existing system is that it is a fixed output and does not cater for the differing styles of output required for consistency in articles. For example the author fields need to allow for first/last or last/first format, date fields need to be output in day first, month first or ISO format. At the moment if the output style does not match that of the article the template has to be substed and modified to match. Keith D (talk) 11:18, 9 July 2014 (UTC)
4. Support difficulty monitoring, problems with cite var, difficulty in verifying etc. I like the idea in theory but in application it just doesn't seem to work so well, SUBST would allow continued use of function but resolve some issues. - - MrBill3 (talk) 20:52, 13 July 2014 (UTC)
5. Support. Not only I agree with the OP, my incentive for support is also some of the things that the opposing people have said. For example, User:RomanSpa says that some of the articles implementing the templates receive very little traffic. However, I do know for a fact that articles always receive more traffic than templates, as the templates have virtually no standalone readership. In addition, when a piece of info is brought out of an article an into a template, the weak link becomes two: The calling code in the article and the template itself; both need monitoring and maintenance. And since most of these are transcluded only once, the burden of monitoring soars. Also User:Headbomb says "newbies would not use the system in the first place" which is actually promoting elitism and is actually a big flaw. Coupling this with RomanSpa's comment means that these templates' benefit of readership through transclusion is curtailed.
   Best regards,
   Codename Lisa (talk) 03:55, 20 July 2014 (UTC)

   Comment. I think you may have slightly misunderstood me. When I remarked that the template pages receive little traffic, I meant what we might call "direct traffic" - that is, people accessing those pages specifically to look at the content of the templates. I didn't mean that they don't receive negligible "indirect traffic", by which I mean views arising from their information's transclusion in other pages. Every time an article that includes information from a template is read, that obviously creates "indirect traffic" for the template page.

   I'm sorry to say that I feel your remark about "elitism" is way off base: I think you're confusing elitism with experience. To take myself as an example, I didn't know about citation templates at all before being invited to participate in this discussion. Now I do, and with the appropriate learning curve will now be able to incorporate them into any articles I write. I don't think you guys who already knew about them are "elite", and I certainly don't consider myself "elite" just because I now know about this kind of template - I just think I've learnt a bit, and gained a little more experience. This is just how life goes. RomanSpa (talk) 04:46, 20 July 2014 (UTC)

   Let me just interject here that the statement 'Every time an article that includes information from a template is read, that obviously creates "indirect traffic" for the template page" ' isn't right. Because of caching, if article A refereneces template T, template T gets a hit only as often as A is edited (not every time someone just looks at A), and maybe (for technical reasons) a bit more than that. EEng (talk) 18:35, 26 July 2014 (UTC)

   Hi. Reading your clarification, I am becoming more and more certain that I didn't misunderstand at all. No matter how you look at it, putting the citations in the article exposes them to more scrutiny. As for your note on elitism vs. experience, let me put it frankly: You are sugarcoating elitism. Verifiability is one of the founding policies of Wikipedia. Every single person on this planet must be able to do it. This layer of obfuscation is only deterrent without benefit. Best regards, Codename Lisa (talk) 14:38, 20 July 2014 (UTC)

   Hi. I don't think I understand what you mean by "sugarcoating elitism": so far as I can work out, nobody in this discussion is advocating anything elitist, and all that I meant to say is that, as with everything else, you get better at Wikipedia with experience, and as you learn how to use more of the tools at our disposal. RomanSpa (talk) 17:41, 20 July 2014 (UTC)

   The only tool that should be necessary for the novice editor to update encyclopedic text (which includes citations) should be the edit window. -- Netoholic @ 20:10, 20 July 2014 (UTC)

6. Support. While I think vandalism occuring on these templates is highly unlikely, the possibility is still there. Furthermore, it's inconvenient to have to go and edit another page just to fix a journal citation. Per Keith D, there is no provision in the existing system for slight variation in citation styles between articles. Substituting the templates would be the best solution to these problems. APerson (talk!) 22:19, 30 July 2014 (UTC)

Oppose

1. Oppose. This should be a snow close. This system is widely used and supported. It saves space in article source and allows for the reuse of journal citations. Substitution would mean that the citations would be less likely to be consistent from one article to the next. The onus is on the OP to show that "massive effort" is required to keep these templates from vandalism. The OP posted similar questions at this discussion and received thoughtful explanations. To not link to that discussion here strikes me as surprising. – Jonesey95 (talk) 05:32, 9 July 2014 (UTC)
2. Oppose. Proposal is unclear: is the doi subtemplate to be deleted after subst:'s?, must the bot be shut down?.

   Proposer invokes argument WP:T3. But that is a criterium for speedy deletion, explicitly about deleting duplicate templates. One thing this bot does good: there are no duplicates to begin with (and no minor variants either). So WP:T3 is out.

   Indeed I am impressed by the numbers, but for the opposite outcome. If 50,037 doi templates exist, the bot must be adding something good to this wiki. Whatever that is, we'd not want to miss it. Leaving it to the editor who, by good intention, wants to add a doi citation to a page, this number would not be reached and the quality of the citations will be lower and more inconsistent. As for the unused doi subtemplates: a bit of research with the botoperator (like: where do they come from?), and then a mass speedy might be applicable.

   Netoholic points to one of the very oldest guidelines "Templates should not be used to store article content". Must say, this is the toughest nut to crack. In short: I claim to ignore this rule (guideline). Being a 'very old' guideline is not making it better. This even apart from the fact that "article content" has meanings with grey areas (cf., navbox vs see also section). The rule ignores two basics of the web: 1. Don't copy, reuse existing code/pages and 2. Transclusion is a core principle for webpages (and for wiki no less). It is way too primitive to define content by "what is on this physical webpage". That is where HTML classes are for. Using a template to store "article content" is, at least in this situation, a very, very convenient way to do additions, improvements, and maintenance. Then Netoholic also mentions: 'makes it harder for the content to be reused elsewhere' - maybe, maybe not. 'The other major reason for the policy [sic] is that templates are inherently hard for new users to understand'. Agree on the difficulty, though not sure about whether it is actually a reason. To reduce edit complexity, is there any other way, easier, to add template functionality to an article? And that for a (doi) citation? At least the bot helps the editor. This 'cleanup' idea is throwing out the baby with the bathwater.

   I have not read from an editor who actually works with these doi citations and this bot and says: "yeah, sure, bad thing". If one cannot convince working editors of the badness of this and its background, shouldn't that tell something? -DePiep (talk) 09:15, 9 July 2014 (UTC)

3. Oppose / Snow close This is a place where WP:CSD#T3 does not apply / and where WP:IAR applies. The cite doi system is designed this way on purpose to facilitate the maintenance of citations and cleanup the edit window. I do not use this system myself, but many do and I can't think of a good reason to remove this way of doing things. The argument that "this is confusing to newbies" makes no sense, as newbies would not use the system in the first place. While the risk of template vandalism exists on paper, in practice the cite doi templates are very very rarely vandalize because the average vandal focuses on random articles / high visibility articles, and those seeking to disrupt Wikipedia will pick on higher visibility templates. Headbomb {talk / contribs / physics / books} 19:48, 9 July 2014 (UTC)

   Its not vandalism that is the only concern. Any change to a shared citation (formatting, updated edition, etc.) could negatively impact or invalidate any article text that is sharing that cite_doi. Now, its actually very rare that templates are on more than one page, but that points toward the other bad aspect of this system. --Netoholic @ 20:40, 9 July 2014 (UTC)

   I read no reason for a snow close. -DePiep (talk) 00:23, 10 July 2014 (UTC)

4. Oppose. Proposal is unclear, but I'd oppose the shutting down of the bot using Template:Cite doi, which makes valuable edits "that would be extremely tedious to do manually". The bot also makes the citations consistent when they are re-used over multiple articles, which is common. I've never come across vandalism of an article via the template cite doi or Template:Cite pmid so can't see this adds any weight to the argument.
5. Oppose. I would like to keep this as an option until we have a better central citation database (as planned by https://meta.wikimedia.org/wiki/Grants:IdeaLab/Reform_of_citation_structure_for_all_Wikimedia_projects and https://meta.wikimedia.org/wiki/Wikicite). I see the cite doi template system as an acceptable temporary solution and as half a step to this (optional) final solution. So, i agree its not perfect but as a consequence I would like to substitute it for a better centralized system rather than deleting it.

   Such a central citation database on wikidata (for optional use) seems to solve most of the concerns mentioned here so far, right? Wikidata was planned for this kind of data (so guidelines support it), it was planned having transwiki usage in mind (other languages), datasets are put on the editors watchlist against vandalism, having different choices how to display the data could be implemented.

   To me it looks like the data could be much easier read/transferred from the uniformly structured cite doi subtemplates than it could be extracted from individual citations in articles (different formats /slightly inhomogeneous content). And I think authors who had chosen this central type of citation management agree with the migration to wikidata in future.--Saimondo (talk) 17:09, 10 July 2014 (UTC)

6. Tentative Oppose. I am absolutely not an expert in this area, having been brought here by an RfC invitation. However, I have read the discussion, and have done my best to learn how to use the template (though I understand that there are problems with the bot as at the time of writing). If I understand the discussion correctly, the idea is to move individual cite journal citations into the articles in which these citations are made, rather than maintaining a large number of templates, most of which receive negligible traffic. My strong intuition is that such a move would be unwise, because the present approach allows better control of the relevant data, since it is held in a single obvious place (the particular subtemplate), and articles using this data simply incorporate it by transclusion. This seems better for consistency purposes than having several copies of the relevant data in different articles. I do agree with some of the comments made above - I can imagine a better "central database" than a large number of separate subtemplate pages - but the system as it currently stands appears to work, and doesn't appear to be reaching any obvious capacity constraints, so changes should be made very cautiously.

   I do understand the concerns of the proposers of this idea, I think, and I'd like to address one concern. It has been noted that it is general practice that "templates should not be used to store article content", and this seems like a very sensible general rule. However, it seems to me that in this case this is not what is happening. Rather, the article contains a reference, and what is held in each subtemplate is more a sort of descriptive information. Each article could just give the doi, and leave it to the reader to disentangle what that doi means; the subtemplates, as I understand it, simply provide a sort of sprinkle of sugar to make it easier for the reader to swallow.

   If I've made any errors in my understanding, please let me know. I shall, in any case, add this discussion to my watchlist and review it again in a few days. Thank you for inviting me to contribute, and I hope I haven't sounded too stupid. RomanSpa (talk) 07:22, 11 July 2014 (UTC)

   A very thoughtful contribution that gives a comprehensive description of the topic. Worth reading, and a pleasure to read. (If I show jealousy here, that's OK with me ;-) ). -DePiep (talk) 21:16, 11 July 2014 (UTC)

7. Tentative oppose in plagiarism of Romanspa I am barely acquainted with the topic, but I go with Romanspa's argument. If someone will come up with a better-designed and well-accepted tool, fine. Until then however, it generally is an unhealthy practice to lump together multiple functions with different internal logical structures into a single metafunction. As a rule a better approach if a need for such a combined tool became recognised would be to write a hierarchically higher level of tool that invokes lower levels without distorting their specific intentions. Established functional practice I'd say. (Still tentatively!) JonRichfield (talk) 13:10, 29 July 2014 (UTC)

Threaded discussion

• Comment Similar schemes are used for the approximately 11k Category:Cite pmid templates and 1400 Category:Subtemplates of Template RussiaAdmMunRef. This RfC doesn't concern them but it's a parallel situation. -- Ricky81682 (talk) 05:00, 9 July 2014 (UTC)

• Comment The OP has notified a few editors on their talk pages. This may be perceived as canvassing unless the OP is careful to notify all editors involved in previous discussions, including this one. – Jonesey95 (talk) 05:32, 9 July 2014 (UTC)

I forgot that discussion. I only went off the previous TfD discussion. I'll notify everyone left. -- Ricky81682 (talk) 06:13, 9 July 2014 (UTC)

• Request To get some perspective, I suggest that everyone thinking about this RfC look at the reality of how this system is being used. Pick a couple dozen random Category:Cite doi templates from random category pages, and check the "What links here". A shocking number of them are completely abandoned, and most have only a single use. Any thoughts that we gain any significant level of consistency among articles is going blown away by how very rare that happens. -- Netoholic @ 11:05, 9 July 2014 (UTC)

Unused doi subtemplates? Start a blanket TfD for them, get arguments from the bot operator, and when agreed a bot can delete those. Now please stop panicking for an unused template. In now way that says the system is broken. -DePiep (talk) 23:50, 9 July 2014 (UTC)

I'm not sure that makes sense really. In the last discussion, the orphaned template was unorphaned by this edit, reinstating the template. Are the templates that are orphaned orphaned because they were substituted or orphaned because the citation has been completely removed? I'd assume the supporters of the current system would want the templates used as much as possible, so we'd need to check every 'orphaned' template to see if the template has been abandoned or the citation. I'm not making a strawman argument but it would seem strange to keep the template in its current form but not to actively search for all DOIs to be templated: either the system improves citing or it doesn't. Keeping it for the current uses only or for whoever feels like using it seems like a haphazard idea. -- Ricky81682 (talk) 04:00, 10 July 2014 (UTC)

As I said, it takes some thought and I already advised twice to seek advice from the bot operator. And yes, you do make a strawman. You are creating by your own reasoning that supporters of this template actually want to have them all, and so unused ones. Nonsense. Stop panicking for an unused template. -DePiep (talk) 08:00, 10 July 2014 (UTC)

• Comment – There is consensus within the WP:MED project that the {{cite pmid}}, {{cite isbn}}, and {{cite doi}} templates are a bad idea and should be substituted. Boghog (talk) 11:21, 9 July 2014 (UTC) There was also a related discussion here. Boghog (talk) 11:25, 9 July 2014 (UTC)

yeah but they didn't do anything so screw that. """Oppose"""" any changes and close and warn the crooked supporters for wasting people's time here. Biased violators have no business ruining the encyclopedia for their laziness. — Preceding unsigned comment added by 166.137.210.28 (talk) 23:14, 9 July 2014 (UTC)

In fact the project did do something (see BRFA). In the top 1500 most accessed pages within the WP:MED project, {{cite pmid}}, {{cite isbn}}, and {{cite doi}} templates have now all been substituted. Boghog (talk) 02:50, 10 July 2014 (UTC)

Beside the advantages of a central citation database mentioned here and here I understand why the MED people dislike cite doi. In the source code itself you see only the doi number and cannot see directly if it is really the right publication. You only see the citation data (title, name.. of the publication) when you have the tooltips option enabled in your preferences (is it by default?) when you hoover over the citation number in the preview mode (screenshot ). I wonder whether the MED guys were aware of that, I think its not that uncomfortable. Personally, I appreciate the cleaner looks of non-inline references, especially in case of several citations in one sentence. Example Although the source code coloring by the syntax highlighter helps, a higher number of detailed in line references can still be confusing (and impairs source code readability permanently / in the long term for the sake of one time checking the ref values within the source). So I would like to leave it to the author to choose. Maybe we can brainstorm about technical improvements in the citation database projects.--Saimondo (talk) 17:37, 10 July 2014 (UTC)

• Comment I don't know enough about the technical aspects involved to give a firm opinion at this moment. However, I have avoided using these templates for a while now, for 2 reasons. 1/ I don't think anybody has these templates on their watchlists. Any vandalism could go undetected for a very long time. 2/ If the name of a journal is linked to the WP article on that journal, the layout will be incorrect (name in bold) if the template is used in the article on the journal itself. As for the comments above about the bot, citationbot is indeed a great help (although recently it often seems to have problems completing). However, the bot works just as well with the {{cite journal}} template as with the cite doi template. When entering a reference, I use the cite journal template with "doi=xxxxxxxxxx" as only parameter and then, after I save, run the citationbot, which normally then fills in the rest of the template. --Randykitty (talk) 10:38, 11 July 2014 (UTC)

• Comment Maybe a year ago someone pointed me to the doi machinery, and at first I was delighted -- all that work saved! But after a month or two I gave up and subst'ed all the templates into the article and never came back. Why? Because the bot insists on doing things its own way, overwriting data I corrected with its own incomplete or incorrect data, blocking fixes to formatting problems (such as when an article title itself contains quote marks or an apostrophe) etc. I spent a lot of time on the bot talk page trying to get those active there to understand that there need to be ways for editors to take what the bot provides and formally & permanently alter it, but what I got was a lot of insistence that I shouldn't want to do what I wanted to do -- that I should just submit myself to The Will of The Borg and be happy. It just wasn't worth it.

  On top of that are problems such as those mentioned by others here, like the need to keep all these templates on my watchlist. EEng (talk) 18:43, 26 July 2014 (UTC)

• I understand. This layer obfuscation eliminates the much-needed flexibility. When I had taken my article to WP:FAC, reviewers insisted the referencing style must be consistent. They didn't care about syntactic and technical issues; they wanted a consistent citation style and in all fairness, they were right.

  Best regards,
  Codename Lisa (talk) 11:05, 27 July 2014 (UTC)