Wikipedia talk:JSTOR/Errors

Flat sorted list

Can these multiple groups be converted to in large sorted flat list (also without bare URLs etc.)? There are a number of duplicates present and many of the bare URLs read like gibberish with no indication of the subject matter or which language the content is in.--Cpt.a.haddock (talk) (please ping when replying) 10:07, 20 November 2015 (UTC)[reply]

I don't object, but I'm not sure I understand your rationale. https://it.wikipedia.org/wiki/Bernhard_Karlgren and Bernhard_Karlgren (it) both (to me) at a glance convey that the content is in Italian and the subject is someone named Bernhard Karlgren. Nikkimaria (talk) 15:25, 20 November 2015 (UTC)[reply]

The problem with turning it into one big list is that you'll get edit conflicts every time. By splitting it into sections you lessen the scope of the problem significantly. Also, there shouldn't be all that many duplicates left (maybe ~50 or so), so I don't think that's a big issue. PS. The "strange URLs" are an artefact of the software; the actual article in question is in the "from" parameter (for instance, for one of them it's Googlization). --Xover (talk) 16:02, 20 November 2015 (UTC)[reply]

@Xover: You're right. If the groups are created (next time) after a sort, it will be ideal.--Cpt.a.haddock (talk) (please ping when replying) 16:10, 20 November 2015 (UTC)[reply]

@Nikkimaria: Sorting the entries would have grouped all the duplicate Koh-i-noor articles (for example) together. It would have also grouped the Harappa articles (in no, hr, and en) together and so on. Re: bare URLs, I was talking about stuff like http://uk.wikipedia.org/wiki/%D0%9C%D1%96%D0%B7%D0%B8%D0%BD%D1%81%D1%8C%D0%BA%D0%B0_%D1%81%D1%82%D0%BE%D1%8F%D0%BD%D0%BA%D0%B0 and http://en.m.wikipedia.org/w/index.php?title=Special:ZeroRatedMobileAccess&from=Meridian_(novel)&to=#http://www.jstor.org/view/10624783/dm980390/98p0172n/0%3FcurrentResult%3D10624783%252bdm980390%252b98p0172n%252b0%252c03%26searchUrl%3D#http%253A%--Cpt.a.haddock (talk) (please ping when replying) 16:05, 20 November 2015 (UTC)[reply]

In your latter URL example, the article in question is Meridian (novel) (cf. Googlization, which I've now fixed). --Xover (talk) 16:13, 20 November 2015 (UTC)[reply]

Other domains

@Nikkimaria: @Sadads: I was initially looking for only urls with "jstor" in them, but I accidentally discovered some other domains like "journals.uchicago.edu" redirect to jstor as well. For example, the url "http://www.journals.uchicago.edu/cgi-bin/resolve?AJHG001967" took me to "http://www.jstor.org/stable/10.1086/303028" I'm worried we're overlooking a number of broken links in this effort. Gamaliel (talk) 17:02, 22 November 2015 (UTC)[reply]

Nice catch - likely we'll need to re-review some of the articles marked as not done. Nikkimaria (talk) 17:43, 22 November 2015 (UTC)[reply]

Ouch. It's worse than that. Since the presence of broken direct links to jstor.org does not preclude the presence of indirect links, we'll have to recheck everything. And some of these articles have simply massive References sections (400+) plus External links sections, all potentially with one or more links that will have to be verified. We've now entered bot-job territory here. --Xover (talk) 18:22, 22 November 2015 (UTC)[reply]

Is there a way to rerun the initial report and have it flag the specific url instead of just the article that contains it? Gamaliel (talk) 18:56, 22 November 2015 (UTC)[reply]

The report is a product of bad url errors on JStors end: when traffic shows up at a url that JSTOR can't resolve the reader accurately too. If you aren't finding the url in the page any more: its likely because the bad url has been removed from the article itself.Sadads (talk) 01:32, 23 November 2015 (UTC)[reply]

Sadads, looks like there are non-JSTOR links that redirect to JSTOR - in some cases it's these that aren't being resolved properly, and so someone looking for a JSTOR link would not find them even if the problem persists. The Interior, do you think Gamaliel's suggestion is feasible? Nikkimaria (talk) 02:15, 23 November 2015 (UTC)[reply]

I've now rechecked all the articles in Group 25, following every single external link on the rendered page and checking whether they end up on a jstor.org error page of some description. Of the 30 articles checked (maybe up around a 500 links total), 3 links landed on a jstor.org error page. That is, 10% of articles, but only something like less than 1% of links, are broken jstor.org links. However, the three links in question were all manually added (|url=<…> or [<URL> title]); are all to http://www.journals.uchicago.edu/; are all lookups of DOIs through journals.uchicago.edu; and all three cites also include working links to jstor.org through DOI lookup on dx.doi.org or direct jstor.org links using the stable URL. That is, you could reasonably argue that the problem isn't that Wikipedia's editors have inserted bad links, but that uchicago's DOI solver is broken. And if the root cause is that journals.uchicago.edu is broken, it would be a reasonable approach to target that specifically (remove links to it), rather than assume the problem is general and check every single link in an article.

PS. There is a shocking number of broken links in our articles, and an almost as shocking number of really poor quality links. --Xover (talk) 04:47, 23 November 2015 (UTC)[reply]

@Nikkimaria, Sadads, and Gamaliel: Ok, I've now rechecked all external links in 120 articles (Groups 22–25), and the count of newly discovered broken links to JSTOR stands at 10 (about 8.5% of articles). These represent several thousand links total so we're talking less than .5% of total links. All 10 of the new broken links have been redirects by way of www.journals.uchicago.edu; all of them have been in the |url=… parameter of a cite template; all the cases have had other identifiers (the broken link was redundant); they have mostly been confined to astronomy-related articles (G 117-B15A, Anticenter shell, Wow! signal, Ulysses, Extraterrestrial atmospheres); and, where possible to determine, they appear to have been added around 2007–8. I speculate that this suggests these are an old product of a small group of editors in a limited time period and on a limited group of articles.

Based on this I propose that we go on checking for the specific problems we are aware of (any direct links to jstor.org, and any link through journals.uchicago.edu) and not try to check every external link for other unknown and potential sources of problems (it is incredibly time-consuming and tedious!). This runs the risk of there being other issues like journals.uchicago.edu, but I find the likelihood of that tiny and its probable impact if there to be infinitesimal based on the numbers above.

Thoughts? --Xover (talk) 15:57, 24 November 2015 (UTC)[reply]

Absent feedback to the contrary, I'll proceed on this basis. --Xover (talk) 11:56, 26 November 2015 (UTC)[reply]

subscription= and via= parameters missing in references

@Nikkimaria, Cpt.a.haddock, Xover, Sadads, and Gamaliel: I only tried to fix about ten pages but I found most of the JSTOR citations to be missing the subscription= and the via= template parameters. Subscription should be: subscription=yes, and via should be: via=[[JSTOR]]

Did y'all come across this issue? PS: I had zero luck fixing any pages but I did solve some other reference issues. Cheers! {{u|Checkingfax}} {Talk} 08:01, 23 November 2015 (UTC)[reply]

@Checkingfax: I don't consider adding via/registration parameters a goal of this particular task. I've added them where I've otherwise done a lot of changes, but otherwise I've left articles as is (a lot of them don't even use citation templates for these links). That is, I'm treating this as purely an effort to fix linkrot in articles that JSTOR themselves have logged as containing bad links to jstor.org. --Xover (talk) 10:58, 23 November 2015 (UTC)[reply]

@Checkingfax: +1 to Xover: at the Wikipedia Library we are being proactive about the volunteer who get access to our sources, helping make the citations more transparent . However, as is general policy on Wikipedia: the subscription and via variables are non-standard. However, there will be a day when citations need to be structured data that live in either Wikidata or a similar database: at that point, we will worry more about cleanup of those variables, so that we can do more analysis of those variables (both will be useful for informing bot-related activity, or better access for readers). Sadads (talk) 13:26, 23 November 2015 (UTC)[reply]

Edward Elgar

Well, I tried: [1], [2] Andy Dingley (talk) 18:54, 23 November 2015 (UTC)[reply]

I was reverted as well. Not sure why someone would reject updating a citation. Gamaliel (talk) 19:12, 23 November 2015 (UTC)[reply]

See WP:CITEVAR and discuss if you wish to change existing (and wholly correct and functioning) citation styles. It's not only following WP's rules, it's also good manners. Tim riley talk 19:34, 23 November 2015 (UTC)[reply]

@Andy Dingley and Gamaliel: I've fixed it - in cases where citation templates are consistently not being used, we can just add the correct JSTOR ID to the URL to get the same result. @Tim riley: the citation style is fine but the link was broken - now corrected. Nikkimaria (talk) 19:39, 23 November 2015 (UTC)[reply]

@ALL: WP:CITEVAR says, essentially, that if a style of citation is established in an article you should avoid changing it. So I imagine what Tim objected to was the change in style, not that the broken link was fixed. I would do the same if a wikignome came by one of my pet articles and suddenly made big changes to the cites. We need to be sensitive to this issue and be minimally invasive where a consistent citation style has been established or where the regular editors of an article object to the changes: the goal here is to fix the broken links, nothing more. For most articles you'll find there is no consistent style used, and you should feel free to use the template; but if the regular editors object, try to find a way to address their concerns while still getting that link fixed. --Xover (talk) 19:48, 23 November 2015 (UTC)[reply]

I am much obliged to all concerned for their scrupulous additions and understanding. Tim riley talk 19:54, 23 November 2015 (UTC)[reply]

Recruiting help from non-English speakers through Wikiprojects

@Nikkimaria, Astinson (WMF), The Interior, and Juno: (not sure if this is all on Nikkimaria or what, so pinging all obvious suspects for broader input) I see we are making very little progress on the non-English articles in this list, so I propose to start contacting Wikiprojects related to each language and asking for their help. The message I intend to use is a per-project adapted version of this message:

Example for Afrikaans

Request for assistance from editors proficient in Afrikaans

As part of The Wikipedia Library project's collaboration with JSTOR—a repository of academic journals—whereby experienced Wikipedia editors can get free access to their scholarly resources, we have gotten a list of articles on Wikipedia that contain links to jstor.org which are in some way broken (anything that has been logged as an error on their servers: typically links to nonexistent articles). While a most of these articles are on the English Wikipedia, quite a lot of them are in articles on various non-English Wikipedias, and The Wikipedia Library could use some help from editors that speak the various relevant languages to fix these links.

I'm posting this message here in the hopes of finding editors that are proficient in Afrikaans and are willing to help fix the broken links in articles on af.wikipedia.org.

I am posting a message like this on several Wikiproject talk pages in the hopes of finding editors that can help out will all the 38 languages that are represented in the list. If this message is misplaced on this Wikiproject's talk page, or if there are other or better Wikiprojects I should contact, I would very much appreciate if you could let me know.

Thank you in advance, --

I intend to identify likely Wikiprojects by going to the enwiki article for that language (e.g. Afrikaans) and picking the most likely-looking Wikiproject from among the banners on its Talk page (i.e. WP:ZA for Afrikaans. Given the state of the world I'm a little worried about stepping into some tense situations in my general ignorance of what conflicts are currently raging either onwiki or off, so I'd very much like some level of consensus here before I flatfoot my way into an ArbCom-level situation. And, of course, if someone would prefer to take over this task I would certainly not protest.

PS. Everyone else watching this talk page should of course also feel free to chime in; the ping above is in no way intended to be exclusive. --Xover (talk) 09:41, 29 November 2015 (UTC)[reply]

Hey Xover, this is a cool idea. Another possibility that we might consider exploring first would be to reach out to global branches of The Wikipedia Library to see if any participants would be interested in helping. Nikkimaria (talk) 15:32, 29 November 2015 (UTC)[reply]

So how would we go about doing that? --Xover (talk) 18:25, 29 November 2015 (UTC)[reply]

We could either post it on the talk pages of each branch (eg. m:Talk:The Wikipedia Library/French), and/or I can ask on the coordinator mailing list. Nikkimaria (talk) 19:57, 29 November 2015 (UTC)[reply]

Well, you know the project better, but my first thought is that a message to the mailinglist would be a good way to start. Since the local branches do not have coverage of all 38 languages, the next step would then probably be to contact Wikiprojects. --Xover (talk) 10:43, 30 November 2015 (UTC)[reply]

Good idea. It's easy enough to swap out urls on other language wikis, but it's a struggle to do simple things like create a new citation for a bare url or add a citation needed tag. Gamaliel (talk) 16:27, 29 November 2015 (UTC)[reply]

@Xover and Gamaliel: - per Gamaliel: url changes in other Wiki's rarely cause responses (mostly reverts or "thanks/welcome"), however, if a citation isn't accurate and needs better metadata, then its best done by someone in the local language. In my volunteer time, I have done similar changes with better versions of images, and better citations/urls-- and have never gotten a negative response. Pinging EN-Wiki WikiProjects will usually get a few people who are multilingual, or are interested - per Nikki, we can also send an email to the coordinators list :) Astinson (WMF) (talk) 15:35, 30 November 2015 (UTC)[reply]

@Nikkimaria: (and everyone else): since the enwiki articles have now all been completed, but I'm still seeing very little movement on the non-English list, I'm going to go ahead and start contacting WikiProjects as per above. I'll be doing it somewhat piecemeal, and probably not too many at once, so it's bound to take a while. I'll probably put up some kind of contact list here (on Talk) to keep track of which languages and associated WikiProjects have been contacted. --Xover (talk) 19:31, 14 December 2015 (UTC)[reply]

Great, thanks so much! Nikkimaria (talk) 00:10, 15 December 2015 (UTC)[reply]