Category talk:CS1 errors: URL–wikilink conflict

Anyone else working on this category?

I started working on this category on April 22, when it had about 7,100 entries. Some have been added since then, and now we're just under 6,100. I have edited at least 842 articles since then to remove these errors; including articles affected by templates that I've fixed, I figure I've removed about 1,000 articles from the category.

I've noticed that the count sometimes goes up; it looks like a bot is adding articles, since when I look at the new articles, they usually have not been edited recently. So far, I have fixed all of the articles starting with: numbers; letters before A; Q, U, V, X, Y, and Z; and letters after Z.

I welcome any additional help. I also have some ideas about enhancing the "how to resolve" documentation on the CS1 help page, but I haven't gotten to it yet. I have found about five or six kinds of errors that cause 95+% of these problems, and I'd like to find time to explain how to fix them easily. Jonesey95 (talk) 03:52, 6 May 2013 (UTC)[reply]

If you can make a quick list of the most common errors and hints for repair here or perhaps at Help talk:CS1 errors, I'll integrate them into the help text. If you find any really oddball citations, consider adding them to Module talk:Citation/CS1/Rogues gallery.

Do you know which robot is creating the erroneous citations? Perhaps a note to its maintainer is in order.

—Trappist the monk (talk) 11:10, 6 May 2013 (UTC)[reply]

I don't know for sure that it's a bot. All I know is that I've cleared out all of the "Z" entries, for example, and then one day a new entry starting with Z shows up. I look in the article's history, and there are no edits. It's odd.

Here's a quick and dirty list of the main things that cause this error and how to fix them:

Wikilinked text in title: {{cite|...|title=Interview with [[John Doe]]|url=...}}. Fix like this: {{cite|...|title=Interview with John Doe|url=...}}. There is usually a mention of the wiklinked person or other noun in the main article text, and the wikilink does not work anyway, so remove it.
{{XX icon}} or {{subscription}} template in title: <ref>{{cite|...|title=El Oso Grande {{es icon}}|url=...}}. Fix like this: {{cite|...|title=El Oso Grande|url=...}} {{es icon}}</ref> (Put the XX icon template before the closing ref tag. Put a space between the two sets of braces. Use the same method for {{subscription}}.)
{{lang}} template around title: <ref>{{cite|...|title={{lang|es|El Oso Grande}}|url=...}}. Fix like this: {{cite|...|title=El Oso Grande|url=...}} {{es icon}}</ref>
Title and translated title in title parameter: <ref>{{cite|...|title={{lang|es|El Oso Grande}} ("The Big Bear")|url=...}}</ref>. Fix like this: <ref>{{cite|...|title=El Oso Grande|trans-title=The Big Bear|url=...}} {{es icon}}</ref>
More than just the title in the title parameter: {{cite|...|title=Interview with John Doe in [[Time (magazine)|Time]] by [[Walter Wilson]], June 8, 2008|url=...}}. Fix like this: {{cite|...|title=Interview with John Doe|work=[[Time (magazine)|Time]]|last=Wilson|first=Walter|authorlink=Walter Wilson|date=June 8, 2008|url=...}}
Rare: The link is embedded inside a template, so you can't find sample text from the reference in the article. This can happen with {{cite doi}} and similar templates. Instead of editing the article itself, click the (edit) link at the end of the reference in the article. Edit templates with care, as they can appear in many articles.

These may seem like they are similar enough to merge into a single piece of advice, but I think it may be helpful for inexperienced editors to have explicit instructions. I have found very few situations in which removing the url is the right solution. Jonesey95 (talk) 14:08, 6 May 2013 (UTC)[reply]

Null edits causing pages to be added to category?

I said above that I thought a bot was adding pages to the category without making an edit. I now think that these additions are caused by a user making a null edit, or by a background process purging page caches, which would (may?) have the same effect. The description of null edits says that if you edit and save a page without making any changes, it will purge the page cache and will not leave an edit summary in the history. This sounds like what might be happening. I suppose there may also be some sort of background process that periodically purges page caches, having the same effect. Just a thought.

For an example of a page that was added to the category in the past couple of days, see VirusBlokAda. I had previously fixed all articles starting with the letter V, except for a couple of Talk pages. VirusBlokAda showed up in the category during the past couple of days, and its edit history shows no edits since December 2012. Jonesey95 (talk) 03:09, 12 May 2013 (UTC)[reply]

It's not just null edits. About 300 articles have been added to this category in the past 36 hours. I don't know enough about how WP works to know what's going on here, but it seems like some sort of refresh or cache clearing is happening to these articles, and they are getting tagged as belonging to this category.

I had the category down from 7100 articles to under 5600, even with a trickle of new articles coming in, but now the trickle has turned into a steady stream, and it's up to 5800 again. I'm still keeping everything before A and after T clear, but the rest of the category is growing. Jonesey95 (talk) 12:47, 17 May 2013 (UTC)[reply]

In principle, whenever one of the citation templates / module is edited all of the pages that use that template should be updated. This should allow any error checking to be repeated and appropriate categories to be processed, etc. In practice though, the system often doesn't work quite right for templates that are used on very large numbers of pages. Some of the pages get skipped, and they aren't actually updated until some future process causes the page to be regenerated. In large part, what you are seeing are those stragglers that should have been added to the category weeks ago but because of glitches in the update process have only been added recently. As it has now been several weeks since this error checking was added, I would generally expect that most of the stragglers have now been dealt with. So hopefully things should be pretty stable now. Thanks for your help in cleaning up these issues. Dragons flight (talk) 07:47, 20 May 2013 (UTC)[reply]

Template changes to the cite templates could explain additions to the category without article edits. I can't explain most of the articles that just popped into the category last week, though. Like Z. D. Scott, for example. Or United Arab Republic at the 1965 All-Africa Games, or Unusual eBay listings, or Usta Gambar Garabaghi, or Walter Mengden. I had all of the articles starting with the letters U-Z fixed, so when about 400 articles popped into the category last week, I was 100% sure that the above articles had not been listed in the category before. I suppose they were stuck in some queue somewhere and got jostled loose by some stray process or edit.

I now have all articles starting with T-Z, as well as Q and all numbers, cleared out of this category. I'm continuing to work backwards through the alphabet at a rate of about 50 articles per day. People are welcome to join me. Jonesey95 (talk) 16:55, 20 May 2013 (UTC)[reply]

Category clearing update

I have removed all articles starting with N-Z, along with everything before A, from this category. Other people have cleared a few hundred articles (great work!). The category is down from almost 8,000 articles to 3,200. A tiny handful of articles, maybe ten or so, are added every week, but it looks like the red error messages are helping to prevent the category from growing.

In the N-Z sections of the category, I have left the Talk page articles alone, on the principle that I don't mess with other people's writing on those pages. Should pages in the Talk space really appear in this category? It's not a problem now, but it's something to think about for this and similar categories as they are cleared out.

The overarching category of "Articles with incorrect citation syntax" has about 135,000 members, many of which are duplicates (i.e. some articles have multiple citation errors). I have removed 4,000 – 5,000 articles from its subcategories on my own in three months, so it's reasonable to believe that we could clear out these categories, especially with a little bot-assisted editing. Jonesey95 (talk) 15:42, 24 July 2013 (UTC)[reply]

As of September 18, 2013, this category is empty! In just under five months, I fixed 5,000+ out of the original 8,000+ articles, and User:Gilo1969 fixed at least 1,350 articles. There are 96 articles remaining, mostly Talk space articles that I believe should be prevented from showing up in the category, since we shouldn't fix them. Can Talk space articles be set to display the error (it is useful to alert editors of their errors and demonstrate problems) but exclude the articles from this category? – Jonesey95 (talk) 14:10, 19 September 2013 (UTC)[reply]