False positives for Category:Pages that use a deprecated format of the math tags
Open, Needs TriagePublic

Description

The category
https://en.wikipedia.org/wiki/Category:Pages_that_use_a_deprecated_format_of_the_math_tags
created to detect deprecated syntax in related to T197842 is giving some false positives.

The minimum false positive example is <math>\%<math> indeed anything with an escaped % seems to trigger the category.

Another problem is mentioned in https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:WikiProject_Mathematics&oldid=994978378

<math>E\lor F</math> E ∨ F {\displaystyle E\lor F} {\displaystyle E\lor F} is triggering the error category, but <math>\lor</math> is fine.

Event Timeline

Another problem has been spotted with some formula like \mathbb{C}^7, which was failing but others like \mathbb{C}^6 worked fine.

To reproduce first pick some high index say 67. Create a page with <math>\C^{67}</math>, and save, triggering the deprecation error. The change it to <math>\mathbb{C}^{67}</math> this keeps the error. https://en.wikipedia.org/wiki/User:Salix_alba/sandbox1 has this error.

Entering a new expression using <math>\mathbb{C}^{68}</math> does not trigger the error, forcing a mathpurge also clears the error.

This was discussed at https://en.wikipedia.org/w/index.php?title=Wikipedia:Help_desk&oldid=995140523

I checked it with https://mathoid2.wmflabs.org/info.html . Seems to be restbase / caching problem. I have discussed with @WDoranWMF @AMooney @Pchelolo @daniel and others about the plans to avoid restbase T252389. Unfortunately, I am uncertain what the speed of this activity is. We targeted bi-weekly progress checks. However, due to the global pandemic, things did not work out as planned. We touch base from time to time, but the innovation speed is approximately zero as we also need to deal with the foreseeable headwinds like deprecations, incompatibilities, browser-bugs, and code maintenance.

Here is another example of a false positive: <math>\partial</math> in the article https://sv.wikipedia.org/wiki/D

Here is another example of a false positive: <math>\partial</math> in the article https://sv.wikipedia.org/wiki/D

Same example here (French Wikisource).

Another problematic expression cached as a problem is x \in \mathbb{C}, see https://test.wikipedia.org/wiki/Cached_deprecated_math, any change (e.g. to y \in \mathbb{C}) makes the deprecation to disappear.

This last one is cache problem. Trying to purge things with action=purge&mathpurge=true does not clear this error.

It is not a cache problem (or if it is, it is a cache problem that has lasted six months!). I am consistently seeing hidden category "Pages that use a deprecated format of the math tags" added for any expression that includes \mathbb{C} (but not for other mathbb letters). This format should not be tagged as problematic. Purging does not clear the error, and I get the error on multiple pages (including in my sandbox on preview).

This occurs, for instance, in https://en.wikipedia.org/wiki/Template:Number_systems, and appears to be the cause of the deprecated format category on that template and on the many Wikipedia articles that use that template.

What I mean by a caching problem is that it depends on historical information about the formula rather than recalculating whether formula should cause the page to be placed in the category.

Somewhere in the system the exact formula

\mathbb{C}

is recorded as being a deprecated syntax. Every time that exact syntax is used the page gets put into the category. There does not seem to be a user side way of clearing this information. Other formula like

\mathbb{C} \,\!

that appears the same but as it is a new formula causes the algorithm to recalculate it is deprecated syntax.

As the priority of removing deprecated formula is low, adding workarounds to the formula to remove the page from the category is counter productive.

I'm experimenting with a whitelist system. I've edited https://en.wikipedia.org/wiki/MediaWiki:Math-tracking-category-texvc-deprecation so that specific pages can be excluded from the tracking category. The way this works is that there is a switch statement

{{#switch:{{FULLPAGENAME}}|Template:Number systems=|Pages that use a deprecated format of the math tags}}

any page mentioned in the switch will not appear in the tracking category. I think you need to be an administrator to edit the page so requests can be placed on the talk page.

I have implemented a solution to this problem in December 2020. Some parts are "already" merged. However, at least 2 steps are missing. I hope someone from WMF will find time to look at my code.

https://en.wikipedia.org/wiki/Template:Or- has <math>A\lor B</math> which is correct syntax but triggers the category. Lots of pages include this and they all get included in the category.