Talk:New requirements for user signatures

About this board

Welcome! Please use this page to ask questions and share information.

This is a Structured Discussions (Flow) page. Click in "Start a new topic" or reply to an existing thread. Use the pencil icon to switch to wikitext mode. If you find a typo after you post, then the ••• menu is where you find the Edit button. Signatures are automatic.

Whatamidoing (WMF) (talkcontribs)

Just a quick heads-up to say that the team decided to implement most of this, but not all of it. Official announcements will be forthcoming, but the overall notion is this:

  • They will not disallow all obsolete HTML tags. That could be done later, but they wanted to keep it simpler at this stage.
  • Most of the work will happen during 2020.
  • After they make the initial changes to MediaWiki, we'll notify active editors if their signatures need to be updated.
    • I think that the most likely approach is a MassMessage for each separate problem ("you need to un-tick that box" to one list, "you need to get the template out of your sig" to another list, etc.). @AntiCompositeNumber has a good system for generating these lists, which means that anyone can start this work at any point, or you can wait for me to send messages.
    • If you can help with translation and localization of these messages, please watch this thread (click the star in the upper corner of *this* thread to make it blue, even if you're watching the whole page [which I recommend, too]).
WereSpielChequers (talkcontribs)

It would be good if you could composite media messages where editors are going to get multiple messages, alternatively phase the messages apart so that someone with three problems gets their messages a month or more apart - that would also spread the support side of this. But I do hope you change the signature box validation process first - so any new or changed signatures have to be fully compliant. We don't want a situation where people can fix one problem in their signature by replacing it with something that is about to be a problem

Whatamidoing (WMF) (talkcontribs)

Yes, I agree that the software changes come first.

Are you aware of any editors with three separate problems in their signatures?

WereSpielChequers (talkcontribs)

Hi WhatAmI, I would predict that if you make three changes to the software as to what is a valid signature there will be some editors to whom you cause three problems, but i don't know - I would hope the people who want to make these changes would know. However I would suggest a slightly more editor friendly approach re this - remember the foundation is making changes that are causing a problem to these people's signatures, not that these are editors with problematic signatures. A subtle nuance, but a very important one.

Jonesey95 (talkcontribs)

Some of these editors *do* have signatures that cause problems, like unclosed tags that cause the whole rest of the page to be rendered in a smaller font, or with strikeout text. I provided links on this page or a related page.

WereSpielChequers (talkcontribs)

Hi Jonesey, I don't know about all other projects, but on the wikis where I am most active, if you use a signature that causes the sorts of problems that you describe, people will point out that you have an invalid signature. Dealing with such signatures is not the problem area, the problem is with signatures that are currently valid, and have been valid for some years. Telling people that a change has been made that requires them to make a change to their signature is different than telling people they have made an invalid signature.

Jonesey95 (talkcontribs)

I think we agree then. This page is only about signatures that are invalid in some way.

Whatamidoing (WMF) (talkcontribs)

One of the problems is ticking a box to say how you want your custom signature interpreted, without having a custom signature for the software to interpret. This obviously won't be found in combination with any of the other problems, since the existence of any custom signature, including a problematic one, would make this non-problematic. (Now that I think of it, those messages could go out at any time.)

Nested substitution has been banned in policy since before any of us were baby editors, so I don't expect to find many unblocked/unbanned accounts with that problem. I can imagine the other two problems being combined in the same sig, but I don't remember anyone actually reporting an example of it.


MaxHarmony (talkcontribs)

To clarify, is this change being rolled into the publicly-available software, or is it an internal revision for Wikimedia? If it's being rolled in, will it be a configurable option?

Whatamidoing (WMF) (talkcontribs)

This will affect WMF wikis, e.g., the English Wikipedia. I am not certain if it will affect the default MediaWiki core for anyone who upgrades their private MediaWiki servers, but I believe that it will be. Do you have your own wiki?

Tacsipacsi (talkcontribs)

I think this should certainly be configurable for non-WMF wikis. I’ve had hard time disabling all password complexity rules, so that I can have 1234 as a sysop password in my own wiki, running on my own computer, in my own local network—nobody will ever try to hack that account, as only my family has even a chance to access the wiki. Similarly, different wikis may have different rules and customs; most wikis will probably never use DiscussionTools, so this change is not technically needed on them etc. (Just to reassure: I have a much stronger password than 1234 here on MediaWiki.org, as it’s a real possibility that someone tries to hack into my Wikimedia account.)

Reply to "Outcome"

Do you have an example signature?

28
Whatamidoing (WMF) (talkcontribs)

Is there a custom sig that you think should be accepted, or rejected, or you're just curious whether it would be accepted? You can post it here if you want to.

AntiCompositeNumber (talkcontribs)

For the requirement for a link to user/talk/contribs: can the link be an interwiki or interlanguage link? For example, users @GB fan and @DerSpezialist on enwiki currently use the signatures [[:en:User talk:GB fan|~ GB fan]] and [[:de:Benutzer:DerSpezialist|Spezialist]]<sup>([[:de:Benutzer Diskussion:DerSpezialist|talk]])</sup>, respectively.

Whatamidoing (WMF) (talkcontribs)

@Matma Rex, what do you think of this? It's not unusual for a MassMessage to be sent "from" m:User:Whatamidoing (WMF). Although that's done manually per-message, it's possible that some people might set their prefs that way, especially outside their home wiki.

AntiCompositeNumber (talkcontribs)

Note: I did see one editor in my dataset with an interlanguage link like [[en:User talk...]]. If IWLs/ILLs are allowed, an unprefixed ILL like that (which would show up in the sidebar, not as an inline link) should generate an error.

Whatamidoing (WMF) (talkcontribs)
AntiCompositeNumber (talkcontribs)

Yes (the one mentioned here). That specific link works, since it's already on enwiki. However, when that exact signature is copy-pasted to another wiki (or set globally), it might not work (depending on how the interwikis are set up).

Whatamidoing (WMF) (talkcontribs)

I'm wondering whether we might see more interwiki links at Meta.

AntiCompositeNumber (talkcontribs)

@Whatamidoing (WMF) They are more prevalent on Meta, but there are fewer signatures overall and 36 times fewer potentially problematic signatures of recently-active Meta users. data

Matma Rex (talkcontribs)

The main problem with these is that namespace names are translated, so we'd have to load localisation data for all the hundreds of languages, which is actually way slower in MediaWiki than you'd think (once upon a time I tried to do that in some unit tests, it took several seconds).

If folks really want this, then I suppose we could add an exception so that any signature with an interwiki link is valid, but we realistically can't validate that the link points to a user page (and DiscussionTools, Echo etc. won't be able to detect this signature).

I'm not sure if having a signature like this makes much sense these days. We have cross-wiki notifications, so why ask other users to reply to you on another wiki? (I actually used to have a signature like that on en.wp, e.g. … good times.)

Note also that you could easily have both a local and an interwiki link, which to me seems like the best of both worlds.

AntiCompositeNumber (talkcontribs)

Another question: what about links to redirects? On nl.wiktionary.org, @~riley uses -[[User:Riley Huntley|Riley Huntley]] [[Meta:SWMT|(SWMT)]]. That page is a redirect created during renaming, and I expect that there are plenty of similar signatures on other wikis.

Matma Rex (talkcontribs)

DiscussionTools would not follow the redirect, so it will detect the "wrong" user as the author of a comment. Right now that doesn't really cause any problems, but in the future we might want to send notifications about the reply or something. (And remember that after an account is renamed, a new account can be registered under the old name, so I don't think it would be correct for us to just follow the redirect.)

I think we should treat this as invalid.

Cabayi (talkcontribs)

that doesn't really cause any problems, except that the Navigation Popups gadget will be showing the credentials of the previous account (either a non-existent account, or zero if the user created a new account there to block misuse), and tools which highlight users based on their permissions won't work.

Perhaps the change username process should also blank non-default sigs to ensure they use the new username?

AntiCompositeNumber (talkcontribs)

The most user-friendly option would be to automatically replace User:OldUsername with User:NewUsername in fancy signatures, but that is likely to take the most development effort. Blanking the signature from an active user isn't going to be a great idea (I think we could just about get away with it with old accounts, but someone who just had their account renamed is likely to notice and complain), but the "ignored until fixed in preferences" state proposed by others here would work. Adding a message to the rename complete notification for users with fancy signatures would also be a good idea.

Tacsipacsi (talkcontribs)

As tools read existing signatures, the biggest problem is the comments signed before the renaming. This is a problem also for the vast majority of users who don’t use fancy signatures. This may be out of the scope of this project, but I think it has a much bigger impact than fancy signatures placed after the renaming. The latter will, in fact, be disallowed if the current proposal gets implemented, as not updated fancy signatures no longer contain a link to the new user page.

Cabayi (talkcontribs)
~riley (talkcontribs)

I honestly just don't have time to change my signature across 800 wikis. There should be technical support from Mediawiki to make this possible through preferences, a special page or a user script.

Tacsipacsi (talkcontribs)

Or a Toolforge tool. There is a grant named editmyoptions, which allows OAuth tools to change preferences.

Cabayi (talkcontribs)

Wondering, we have global user pages, global js, global css, why not global sigs?

~riley (talkcontribs)
AntiCompositeNumber (talkcontribs)
AntiCompositeNumber (talkcontribs)

Yeah, I've got another one. On de.wikipedia.org, @Doc Taxon has <small> {{ers:#timel:H:i, j. M Y (T)}}</small> at the end of their signature and appears to sign with ~~~ instead of the usual four tildes. This creates a correctly formatted (for dewiki) timestamp, but surrounds it with <small> tags. Looking for a properly-formatted date is one of the current methods for signature detection, and the extra markup might confuse that method.

Whatamidoing (WMF) (talkcontribs)

That's an interesting one. I don't know whether the Reply tool is looking for the timestamp to be at the very end of the line, but other tools might.

Doc Taxon (talkcontribs)

The reply tool handles my signature correctly. Is there a problem with it?

Whatamidoing (WMF) (talkcontribs)

In terms of this product, I don't know.  @Matma Rex would be the best person to answer that question. In terms of other tools (e.g., an archiving bot), it's possible that some of them are confused, but if you're not getting complaints, then it's likely that the more popular tools at the German-language Wikipedia can handle it.

Matma Rex (talkcontribs)

@AntiCompositeNumber @Doc Taxon I tried a similar signature locally and it doesn't seem to work with DiscussionTools (there is no "Reply" link for it). Are you talking about a different reply tool?

The code in DiscussionTools is intentionally quite strict to avoid incorrectly matching random userpage links as if they were signatures; we pretty much expect exactly the output of ~~~~. I don't think we should be adding special cases for other things, except possibly auto-signing bots with thousands of edits.

In general I'm not a fan of customizing the timestamp. I guess <small> is harmless, but I've seen user signatures that put the timestamp inside of the userpage link, or where the timestamp has a different date format, and I'd personally consider those disruptive. And it would be difficult to draw a line between those.

For the record, the proposed new validation here would not reject your signature or any of those I mentioned, though.

Doc Taxon (talkcontribs)

Oh, because I have been mentioned above to this topic, I thought, you're talking about tools like Ping or Replyto. I think it was a misunderstanding. Sorry!

Tacsipacsi (talkcontribs)

I think any tool should allow some closing HTML tags after the signature, as they might appear for several reasons: for example, when I post a side comment, I usually write it in small, including my signature ~~~~, or when someone reconsiders their opinion, they might strike through the old one ~~~~ and underline the new one ~~~~. In all these cases, the comment ends with a closing HTML tag, which doesn’t come from the preferences, by the way.

Pelagic (talkcontribs)

French Wikipedia's equivalent of sinebot wraps the date in a link to the revision. That could be reformatted to work with Discussion Tools, but @Tacsipacsi's use cases are something that hadn't occurred to me.

Ideally we'd have some explicit start-of-comment &/or end-of-comment markup, but whilst we rely on detecting timestamps, anything that doesn't end in Pelagic (talk) 21:54, 7 May 2020 (UTC) ~~~~ [good grief, I wasn't expecting Structured Discussions to translate the tildes!] has the potential to misbehave.

As @Matma Rex pointed out, sig. validation won't affect these, though.

Reply to "Do you have an example signature?"
156.57.13.33 (talkcontribs)

Signatures should be in a proper format. This is long overdue. ~~~~

Pine (talkcontribs)

I'm OK with the proposal, and I would explicitly give communities the option to enforce the requirement on existing signatures. My guess is that some communities would be willing to enforce the requirement on existing signatures after a grace period and after providing notifications to affected users who have noncompliant signatures.

Anomalocaris (talkcontribs)

The software should not allow any changes to a signature that leaves a signatures with any External links, line breaks, horizontal rule, unescaped pipe character, images, or lint error including font tags, or lacking a link to some user page. Users with signatures not in compliance with any of these should be given 60 days to fix their signature and after that the signature should be wiped and they get the default signature. There should be a link to a help desk where you can ask, among other things, "Please help me create a signature string that renders like my existing string but is in compliance with the new rules." On wipe-noncompliant-signature day, non-compliant signatures should be stored in an "Old Sig" field to help those on Wiki-breaks to develop a compliant signature when they return.

Whatamidoing (WMF) (talkcontribs)

Presumably, if you have a custom sig, you have also signed a comment. That means that you could find a copy of your custom sig in Special:MyContributions.

Pine (talkcontribs)
Reply to "Whole hearted support"

Support, but not sure about font

29
Headbomb (talkcontribs)

<font face="Helvetica">

is a lot shorter than

<span style="font-family: Helvetica">

Jonesey95 (talkcontribs)
Izno (talkcontribs)

I am on the same page as Jonesey. It is nice that it is short but browsers may stop supporting it at any time, in which case all those beautiful customized signatures will not be beautiful anymore.

AntiCompositeNumber (talkcontribs)

I definitely agree that <font> shouldn't be used anymore, but I'm also aware that there are probably still editors that are a) still using it, b) close to the character limit, and c) would run to the dramaboards because the evil WMF banned their personal signature out of spite. If anyone has any sort of data to say how prevalent font tags are in signatures, it would probably help guide this decision.

AntiCompositeNumber (talkcontribs)

Alright, I ran the numbers: there are 6705 enwiki editors with a font tag in their signature already. Some of them clearly haven't been changed in a while, since they predate the software-enforced 255 character limit. There are 1916 enwiki editors with a font tag and a signature longer than 200 characters. Nothing would break immediately for these editors if they did not change their signature, but many would be unable to make any edits to their signature without removing the existing formatting (because the span tags are too long).

Jonesey95 (talkcontribs)

Do you have a way of knowing how many of those 1,916 editors have made an edit in the last year? That would give us an idea of how many people may actually be affected by font tags being deprecated.

And FWIW, changing a font tag to a span tag if you are setting the font face, as in Headbomb's example above, adds 14 characters, but if you are just setting the color, it adds only 7 characters. The vast majority of editors should be fine, and a few might have to make do with a bit less customization.

Headbomb (talkcontribs)

It would probably also be a good idea to post suggested fixes to the signatures. For example a bot message like

Blah blah blah, linter error, html 5 blah blah blah, which is bad because of explosions, murder, and possibly jaywalking

  • Current signature: <b>[[User:Example]]</b>, which renders as User:Example
  • Suggested signature: '''[[User:Example]]''' which renders as User:Example

A message from FIX-YO-SIG Bot 00:41, 5 March 2020 (UTC)

Izno (talkcontribs)

I think I would tend toward a "see Project:Signature fix page for a list of common optimizations and feel free to ask on the talk page if you need help changing your signature."; even in your example, there might be context such as <span> elements with styling that could instead be turned into <b> elements with styling.

But, that's implementation details.

Headbomb (talkcontribs)

That's also a good potential workflow. Possibly with a bot posting basically a similar message and pointing them to that signature fixing page.

TheDJ (talkcontribs)

Well lint fixing experiences of the past show that you can't even write a bot for some of these things, so either we need a really smart AI, or we just have a page where some humans help out.

AntiCompositeNumber (talkcontribs)

363 editors have font tags and have made an edit since 2019-03-04T00:00:00Z.

Izno (talkcontribs)

I do not think 363 is really a sufficient quantity to support font, and eyeballing the majority of the cases pulled up in that query, I don't really think the signature lengths I'm seeing need it either given a number of optimizations I can see being made to most of those signatures. For the others, I don't think they should really be running around with such individuality as I'm seeing.

(But thank you for the tasty data.)

Headbomb (talkcontribs)

Also this editing interface is garbage. ~~~~

Whatamidoing (WMF) (talkcontribs)

This community prefers that people start in visual mode. The pencil icon in the corner by the "Reply" button will let you switch to wikitext mode (automatic sticky preference, so make it in your next reply, and it'll remember it until you change it).

AntiCompositeNumber (talkcontribs)

I'd definitely agree that prohibiting <font> tags will cause unnecessary complaints. The other requirements seem to be reasonable and enforced by policy and/or angry editors in many places already. The overall decision to deprecate and remove font tags as lint errors is probably outside the scope of this proposal though.

Headbomb (talkcontribs)

It's not only the familiarity of the font tags themselves, but also the length saving. The difference between font/span example above is a whopping 14 characters.

Everything else seems kosher.

Headbomb (talkcontribs)

It's really not evident in the least that a pencil icon switches between different editing interface.

AntiCompositeNumber (talkcontribs)

It is the standard button for it throughout OOUI.

Headbomb (talkcontribs)

I'd say it's equally confusing everywhere it's used then. Anyway, that's irrelevant to signatures, so let's stop here. Other fun fact about VE discussions, you can't collapse/archiving side discussions or re-indent them to make them follow one another. Yaaayyyy.

Whatamidoing (WMF) (talkcontribs)

(Technically, this is Flow, which happens to use a version of VisualEditor as one of its editing environments. I'm not sure that there's even a name for Flow's wikitext mode. I think that both desktop and mobile use the same tools in Flow, so you should only have two, rather than four, editing systems here.)

RexxS (talkcontribs)

I'd definitely get rid of <font>...</font> tags. All too often you get something like <font face="Helvetica">''' ... '''</font>, which translates to <b style="font-family:Helvetica"> ... </b>, which is only one character more. It's often possible to use style="font:Helvetica" if the other font property defaults are okay at that point, and that actually saves six characters net. When you get colour changes at the same point, you increase the savings, and because styles can be applied to <i>...</i>, <sup>...</sup>, <small>...</small>, etc. there are lots of opportunities for applying a style without adding a span. That very often results in a net decrease instead of a net increase in sig length.

Anomalocaris (talkcontribs)

Can you provide an example where style="font:Helvetica" actually works?

AntiCompositeNumber (talkcontribs)

It has to be "font-family:Helvetica", not "font:Helvetica". You probably won't see any difference with Helvetica anyway, since there's a good chance that it is or is very similar to your default sans-serif font. If we use "font-family:serif", like this, there is a clear effect.

Anomalocaris (talkcontribs)

In other words, when RexxS said,

It's often possible to use style="font:Helvetica" if the other font property defaults are okay at that point, and that actually saves six characters net.

this was entirely incorrect.

Izno (talkcontribs)
Anomalocaris (talkcontribs)

Izno: That shorthand requires a font size. Examples:

  • <font face="Times">The quick brown fox jumps over the lazy dog.</font>
The quick brown fox jumps over the lazy dog.
  • <span style="font:14px Times">The quick brown fox jumps over the lazy dog.</span>
The quick brown fox jumps over the lazy dog.
  • <span style="font: Times">The quick brown fox jumps over the lazy dog.</span>
The quick brown fox jumps over the lazy dog.

Without a font size, the markup doesn't work. In English Wikipedia, rather than 14px, I need 12.7px to make the lines the same, and although English Wikipedia uses a default font size=small, using "small" instead of 12.7 results in a slightly larger-than-default font size, unlike <span style="font-size:small">, which has no effect on size.

RexxS (talkcontribs)

Entirely incorrect - apart from all the documentation on the CSS font property, of course.

For example:

https://developer.mozilla.org/en-US/docs/Web/CSS/font

"The font CSS property is a shorthand for font-style, font-variant, font-weight, font-stretch, font-size, line-height, and font-family. Alternatively, it sets an element's font to a system font.

Nevertheless, as you say, it doesn't work if you only specify the name(s) of the font-family. Oddly enough, style="font:1em serif" actually works.

This is serif text

So I can only save 2 characters net (unless you also want to change the font size at that point). I have no idea why some combinations of properties work and others don't.

AntiCompositeNumber (talkcontribs)

because css

RexxS (talkcontribs)

Indeed

Reply to "Support, but not sure about font"

Only link to your own userspace

3
ToxiBoi (talkcontribs)

I don't know how hard this could be, or if I'm potentially late to the party, but I think we shouldn't be able to include links to someone else's userspace; since it will mostly be used for signature forgery or other bad-faith matters.

WereSpielChequers (talkcontribs)

I'm sure I have seen people link a signature from their alternate account to their main account. Plus when people change their username all their old signatures still link to their old name. As for signature forgery and other badfaith stuff, if and when that happens our admins can deal with it. No need for a software solution until and unless something becomes such an issue that a software solution is needed.

Jonathunder (talkcontribs)

Agree with WereSpielChequers. Some alternate accounts, such as bots, might be better linked to the main account than to themselves. ~~~~


Reply to "Only link to your own userspace"

Role of auto-assigned User ID?

6
Ceyockey (talkcontribs)

I think that the entire discussion around signature formatting could be set aside if the identification and reply functionality were based on the User ID. On my page https://en.wikipedia.org/w/index.php?title=User:Ceyockey&action=info my User ID is 150564 . It is my belief that this auto-generated ID is a) unique for every newly created account and b) independent of the signature. I would suggest that functionality be put in place that would embed the User ID into the signature, as hidden text, upon use of any of the 3, 4 or 5 tilde signature indicators. I am not familiar with the details of the wikimedia software, so my understanding of the feasibility of this is 0. Thanks for considering, nonetheless. --~~~~

Sunpriat (talkcontribs)

This id is different across all wikis. The nickname is the same in all wikis (upon request it is possible to make a different name for one of the wikis, for example, there is "nickname (xxwiki)", but there are not many such cases). Signing with an id on one wiki, on another wiki it will be the id of another person. When registering a nickname, software guarantees uniqueness and discrepancy with old nicknames. So id can be embarrassing, they are less croswiki universal and some will think with prejudice towards someone whose number is less.

Ceyockey (talkcontribs)

Maybe this is an area where WikiData can help: Use WikiData as a place to store the UserID-nickname matrix for each user. Is a UserID created for a person if they have never visited a wiki? For instance, I don't think I've ever visited the Thai wikipedia (th) and just went to https://th.wikipedia.org/w/index.php?title=User:Ceyockey&action=info which shows my user ID as 380,192; I confirmed after checking the ID that I have no contributions on that wiki. I then looked at the Statistics page and found that the number of registered users is listed as 378,272. This - by a looong stretch - leads me to speculate that my ID was not created until I first visited the site, assuming that the Statistics page is not updated in real-time and assuming that User IDs are created in sequence. So, a routine could be added that would append a user page in WikiData with a newly created User ID when they first visit a particular wiki. Thoughts on this noodling?

Ceyockey (talkcontribs)

Oh -- and I just took a look and I got a Notification that a bot has just put a new user welcome on my talk page on the Thai wiki. So there are already routines in place which could be exploited as I noted above.

Tacsipacsi (talkcontribs)

Wikidata is a freely editable database of mostly data interesting for the public. The user ID–user name connection is not interesting for the public, and certainly not freely editable (what if I change the enwiki entry for Jimbo Wales to point to my user ID?). By the way, I don’t understand why is it worth using the cryptic user ID instead of the user name. The user name can be added as a hidden text just like the user ID, and software will be unable to understand it if the wikitext is so broken, just like if the user ID was there. (And the link requirement is not only for software’s benefit, but also for users’—if I want to contact the comment’s author, I’d like to click a link and get there, not search in page history or—in case of your proposal—comment wikitext.)

Matma Rex (talkcontribs)

There have been similar ideas in the past (most recently: T230653). This would be possible to do, but:

  • it would be a change to the syntax of all signatures and so we're a bit wary of making it
  • it isn't clear what should be embedded (as you've already realized here) and it would be difficult to change in the future, since the results are saved as wikitext

We might come back to this at some point, but we're not planning to work on it at the moment.

Reply to "Role of auto-assigned User ID?"
Whatamidoing (WMF) (talkcontribs)

It's the end of March. I think the Editing team has all the information they need for this stage, but they are not in a hurry. If you've got comments/questions/ideas, please keep posting. Expect to see a bit more (e.g., perhaps a timeline) in the next few weeks.

On the community side, I think the main tasks will be:

  • updating any help pages (if your community has a page saying that you can do something that will no longer work), and
  • asking editors with soon-to-be-invalid signatures to update them. In some wikis, this is just a couple of people, and in others, we'll want to use Special:MassMessage.

Also, Thank you. There's a lot of good information in these discussions, and I really appreciate it.

Spinningspark (talkcontribs)

I've only just been notified about this (through the WP Administrators' Newsletter) today, 1st April. I see comments are required before 31st March. How was that meant to work?

Whatamidoing (WMF) (talkcontribs)

This was announced in Tech News, and on your home wiki, at the Village pump (technical) and the bot operators' noticeboard. If you are interested in technical matters, including software changes that affect non-technical users, I recommend subscribing to (and then actually readinng) Tech News each week. I appreciate the help of anyone who decided to share the links further, but naturally the WMF doesn't control their publication schedule.

As for the stated deadline, it was an estimate of when the devs would actually be able to work on this. As you can see from my earlier comment, they're busy with other things this week, so there's no need to stop providing information just because of the earlier estimate. The "real" deadline is a few seconds before they start coding.

Spinningspark (talkcontribs)

This is not a technical matter. It potentially affects all users. The issue may have a technical solution, but that is no reason to hide it in tech related venues. I'm not going to sign up to tech newsletters just in case something comes up. When you are at the stage of calling for comments, a watchlist notice would be appropriate.

Whatamidoing (WMF) (talkcontribs)

Solutions to almost all technical problems have the potential to affect all users, e.g., by making it harder, easier, faster, or slower to use the sites. The problems addressed here are:

  1. invalid HTML,
  2. accurate machine detection of signatures, and
  3. how to parse signatures that contain certain characters (i.e., how to turn what's stored in the database into HTML that your web browser can understand).

These all sound like technical problems to me.

Reply to "Status update"

Disable "fancy" signatures entirely

8
ToBeFree (talkcontribs)
AntiCompositeNumber (talkcontribs)

I wish.

Pine (talkcontribs)
Stjn (talkcontribs)

On the other hand, ‘diversity of user signatures’ makes our wikis less welcoming to people with colour blindness and disabilities that require using a screen reader to edit (since, I imagine, they have to go through a lot of useless code to know, say, who to ping in a discussion). You have to wonder what kind of diversity we should value.

Stjn (talkcontribs)
Spinningspark (talkcontribs)

Imagine that a postal service started demanding we sign our letters in a particular way. Don't be so restrictive. It is unnecessary and condescending to volunteers

ToBeFree (talkcontribs)

No imagination needed: Emails sent via Wikipedia are text-only

Spinningspark (talkcontribs)
Reply to "Disable "fancy" signatures entirely"

Support link, oppose linter

8
Alsee (talkcontribs)

First I want to note that I'm not a fan of putting this on display in the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying Beware of the Leopard. I only saw it because it accidentally turned up in the middle of a list of Phab search results.

The link requirement is helpful to users, the error message would be clear and simple, and it's surely in line with policy or practice at probably every wiki.

The linter portion is not remotely supported by the why rationale offered for the proposal. It would pointlessly disrupt users with confusing error messages on perfectly valid wikitext. It would be insane to suggest things like <small><sup>this</small></sup> could or would ever be banned. @Whatamidoing (WMF) Please ping me if linter is dropped from the signature proposal. In that case I can skip incorporating this signatures item in a linter-related discussion planned for the new EnWiki Village Pump page.

AntiCompositeNumber (talkcontribs)

@Alsee That example isn't perfectly valid wikitext though. It's a side effect of a old, broken parsing system that is being replaced, and shouldn't be used. If you've got a suggestion about where this proposal should be further publicized, I'm sure we'd all be glad to hear it.

Alsee (talkcontribs)

@AntiCompositeNumber your argument is completely unrelated to the rationale given for this proposal.

Instead of trying to push through the linter issue under a deceptive rationale, how about we allow the Require a link to user page, talk page or contributions portion sail through uncontroversially. Then you or anyone else can make a separate linter proposal with the authentic rationale. Then I can more neatly wrap the linter topic inside a discussion I plan to open on the Foundation's broader strategy around VisualEditor. There has been too much conflict between the Foundation and the community, too many failed products, and it primarily traces back to the strategy around VisualEditor. The Foundation and the community need to get in better alignment on our broader strategy, and one possible outcome is that the rationale you just gave for linter vanishes.

Kaldari (talkcontribs)

I don't see what this proposal has to do with VisualEditor. Regardless of whether the WMF uses VE or Parsoid for talk page replies, they will still need to be able to reliably detect signatures in order to reliably add reply buttons, which is impossible right now. Requiring signatures to use sane HTML makes detection much easier. Of course it also has lots of other nice side-effects like making signatures render consistently regardless of HTML doctype and making it easier to migrate to Parsoid (or any sort of modern DOM-based parser), but that's not really the point.

Alsee (talkcontribs)

Setting aside the other issues, you can't actually impose any constraints at all on signatures. Anyone can type anything as a signature - or even automate it with a userscript. -- Main (talk) 25:68, 44 March 4130 (UTC)

AntiCompositeNumber (talkcontribs)

Sure -- and people can just not sign their messages at all. That's like saying Wikipedia shouldn't have verifiability standards because anyone can edit.

Matma Rex (talkcontribs)
Spinningspark (talkcontribs)

In general, I'm in favour of grandfathering in signatures when the rules change. However, there really is no excuse for having linter errors in a sig, and I'm speaking as someone who had errors in my sig for years until it was pointed out. They might still be there in wikis I visit only rarely.


Unclosed bolding, italics, small etc can cause enormous problems on pages that are transclusions of many other discussions, messing up the entire page. Examples are Wikipedia's DYK and AFD log pages. It is annoyingly time consuming to track down which discussion caused the problem and which post in that discussion, let alone find the actual error.

Reply to "Support link, oppose linter"
WereSpielChequers (talkcontribs)

The WMF is conceding that there are a few signatures that would be non compliant, but has not quantified this, notified those whose signatures would need to change, or told those individuals what change they might need to make. Most wikipedians are reasonable. I believe that many would simply change their signature if they received a polite request "Hi, the developers are looking to change the signature rules slightly. If this goes ahead you would need to change your signature from "this bunch of text" to "this bunch of text", there is a discussion "here". The rest of us would probably relax if we knew that it was being handled this way.

AntiCompositeNumber (talkcontribs)

By my count, about 8200 enwiki editors with edits in the last year. Most of those editors are new contributors who put plain text into the signature box and clicked "Treat this as wikitext" without reading the warnings around it. The original proposal was to leave these signatures in place, but there has been (at least initial) support for a process of depreciation and removal.

Matma Rex (talkcontribs)

To be clear, with the plans we have right now, no one will need to make any changes to their signature. Even if your existing signature would be invalid under the new rules, it will still be allowed, and you'll only need to make changes if you're trying to change your signature anyway.

We're considering doing something about existing signatures (see the discussion other sections on this page), but right now we have no plans, and if we end up doing anything, we should notify the individual users as you suggest.

Whatamidoing (WMF) (talkcontribs)

Even though there is no plan to invalidate signatures yet, I think that some users would like to be able to check whether theirs (or their friends') are "on the list". When Tidy was removed, I remember that @Anomalocaris talked to a lot of individual editors about fixing their signatures, and I think most people were cooperative about it.

@AntiCompositeNumber, do you think it would be worth turning that into an editor-friendly list with sections like "These people should untick that button" or "These people should add a link", and post it on the relevant wiki?

AntiCompositeNumber (talkcontribs)

A tool to check a single person's signature wouldn't be very difficult. I'll work on that.

Anomalocaris (talkcontribs)

I started my effort to ask enwiki editors to fix their signatures on October 15, 2017. I tracked this effort in multiple spreadsheets, and there may be some names that appear in more than one. I estimate that by now I have made signature requests of about 990 editors. In descending order, the four most common issues were (1) obsolete font tags; (2) Tidy bug affecting font tags wrapping links, now known as Old behaviour of link-wrapping font tags; (3) enlargement with font size or big markup; (4) other lint issues including misnested tags, missing end tags, other obsolete HTML, and stripped tags. In no particular order, there were also (5) use of subst to exceed the 255-character limit; (6) excessive text shadow; (7) use of the unescaped pipe character; (8) signatures with images. I never observed a signature that lacked a link to user page, talk page or contributions. I never observed a signature with nested substitutions.

About 1% refused to comply. About 9% ignored my request and continued to use their non-compliant signature. About 15% may have fixed their signatures, but I never saw and recorded that they used their signature after my request. About 75% I recorded as having fixed their signatures.

"We're looking for feedback as to whether you would like existing invalid signatures to be disallowed. If invalid signatures are disallowed, the default signature would be inserted when affected users sign their comments, until they correct their personalized signatures." Yes, let's notify users that this change is coming, and then use default signatures in place of existing invalid signatures. Is this the right place for me to put this comment?

Whatamidoing (WMF) (talkcontribs)

> Is this the right place for me to put this comment?

Yes. Pretty much any place on this page is the right place to put your comments. MediaWiki.org isn't a bureaucratic place. :-)

PerfektesChaos (talkcontribs)

@AntiCompositeNumber You might look closer to the time of the most recent edit or editcount when informing projects about a list of problematic signatures.

  • I took your analysis and did inform the 8 most seriously affected users at dewiki.
    • These were missing end tag and 1 nested-subst.
    • The nested-subst has been an experiment by a user who made one article edit and never showing up again, obviously just playing around.
    • 4 users are editing on a daily base, they did change the signature immediately.
    • 4 users made their last edit 2, 5, 8, 10 month ago, some with lower total count.
    • All were actually misnested and remedied anyway by REMEX (tidy) cleanup. They did not harm to subsequent contributions.
  • In general it is a very good idea to feed a short list to communities in order to support people in fixing and explaining in their own language and by local techies.
    • However, the list should be as short as possible to keep the task feasible.
    • I would limit myself to 1 or 3 months; who is off for a longer period does not sign and will not respond on talk page notifications.
    • Only most disturbing effects should be listed (non-empty without link, closing tag missing, subst etc.).
    • You may provide a JSON list offline with complete data, but many projects might be unable to derive conclusions from that.
  • Decorative issues, styling, obsoleted HTML which will work for decades should not be mentioned in first place project information. It is up to projects to check that with their local guidelines.
  • If someone is using an inappropriate signature, other users will note that and start reminding and demanding earlier or later.
    • Affecting subsequent contributions is not obvious to the user itself, but will be observed by the next editor at the last section of thread. Here it is really helpful to get early notification and rejecting storing that preference.
AntiCompositeNumber (talkcontribs)

Hey, I don't write the requirements, I'm just trying to help everyone make a more informed decision. I do agree that most of the users in the initial lists were just experimenting and have now left. A potential bot making notifications would be in a better position to evaluate the activity of users before notifying, as total run time is less of a concern.

Anyway, I put https://tools.wmflabs.org/signatures together. It allows you to check a single user's signature (with basic explanations of the problems) or view a nicer form of the aggregate listings. Those lists now only cover editors with an edit in the last 30 days. Remember that this tool still only shows preliminary findings, and may not totally represent what future MediaWiki developers may implement. (This is especially the case for nested substitution, as I don't have a reliable way of implementing it without many false positives or false negatives). PRs/suggestions/bug reports are of course welcome.