UDWiki talk:Administration/Policy Discussion/Signature Compatibility

It doesn't look like that anymore on IE. If that happens, it means you're not using the latest version of IE, but rather an older version. --•▬ ▬••▬ • •••• •▬ ▬•▬• ▬•▬ ^{#nerftemplatedsigs} 23:46, 14 October 2010 (BST)

Here's an example of what it looks like now in the latest version of IE: [1]

Note Mis's signature and the fact that the screenshot is of IE. --•▬ ▬••▬ • •••• •▬ ▬•▬• ▬•▬ ^{#nerftemplatedsigs} 23:50, 14 October 2010 (BST)

I really had the idea certain wiki code did not work on Explorer, even the latest versions. How does Aichon's userpage look like?--Thadeous Oakley 23:59, 14 October 2010 (BST)

Still looks fine. --•▬ ▬••▬ • •••• •▬ ▬•▬• ▬•▬ ^{#nerftemplatedsigs} 00:00, 15 October 2010 (BST)

It looks fine but the code doesn't work in terms of linking to the user.--Yonnua Koponen ^{Talk ! Contribs} 15:57, 15 October 2010 (BST)

I actually "fixed" (i.e. added unnecessary code to) my userpage to make it not look like crap in IE a month or two back, since IE lacks support for some very simple stuff (e.g. they don't even support rgba...what the hell, Microsoft?). At this point, my page should be mostly together in IE7+ and shouldn't have random shapes and text flying all over the place. I didn't check it in IE6, but I assume it still works okay-ish. And certain wiki code/HTML/CSS does indeed not work in even the latest versions of IE (hence why there are an abundance of hacks out there that web developers use to try and make things look halfway decent in IE when every other browser supports the features natively), but Rooster and I (mostly Rooster) finagled and massaged it and then beat it with a stick after shoving a load more code in until it worked nicely with at least IE8 and didn't break the page for IE6 and IE7. —Aichon— 18:12, 16 October 2010 (BST)

Rather than "for internet explorer" why not make it all-friendly and make it "for all browsers"? I mean, most browsers are better than IE, but I think it's more well-rounded that way. Also, I really like this policy because those signatures are ridiculously annoying and essentially bypass the "you must contain a link to your userspace" rule if the viewer is on IE.--Yonnua Koponen ^{Talk ! Contribs} 23:55, 14 October 2010 (BST)

Hey, I use Firefox myself, the reason I mentioned IE and not, for example, Opera or Safari is that IE is the most popular one out there. But yeah, that works too. --Thadeous Oakley 23:59, 14 October 2010 (BST)

Seems a moot point now given Axe's screenie. What versions of IE do and don't work? 00:11, 15 October 2010 (BST)

I agree with Yonnua, we should spread this to all browsers, or at least the main ones: Chrome, Safari, IE and FF. If something breaks a page on any of those it's worth forcing them to change it. -- LEMON #1 00:42, 15 October 2010 (BST)

Perhaps the ones set out by the European Commission? I think it's mainly those you've listed, but they're alledgedly the general ones. If there's anything barmy on there, then we probably shouldn't, but at least DDR's four should be included.--Yonnua Koponen ^{Talk ! Contribs} 08:28, 15 October 2010 (BST)

Thanks to Iscariot I've always been sketchy on relying on the governing by a national polical system tbh, but as long as everyone can agree on something, I'd like to think those four would be given as a minimum. -- LEMON #1 11:54, 15 October 2010 (BST)

Right, other than the four above the EC give: Opera, iron, flock, Maxthon, K-meleon, Lunascape, Avant and Flashpeak. I've heard of maybe one of these? Does anybody here use any of these, or even know anybody who does?--Yonnua Koponen ^{Talk ! Contribs} 15:30, 15 October 2010 (BST)

Much like most of the Commission's work, they're probably insignificant or fictional. 15:53, 15 October 2010 (BST)

I use Opera (with Chrome as back-up) and I've never experienced any signature problems. ~~ Chief Seagull ~~ ^talk 15:57, 15 October 2010 (BST)

Thadeous Oakley said:

What this means in practice is that people who use a different browser, like Firefox or Chrome, should always check whether or not their signature does not break when viewed through Internet Explorer.

I don't have IE on my machine. How do I check?--~ 16:22, 15 October 2010 (BST)

It's fine (I'm using an old IE right now). A general rule for the kind of breaking in Mis' sig (as well as AH and DDR's old ones) is if you post on my talk page and the page breaks massively, it doesn't work.--Yonnua Koponen ^{Talk ! Contribs} 16:39, 15 October 2010 (BST)

If there is really a way to check signature compatibility just by posting on a talk page then why not dedicate a page soley to that purpose? I really don't want to spam people and I REALLY don't want to use IE. --~ 18:59, 15 October 2010 (BST)

Well, my talk page (and the code on it) doesn't test for all issues of the code. It's just that the code destroys my talk, and for a different reason, it also doesn't work on earlier IE. I'm sure if somebody was unable to read your sig, they'd tell you about it. :P --Yonnua Koponen ^{Talk ! Contribs} 10:16, 16 October 2010 (BST)

Already Fixed

See this discussion. - Me and Aichon had a discussion on this whilst we were trying to hack together a fix to the Click template. Some people nicked the code halfway though which fixed the template except for the problems in IE. Long story short, it's fixed and works on IE7 or later IIRC. (as well as the other common browsers) If people still need links on images, either do the image redirect thing or use the new template {{ClickInternal}}. If somebody else's is broken, arby them for general breakage since they have no reason not to fix it. --  RoosterDragon  17:14, 15 October 2010 (BST)

I'm using IE7, and it does not work.--Thadeous Oakley 17:56, 15 October 2010 (BST)

Guess I remembered wrong... actually wait. Does it not work, but also not break? Maybe I meant that. Also get IE8. --  RoosterDragon  18:39, 15 October 2010 (BST)

Check the link on the project page. That's IE7. Also, I use firefox myself, but not all people do :(. Also, while were busythis --Thadeous Oakley 19:01, 15 October 2010 (BST)

Frankly I don't care about the people on IE6. IE7 isn't as bad but you're still 4 years behind the curve. Even worse once IE9 is out. This discussion as a whole is a waste of time. We're talking broken code on really old browsers that most wiki users don't use. IE8 and especially IE9 handle stuff just as good as other browsers do. We're only here because a few users grabbed code from a sandbox and decided it stable enough to apply to signatures. Not the smartest move ever, but once those few fix it it's basically like this never happened and we wouldn't be here wasted lengthy paragraphs on stuff our wiki shouldn't really be worrying over. I'm struggling to imagine how this would've ever come about had not the very specific set of circumstances that lead us here occurred. I can't actually imagine this ever coming up again, actually. --  RoosterDragon  20:16, 21 October 2010 (BST)

Rooster, although this came about because of that problem, it's in no way limited to that. It deals with any issue whereby somebody's signature breaks in a major browser. You shouldn't look at it just in the scope of the one problem.--Yonnua Koponen ^{Talk ! Contribs} 20:31, 21 October 2010 (BST)

If you can provide me a list of past instances where somebody's sig has broken in one major browser, but not others, I might agree. Then given that modern browsers are getting better at fixing their eccentricities and only the older browsers which continue to fall out of use are a possible problem, explain how likely it is that we need more red tape and something as fucking annoying as doing a cross browser check of your sig. Nobody will do it. Barely anybody has the inclination or the reason to check their general code in more than one browser, most people don't even have more than one browser for god's sake. If they did we wouldn't have -moz-border-radius hardcoded on every single thing that has ever had a border. If there's an error people will suggest a fix, or decide it's too small to care about. We don't need policy to tell them to use common sense and then make them waste time in other browsers or submitting VB cases.

I'm just annoyed that events lined up in the most unlikely of ways to the point where we're wasting time on the wiki's most red-tape laden discussion format, and discussing ways to add more rules that will have literally no use nor effect. --  RoosterDragon  21:00, 21 October 2010 (BST)

I don't see this as people having to cross-check browsers (may be wrong if the policy actually says that). I assumed it was referring to the current system, whereby if it was screwing up, somebody would tell you and you'd have 7 days to sort it out.--Yonnua Koponen ^{Talk ! Contribs} 21:06, 21 October 2010 (BST)

At which point you'd either need to get another browser and check it yourself or rely on friendly outsider help to fix it. Said help would usually show up without the need to threaten VB, and in either case putting in effort to fix it in outdated software that nobody in their right mind should be using is still pointless to me. --  RoosterDragon  21:17, 21 October 2010 (BST)

I'm just putting my thoughts here

Since I'm kinda the one responsible for creating the code used in those sigs, I figure I should add my 2 cents. And since the conversations above are all over the place, I'd rather just put them in one spot, rather than breaking them into 5-6 comments. I have a tl;dr summary at the bottom of my comment if you want to skip this stuff.

1. First, let's define a few terms so we can speak more easily. "Broken" means that it breaks policy by showing up on multiple lines (e.g. Mis' and DDR's sigs in the image that Thad provided). "Not working" means that it shows up fine (i.e. one line), but clicking on the images takes you to the page for the image (you may be wiki-redirected from there).
2. As Rooster said, the problem should be fixed now (i.e. not broken in any major browser, and working properly in IE8+ and other major browser), assuming that the people using those signatures update to the most recent code, rather than using the old code.
3. In IE6 and IE7, the current version of the code does not work, but it also does not break. If Thad, or anyone else thinks this statement is incorrect, please point me to the relevant still-broken signatures, since I am not aware of them (unless Mis is still using the old code, in which case, shame on him).
4. People using the code in their sigs are supposed to also use image redirects as a fallback method for older browsers that don't support the code. Image redirects have long been permitted in signatures (e.g. Iscariot's Cylon signature), so this should be a non-issue. This fact is explained on the {{Click}} template's page even, though I suppose it could be made more explicit.
5. Mandating support for "Internet Explorer" specifically, or any other specific browser, is ludicrous. What version and why that one? What if IE is still as ass-backwards and broken next year or in five years as it is now? Why not mandate support for mobile browsers while we're at it? It's a reactionary policy with long-lasting consequences to a temporary problem that already has a solution. Also, just because most Internet users use IE does not mean that most UD players use it. Anecdotally, I'd actually say that Firefox is the most used by meta-gamers in UD, who are the most likely ones to be using the wiki. Oh, and on top of that, some of us don't even have IE, so it might even be impossible to test for it for certain users (e.g. Mac and Linux users).

I'm against the policy in general, since broken signatures are already covered under the current policy. This is merely a question of enforcement, not new rules being necessary. Also, if we want to cite browser statistics, let's look at something not outdated by a year and a half. For instance, look at this or this.

Anyway, for the short version of everything I said, there's no need for a new policy since the old one already covers any of these problems, IE sucks, the issue with the code should already be fixed, people using the old code should be shot/warned of sig violations, IE still sucks, everyone using the new code in their sigs must use image redirects as a fallback or else be warned of sig violations, I'm not aware of the new code being broken in any major browser from IE6 onwards, and IE still sucks. Badly. Does this handle everything? —Aichon— 18:04, 16 October 2010 (BST)

Oh, one additional point: none of that addresses the issues these sigs create on Yonn's talk page (that is, that in certain browsers using these sigs creates a lot of blank space at the bottom of his talk page due to the way that scroll boxes work). Personally, I'm inclined to think that even though they do have issues on his page, that the problem is his to deal with, just as he has been doing, since it's more a case of his page breaking the sigs than the sigs breaking his page. In fact...looking through some documentation, I think I just figured out a fix for his page that lets those sigs work on it, which just reinforces my point. Off to his talk page! —Aichon— 18:44, 16 October 2010 (BST)

I just used my talk as an example because it correlated with the sigs breaking in earlier IEs. :P Also, I feel that by qualifyign the terms "break" and "work" in the way that you did, you've made the argument massively biased. In my opinion, a signature is broken if it doesn't link as it should do. I'm not going to change my terminology just because doing so sounds better for your side of the argument. I'd also disagree that this policy is reactionary. Reactionary means a policy with conservative elements reversing or halting reform. I see no way in which this policy is reactionary. Unless you meant: it reacts to things, in which case I agree. It's reactign to a problem on the wiki with a simple solution. I'm afraid I'll have to disagree with you on this one Aichon.--Yonnua Koponen ^{Talk ! Contribs} 19:48, 16 October 2010 (BST)

Change "reform" to "progress" and your definition of reactionary is perfect for this. :/ 20:07, 16 October 2010 (BST)

Reform and progress are massively different things.--Yonnua Koponen ^{Talk ! Contribs} 20:14, 16 October 2010 (BST)

Yes, but the "conservative elements reversing or halting X" is apt. As Aich's pointed out, there's nothing borked any more, so curtailing a novel approach to things like sigs because of outdated and replaced software not working with it is backwards. 20:21, 16 October 2010 (BST)

Sorry, I should have been clearer with my definitions, since it was not my intent to bias anything. I wasn't talking about the signatures being broken or not working (which is a nonsense distinction, as you rightly point out). I was talking about the code being broken or not working. These are terms that Rooster and I have been using for months in our talks to refer to the code in a technical manner, as the conversation he linked clearly shows. I just wanted to get everyone else on board with using the correct definitions to describe the various problems we're describing, that way we didn't argue over something we actually agree on (and I think you and I agree mostly, though you don't seem to have realized it yet :P).

Anyway, if you apply those terms to the signatures themselves, then yes, of course you're correct, since a sig with code that is not working is almost always a broken signature (with one important exception). The reason why we need the distinction between the code being broken or not working is because we have four possible scenarios that need to be dealt with, and we can't distinguish between #1-3 without those definitions:

1. The code is broken, so the signature is broken.
2. The code is not working (i.e. it does nothing noticeable), and image redirects are not used, so the signature is broken too.
3. The code is not working (i.e. it does nothing noticeable), but image redirects are used, so the signature is acceptable.
4. The code is working, so the signature is fine.

I'm saying that #1 is already wrong according to the existing policies. Nobody should be running into #1 however, unless they are using the old code in their sigs, which they should not be doing. I am also saying that #2 is against existing policies, since we know that the code does not work in IE6 and IE7, and image-only sigs with no redirects are already against policy. So, basically, if we run into #1 or #2, we just need to enforce existing policy to handle the problem.

Where I draw the line, however, is with #3, which I believe is perfectly fine, since it behaves the same as an image-only sig using image redirects (e.g. Iscariot's sig). And #4 should be fine as well, since it breaks none of the existing policies, nor does it show up oddly for any of the major browsers.

As for "reactionary", thanks for catching that. I misused the term (as a quick check of the dictionary would have shown me if I had looked). What I meant is that it's a knee-jerk reaction to the situation (i.e. your second interpretation of what I meant). And yes, it's reacting to a problem on the wiki, but the solution is unnecessary, since the wiki already has policies in place to handle this situation. We just need people to enforce the existing rules, not create additional rules that complicate things unnecessarily. —Aichon— 20:39, 16 October 2010 (BST)

The reactionary thing is much a case of people stupidly naming a word. I only know because I made the same mistake the other week in history. :P I'd also agree with your analysis of which of the above points constitutes a broken sig, but I'd say we do need this policy, because it means that we codify what we already have to clarify another circumstance which sigs need to abide - they need to work for everyone (or the majority). This doesn't just cover click, it covers any future issue where one browser is at a disadvantage.--Yonnua Koponen ^{Talk ! Contribs} 21:24, 16 October 2010 (BST)

And that's where you and I disagree, apparently over what policies are meant to be. While you say that we need a policy to "codify what we already have to clarify another circumstance which sigs need to abide", I say that that is the job of precedents. We always use precedents to clarify future situations, not policy. Policy is used to change the rules, precedents are used to clarify them, and this, as you say, is a case of clarification. Otherwise, if we did make a policy of it, you either need to name a specific browser or else use a vague phrase such as "all major browsers" or "most people". In the former case, the problem is that the policy will be outdated and a burden in short order. In the latter, it adds nothing to the spirit of the policy or what we can enforce, while complicating the policy with loopholes and room for lawyering (e.g. you have to define "major browser" or "most people" otherwise someone will worm their way around the rules, and defining them opens up even more problems). Either of those situations is already covered by the current policy as a result of precedent that exists or will exist thanks to future rulings. —Aichon— 22:59, 16 October 2010 (BST)

I completely disagree with that. Case Law / Precedent's job is to interpret policy. If it does, it can become biased and unfair, especially when one is decided by 7 or 8 people and the other is decided by an actual embodiment of the community. This can be counteracted by setting forth what rulings should be made in policy. In this case, that would be as simple as establishing what browsers should be affected (see above for an assumption of IE, FF, Chrome and Safari, maybe opera) and then adding a line saying they should be considered. Saying that we shouldn't make a policy on something because it's handled by precedent is stupid. In an ideal world, all law should be codified, and I don't see why this should be any different. The ultimate difference being this - anybody is free to divulge from precedent (unless we enter the stupid assumption where we say precedent says we can't break from precedent) if they're the highest authority. Since we only have one authority, it makes absolute sense to codify any rules possible and prevent any degree of corruption possible.--Yonnua Koponen ^{Talk ! Contribs} 10:45, 21 October 2010 (BST)

Who's Signitures are at fault?

I'm trying to figure out who's sigs this is aimed at. Like, is mine ok? -Dezonus- ^(talk) 22:40, 16 October 2010 (BST)

Yours is fine. If you want an example of what's wrong see User:Misanthropy/Sig - It uses the floating link code without the underlying page redirects for the images. You don't use the code, so you're fine.--Yonnua Koponen ^{Talk ! Contribs} 22:43, 16 October 2010 (BST)

It's aimed at signatures that have an image redirect. For example...

 

. Note that clicking the image will redirect to my User page instead of the image page. --•▬ ▬••▬ • •••• •▬ ▬•▬• ▬•▬ ^{#nerftemplatedsigs} 22:47, 16 October 2010 (BST)

Except it doesn't.--Yonnua Koponen ^{Talk ! Contribs} 22:50, 16 October 2010 (BST)

What? It doesn't redirect? It seems to redirect for me in both IE and Firefox. --•▬ ▬••▬ • •••• •▬ ▬•▬• ▬•▬ ^{#nerftemplatedsigs} 22:54, 16 October 2010 (BST)

You need to do actual image redirects (e.g. #REDIRECT User:Axe Hack) on the image pages, Axe, for earlier versions of IE. The fancy code only works in IE8+ and the other major browsers. —Aichon— 22:59, 16 October 2010 (BST)

I have IE7 (Code doesn't work) and IE6 (Splits on to multiple lines) as well as Safari (Everything works), so I'm stuck with the annoying situation that I know what everything should look like, it just always looks wrong. :P --Yonnua Koponen ^{Talk ! Contribs} 23:15, 16 October 2010 (BST)

IE6 does split to multiple lines? Drat. I had hoped we had fixed that with the changes made for IE7. I can't test IE6 myself, since I don't have access to a copy of it at all, but I can probably toss in some more changes to the code to make it play nice with IE6 (i.e. make it do nothing at all in IE6). That said, IE6 market share is down below 4% last I checked, and declining rapidly (plus, most of it is attributable to legacy work computers, not home users), so the point may be moot. If you think it's worth looking into though, I'd be willing to do it for you, Yonn. :) —Aichon— 00:00, 17 October 2010 (BST)

I think it's IE6. Could be IE5, but I doubt it. It's really not a bother, considering that computer's completely fucked anyway. Youtube is stuck on Chinese, I can't watch any media content and adobe deleted itself. :P I'm getting a laptop in January, and then I'll have a better browser than all of you combined.--Yonnua Koponen ^{Talk ! Contribs} 00:03, 17 October 2010 (BST)