Talk:Turning off outdated skins

Resources

Saying "because of our limited resources" sounds like sheer mockery, as if WMF would be hard-pressed for money and desperately in need of removing the extremely expensive skins. --MF-W 21:15, 28 March 2013 (UTC)

What the text is about is the cost of the time and energy of paid developers doing support and maintenance based on bugs that are specific to skins. When you add up the hours potentially spent by designers and developers bit by bit supporting compatibility with certain skins or browsers etc., that can in fact add up to thousands of donor dollars. The question begged by it all, and part of the core justification for the change, is whether it is really worth spending donor dollars on supporting skins used by a relatively tiny number of people? If we're not going to really support those skins in MediaWiki, then leaving them up creates an unfair situation where "official" skins listed in preferences are not in reality supported. Perhaps it could be improved in the text if explicitly mention that we're talking about paid employee time supporting skins? Steven Walling (WMF) • talk 22:54, 28 March 2013 (UTC)

Steven is correct, here; resources refers to more than just 'money' (although indirectly it does add up to that). We have a limited number of paid developer and designer hours to build things. If we want to support, instead of four skins, or two skins (Monobook and Vector) nine skins, with vastly different levels of technological complexity and vastly different ways of doing things, we have to choose between delivering incredibly limited software or delivering software that is not only more limited than necessary, but also late. Okeyes (WMF) (talk) 23:49, 28 March 2013 (UTC)

Thanks for clarifying. I understood the reason for stopping providing the 4 skins as outlined in the page, but didn't understand really how the first sentence could be brought in line with that. --MF-W 17:03, 29 March 2013 (UTC)

What if you leave the other skins (especially the "minimal" rather than "alternative") ones but have them marked "unsupported"? That removes the support cost on maintainers but doesn't leave users in the lurch. Vsync (talk) 18:07, 2 April 2013 (UTC)

Except we're implicitly saying "this is fine to use!" by making them available, and setting the expectation that we will support them. Are you telling me that if we mark them as 'unsupported', editors will all abrogate their rights to oppose deploying $extension if they've got outdated skins? I somehow suspect it would cause more drama than it solves for. Okeyes (WMF) (talk) 18:11, 2 April 2013 (UTC)

Numbers

A couple interesting points:

1. Almost 3 millions with monobook in preferences. I don't think this figure realistic and I bet you'd have to divide this by dozens or even hundreds to have "real users": most of them are probably from Special:PrefSwitch that automatically switched you back to monobook on all wikis if you wanted; it almost always failed above a few dozens/couple hundreds wikis, but did its job for many.
2. Subpage says «our conclusion is that Chick, Classic, MySkin, Nostalgia, Simple and Modern are so rarely used as to only questionably justify support»: this doesn't mention Standard, that is used more than Cologne Blue and represents the 70 % of power users affected by the switch off (probably all of them using it since before monobook existed, i.e. some ten years?).

--Nemo 22:26, 31 March 2013 (UTC)

1. Standard and Classic are the same skin. Okeyes (WMF) (talk) 10:08, 1 April 2013 (UTC)

   Right. And I even know that, just thinking with half neuron; will now fix the text. --Nemo 12:38, 1 April 2013 (UTC)

Suggestions for "Classic" user?

Does anyone have ideas as to which on the new approved list would be the easiest transition? I've used Vector every now and then, but I really dislike the grey background, and the fact that it's loaded down with so much scripting (more than my browser can easily handle sometimes)... AnonMoos (talk) 15:50, 2 April 2013 (UTC)

Classic is the only one whose appearance I find tolerable. The rest look awful. This is a stupid decision (although I'm not going to threaten to quit the project over this, as that would also be stupid). DS (talk) 16:08, 2 April 2013 (UTC)

It's a perfectly rational decision; if you can put together an argument about why we should reconsider, I'm happy to listen.

AnonMoos: Personally, I use monobook; tried it? Okeyes (WMF) (talk) 16:16, 2 April 2013 (UTC)

You could also try CologneBlue or Monobook with special CSS... πr² (t • c) 17:24, 2 April 2013 (UTC)

I've been told that some features can be duplicated in .css. If these features -- for instance, the use of serif fonts, or the 'toolbox' links being located in a bar off to the left side of the page instead of up top, or the pale yellow background on all non-article pages -- can be replicated, then I might withdraw my objections. However, I've spent my time on Wikipedia working on articles and patrolling newpages for inappropriate content, so I don't know how to edit CSS myself. DS (talk) 17:35, 2 April 2013 (UTC)

Try cologne blue, Cologne blue? πr² (t • c) 17:45, 2 April 2013 (UTC)

nostalgia.wikipedia.org

Will this affect nostalgia.wikipedia.org? Doesn't it use the "Nostalgia" skin? --Michaeldsuarez (talk) 17:02, 2 April 2013 (UTC)

It does, and so we've forked off the Nostalgia skin into a standalone extension which can be present on Nostalgia.wiki without causing inconsistencies in our deployments. Okeyes (WMF) (talk) 17:05, 2 April 2013 (UTC)

could have used an RFC

Why wasn't there any request for comments on this? If there was, I missed it. It would have been nice if a call had gone out for volunteers to maintain well-loved skins. As it is, only a 2-week notice and it's already set in stone. Vsync (talk) 17:42, 2 April 2013 (UTC)

The message informing people that we were going to do this went out to wikitech-l...I don't have a precise date (I can find one if you want) but: at least 3 months ago. Okeyes (WMF) (talk) 17:45, 2 April 2013 (UTC)

Informing? Still sounds like it was already set in stone by a cabal :( Thanks for the pointer though, is that the place to subscribe to get a heads-up on future changes? To be honest I wouldn't even have known to look for that... Obviously Wikimedia is more than just Wikipedia but I'm primarily a Wikipedia user and even if I visit "Site News" that link doesn't jump out at me. I'll search for it though, thanks again. Vsync (talk) 18:04, 2 April 2013 (UTC)

all skins left are extremely prescriptive

The usefulness of "MySkin" and to a less extent "Simple" aren't in what they impose but rather what they leave up to the user-agent. I have my preference set accordingly because I want to have my browser set up once with my display prefences and not worry about it.

Allowing user CSS on top of these other themes is a nice feature but it still means I have to manually set each preference to what the browser had originally, not to mention that there's now no way to handle different settings on different browsers (for e.g., different display devices). There's a big difference between a skin which sets fonts, colors, etc and as you say has to be maintained; and one which simply stays out of the way and allows the user's preference to work.

I would be interested to see how much work has to go into supporting MySkin since as far as I can see it's so minimal that display just degrades to show the actual content. And "Simple" seems like it just creates the content boxes and lets color highlights show up and again, gets out of the way.

Thanks for all the hard work the designers do. It is notable though that as was pointed out earlier on this page, all the styles left force small sans-serif fonts on a bright white background. I have to wonder how much of this is the work of supporting and how much is ensuring that look is forced on everyone everywhere?

Some of the challenges in this change might be avoidable if checkboxes were present such as "avoid imposing fonts", "avoid imposing background color", "avoid imposing all colors". In this way the overall theming and layout work could still be compatible across-the-board (though, again, MySkin serves a purpose) but non-essential portions wouldn't be forced on the user. This is something I'd be more than willing to contribute code and/or testing to; I just wish I had known about this more than 13 days in advance. Vsync (talk) 17:59, 2 April 2013 (UTC)

In regards to "I have to wonder..." - absolutely zero. Our decision-making process went "which skins are not used by many users? Of those skins, which don't have anyone willing to do active maintenance?". As much as I'd love to be conniving and secretive enough to pull off a conspiracy, you give me more credit than I deserve. Okeyes (WMF) (talk) 18:06, 2 April 2013 (UTC)

Fair point, glad to hear it. I may have missed it but I just didn't see the step between the usage stats and which ones are maintained, and assessing which of them need maintenance as such. I'd love to hear if I'm wrong but I can't imagine the minimal (i.e., nonprescriptive) styles require much. And regardless, how can I volunteer to do active maintenance and then we're both happy? Vsync (talk) 18:13, 2 April 2013 (UTC)

this should help, but frankly I'm not sure if "an untested and unknown developer (if you're known to the devs, my apologies!) is willing to maintain this skin, let's leave it live for N more months" is going to be accepted. Okeyes (WMF) (talk) 18:18, 2 April 2013 (UTC)

Why?

Sorry, but I'm bloody furious about this decision. And it takes a lot to make me swear. Has this been discussed anywhere, or is it just being imposed from above? Are there any advantages in removing skins? You WILL lose editors by imposing this without good reason.Optimist on the run (talk) 19:01, 2 April 2013 (UTC)

The advantages are simple; as explained on the page, at the moment we've got three options:

1. Build things to support 9 different skins, some from 2001 and with 2001-era levels of technology, some of which aren't actually maintained properly by developers, in which case we find ourselves moving at a snail's pace and releasing software that has fewer features than it needs, late;
2. Stop supporting the old skins but still let people use them, in which case we're going to annoy every user who uses those skins every time we hit the on button;
3. Disable the old skins, thus annoying every user who uses those skins once.

We picked the third; we can't cripple our ability to upgrade and improve MediaWiki, and we can't be annoying people constantly. The change was publicly listed in Gerrit from September 2012, and commented on by many developers - the people who ultimately get to shift the situation, since the primary limitation is "nobody maintains these and building code for them slows everything down". Okeyes (WMF) (talk) 19:26, 2 April 2013 (UTC)