This may be controversial but hopefully there is logic in what I'm about to see.
I just took a look at Cologne Blue and there are a huge host of CSS issues. It makes me wonder if anyone is actually maintaining it. All the other skins seems to render nicely. Coordinates overlap the header and the Echo button is overlapping other text. It just looks shoddy [1].
Firstly is anyone actively maintaining it for Wikimedia sites? Secondly, how widely used is it on our production Wikimedia wikis?
I suspect it would be a good idea to at the very least move it into its own extension and possibly disable it on some Wikimedia sites or all Wikimedia sites.
From my personal opinion, the less skins we have to maintain on
Wikimedia sites on the better, and if no one actively cares or uses a skin then it is silly to invest time in doing so...
A good time to revist
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins
and re-run
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats
--tomasz
On Mon, Mar 10, 2014 at 2:10 PM, Jon Robson jdlrobson@gmail.com wrote:
This may be controversial but hopefully there is logic in what I'm about to see.
I just took a look at Cologne Blue and there are a huge host of CSS issues. It makes me wonder if anyone is actually maintaining it. All the other skins seems to render nicely. Coordinates overlap the header and the Echo button is overlapping other text. It just looks shoddy [1].
Firstly is anyone actively maintaining it for Wikimedia sites? Secondly, how widely used is it on our production Wikimedia wikis?
I suspect it would be a good idea to at the very least move it into its own extension and possibly disable it on some Wikimedia sites or all Wikimedia sites.
From my personal opinion, the less skins we have to maintain on Wikimedia sites on the better, and if no one actively cares or uses a skin then it is silly to invest time in doing so...
[1] http://imgur.com/hNZAFUb,bRHdhCS
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Tomasz Finc wrote:
On Mon, Mar 10, 2014 at 2:10 PM, Jon Robson jdlrobson@gmail.com wrote:
Why is Cologne Blue still in core?
A good time to revist
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins
and re-run
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats
Tomasz' links are relevant, but it's on the associated talk page that there's a hint of the real reason this happened: Cologne Blue was originally slated for execution, but MatmaRex (an active developer) stepped up and agreed to maintain it.
Stats would be helpful here, though I disagree with the general premise: most of the old skins were fine to kill because they sucked, they weren't used very much, and/or they weren't maintained. However that does _not_ mean that we should only have one or two skins. Instead, the skinning system should just suck a lot less: it should be easier to add or remove a pre-made skin, it should be easier to customize a skin (site-wide), it should be easier to create a new skin, it should be easier for users to change their skin, and any additional skins we ship with MediaWiki should be more attractive and should cleanly support extensions.
In addition to unfairly limiting user choice, site configurability, and site customizability, the current skins system also encourages confusion between Wikipedia and non-Wikimedia wikis. In other words, if it weren't so dreadful to change MediaWiki's skin, there would presumably be reduced user confusion. As it is, almost every MediaWiki wiki looks the same.
MZMcBride
It might be easier to revamp the skin system if there were fewer skins to port.
- Trevor
On Mon, Mar 10, 2014 at 6:48 PM, MZMcBride z@mzmcbride.com wrote:
Tomasz Finc wrote:
On Mon, Mar 10, 2014 at 2:10 PM, Jon Robson jdlrobson@gmail.com wrote:
Why is Cologne Blue still in core?
A good time to revist
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins
and re-run
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats
Tomasz' links are relevant, but it's on the associated talk page that there's a hint of the real reason this happened: Cologne Blue was originally slated for execution, but MatmaRex (an active developer) stepped up and agreed to maintain it.
Stats would be helpful here, though I disagree with the general premise: most of the old skins were fine to kill because they sucked, they weren't used very much, and/or they weren't maintained. However that does _not_ mean that we should only have one or two skins. Instead, the skinning system should just suck a lot less: it should be easier to add or remove a pre-made skin, it should be easier to customize a skin (site-wide), it should be easier to create a new skin, it should be easier for users to change their skin, and any additional skins we ship with MediaWiki should be more attractive and should cleanly support extensions.
In addition to unfairly limiting user choice, site configurability, and site customizability, the current skins system also encourages confusion between Wikipedia and non-Wikimedia wikis. In other words, if it weren't so dreadful to change MediaWiki's skin, there would presumably be reduced user confusion. As it is, almost every MediaWiki wiki looks the same.
MZMcBride
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
It might be easier to revamp the skin system if there were fewer skins to port.
Touché.
So.. so 2 questions 1) would anyone have any objections to moving it out of core and into its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
I support moving it to an extension and enabling it on deployed sites as to avoid an disruption in service for the users of the skin.
- Trevor
On Tue, Mar 11, 2014 at 12:21 PM, Jon Robson jdlrobson@gmail.com wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
Touché.
So.. so 2 questions
- would anyone have any objections to moving it out of core and into
its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 11 March 2014 15:21, Jon Robson jdlrobson@gmail.com wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
Touché.
So.. so 2 questions
- would anyone have any objections to moving it out of core and into
its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
Speaking from a user perspective, having all of 4 skins available (really? 4? That's a lot?) does make a difference in differentiating which nearly-identical wikis one is working on, particularly if one or more of the wikis involved are non-public ones. Aside from Vector and Monobook, though, I don't see a reason why they can't be extensions, as long as they're still in the preference lists.
I strongly urge maintaining Monobook exactly the way that Vector is maintained: it's the clear favourite of the most active users, it's still faster after years of improving Vector, and it handles a lot of accessibility issues much better than Vector (particularly for the visually impaired, according to those editors I know who have to deal with this).
Risker/Anne
I don't think anyone is suggesting removing or even moving Monobook. I think we are more talking about assigning effort more proportionally to preference of users. Cologne Blue is not the clear favorite of any group of users that we know of.
- Trevor
On Tue, Mar 11, 2014 at 12:36 PM, Risker risker.wp@gmail.com wrote:
On 11 March 2014 15:21, Jon Robson jdlrobson@gmail.com wrote:
It might be easier to revamp the skin system if there were fewer skins
to
port.
Touché.
So.. so 2 questions
- would anyone have any objections to moving it out of core and into
its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
Speaking from a user perspective, having all of 4 skins available (really? 4? That's a lot?) does make a difference in differentiating which nearly-identical wikis one is working on, particularly if one or more of the wikis involved are non-public ones. Aside from Vector and Monobook, though, I don't see a reason why they can't be extensions, as long as they're still in the preference lists.
I strongly urge maintaining Monobook exactly the way that Vector is maintained: it's the clear favourite of the most active users, it's still faster after years of improving Vector, and it handles a lot of accessibility issues much better than Vector (particularly for the visually impaired, according to those editors I know who have to deal with this).
Risker/Anne _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Mar 11, 2014 3:36 PM, "Risker" risker.wp@gmail.com wrote:
differentiating which nearly-identical wikis one is working on, particularly if one or more of the wikis involved are non-public ones.
Do you have specific requests for how to differentiate them?
Either for a whole wiki (could be in main common.css) or your user (your own common.css)
-Jeremy
On 11/03/14 19:21, Jon Robson wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
Touché.
So.. so 2 questions
- would anyone have any objections to moving it out of core and into
its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Aye, it should be an extension. All the skins should probably be extensions - I'd argue that even the default one should be implemented as an extension. Ship it with the main tarball, have it included by default, but keep it as an extension so it's all still modular. It would make working with them so much easier.
And it'd also make it easier for people to make new skins, since that way they'd have a sane default to use as a template. Vector as is is anything but sane.
-I
On 11 March 2014 13:17, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 19:21, Jon Robson wrote:
It might be easier to revamp the skin system if there were fewer skins to
port.
Touché.
So.. so 2 questions
- would anyone have any objections to moving it out of core and into
its own extension? 2) would anyone have any objections for turning it off on Wikimedia wikis
FWIW, having a single skin in core would actually be a good thing for skin development. Currently writing a skin outside core is difficult - I found this in the development of the mobile skin. Having a leaner codebase in core would actually encourage better organisation to make things. skinStyles is a great example - if your skin is not in core, you can't use it.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Aye, it should be an extension. All the skins should probably be extensions - I'd argue that even the default one should be implemented as an extension. Ship it with the main tarball, have it included by default, but keep it as an extension so it's all still modular. It would make working with them so much easier.
+1
J.
On 11/03/14 17:26, Trevor Parscal wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
- Trevor
But that completely ignores the primary users of non-Vector skins - third party users who have developed their own skins. Every time the skinning systems changes, their skins break and it adds yet another blocker for them upgrading. Now this wouldn't be a problem if it were a one-time thing to fix up the system in general, but it isn't - it happens almost every other version, and usually for quite little things. It's like, oh hey, everything broke again. Now we need to redo it all. Again.
The number of skins is not a blocker for revamping the skin system. There will always be too many skins to port. The blocker is that to revamp the skin system, it needs to be done well enough to justify breaking the skins in the first place, however many there are, and that is a whole lot harder.
-I
On 11 March 2014 13:10, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 17:26, Trevor Parscal wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
- Trevor
But that completely ignores the primary users of non-Vector skins
Sure, because they aren't in MediaWiki core. Trevor's point is that there are four skins to fix *in the same commit* as fixing the skins system, and this would be a lot easier if there were instead three, or two, or one.
In fact, this is the *reason* we stopped adding new skins to MediaWiki core, and instead they're treated like extensions and sysadmins are expected to pay attention to breaking changes.
The wider question about fixing "once and for all" the changes to skins is out of scope for this discussion, and would involve some serious thought, but that's something that can be done regardless of whether Cologne Blue is in core. It'd just make it easier.
J.
On 11/03/14 20:16, James Forrester wrote:
On 11 March 2014 13:10, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 17:26, Trevor Parscal wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
- Trevor
But that completely ignores the primary users of non-Vector skins
Sure, because they aren't in MediaWiki core. Trevor's point is that there are four skins to fix *in the same commit* as fixing the skins system, and this would be a lot easier if there were instead three, or two, or one.
In fact, this is the *reason* we stopped adding new skins to MediaWiki core, and instead they're treated like extensions and sysadmins are expected to pay attention to breaking changes.
The wider question about fixing "once and for all" the changes to skins is out of scope for this discussion, and would involve some serious thought, but that's something that can be done regardless of whether Cologne Blue is in core. It'd just make it easier.
J.
Because they aren't in core, my point is that people forget they exist at all. They shouldn't do that.
But yes, having to fix them all in one commit is silly. More on that elsewhere.
-I
If people forget they exist I would say that equates to no one cares about them and no one maintains them. If this is true, we are doing a disservice to our users by providing these skins to our users.
On Tue, Mar 11, 2014 at 1:20 PM, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 20:16, James Forrester wrote:
On 11 March 2014 13:10, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 17:26, Trevor Parscal wrote:
It might be easier to revamp the skin system if there were fewer skins to port.
- Trevor
But that completely ignores the primary users of non-Vector skins
Sure, because they aren't in MediaWiki core. Trevor's point is that there are four skins to fix *in the same commit* as fixing the skins system, and this would be a lot easier if there were instead three, or two, or one.
In fact, this is the *reason* we stopped adding new skins to MediaWiki core, and instead they're treated like extensions and sysadmins are expected to pay attention to breaking changes.
The wider question about fixing "once and for all" the changes to skins is out of scope for this discussion, and would involve some serious thought, but that's something that can be done regardless of whether Cologne Blue is in core. It'd just make it easier.
J.
Because they aren't in core, my point is that people forget they exist at all. They shouldn't do that.
But yes, having to fix them all in one commit is silly. More on that elsewhere.
-I
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 11/03/14 21:21, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them. If this is true, we are doing a disservice to our users by providing these skins to our users.
Developers forget some users exist. They forget use cases exist. It happens. It does not invalidate them. We should not be alienating our users, whoever they may be, for lack of consideration.
We should instead consider things first and THEN alienate them, of course.
-I
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
On Tue, Mar 11, 2014 at 2:38 PM, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 21:21, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them. If this is true, we are doing a disservice to our users by providing these skins to our users.
Developers forget some users exist. They forget use cases exist. It happens. It does not invalidate them. We should not be alienating our users, whoever they may be, for lack of consideration.
We should instead consider things first and THEN alienate them, of course.
-I
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
+ 10000 to what Trevor just said.
Isarra, developers make a _decision_ to forget some users exist. This is why Wikipedia doesn't support IE5. I am suggesting a decision that we forget that users of Cologne blue exist. Although it might seem harsh to the 98,541 readers but justified. It's just not getting any attention by users or developers.
Based on the numbers on https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats - my personal view is if a skin isn't used by more than a million people we shouldn't be supporting it on Wikimedia wikis. This would actually mean just keeping Vector and Monobook. I for one would actually give more attention to Monobook in my testing if this happened.
Can someone who is able to please run the stats for https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats again?
On Tue, Mar 11, 2014 at 3:45 PM, Trevor Parscal tparscal@wikimedia.org wrote:
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
On Tue, Mar 11, 2014 at 2:38 PM, Isarra Yos zhorishna@gmail.com wrote:
On 11/03/14 21:21, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them. If this is true, we are doing a disservice to our users by providing these skins to our users.
Developers forget some users exist. They forget use cases exist. It happens. It does not invalidate them. We should not be alienating our users, whoever they may be, for lack of consideration.
We should instead consider things first and THEN alienate them, of course.
-I
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 11 March 2014 18:48, Jon Robson jdlrobson@gmail.com wrote:
- 10000 to what Trevor just said.
Isarra, developers make a _decision_ to forget some users exist. This is why Wikipedia doesn't support IE5. I am suggesting a decision that we forget that users of Cologne blue exist. Although it might seem harsh to the 98,541 readers but justified. It's just not getting any attention by users or developers.
Based on the numbers on https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats - my personal view is if a skin isn't used by more than a million people we shouldn't be supporting it on Wikimedia wikis. This would actually mean just keeping Vector and Monobook. I for one would actually give more attention to Monobook in my testing if this happened.
Can someone who is able to please run the stats for https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats again?
Ummm. Readers only get Vector, unless they create an account, log in, find their preferences, and make a conscious decision to choose another skin - something almost no reader will do.
Editors, it seems, chose Monobook 2:1 over Vector as of this time last year, when there were many other skins; that is, they make the conscious choice. (I also think there's something really wonky about the "power user" numbers. There's no way there are less than 15,000 users active in the last six months with over 1000 edits across all of the projects.) There's a lot of value in paying attention to Monobook.
Risker/Anne
Risker, I'm not saying there is not value in paying attention to Monobook. I'm saying quite the opposite.
I'm saying I _currently_ pay less attention to Monobook as I currently think of it the same way as Cologne Blue and Modern which I couldn't care less for. Take away those 2, it becomes much easier for me to pay attention to it when I develop as I now only have to think about 2 production skins. It also makes it easier to make radical skin changes that do not break either of these 2 important skins, which on the long term promotes a healthier code base for everyone.
IVector in my mental model becomes the 'reader skin' and Monobook the 'editor skin'.
On Tue, Mar 11, 2014 at 4:02 PM, Risker risker.wp@gmail.com wrote:
On 11 March 2014 18:48, Jon Robson jdlrobson@gmail.com wrote:
- 10000 to what Trevor just said.
Isarra, developers make a _decision_ to forget some users exist. This is why Wikipedia doesn't support IE5. I am suggesting a decision that we forget that users of Cologne blue exist. Although it might seem harsh to the 98,541 readers but justified. It's just not getting any attention by users or developers.
Based on the numbers on https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats - my personal view is if a skin isn't used by more than a million people we shouldn't be supporting it on Wikimedia wikis. This would actually mean just keeping Vector and Monobook. I for one would actually give more attention to Monobook in my testing if this happened.
Can someone who is able to please run the stats for https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats again?
Ummm. Readers only get Vector, unless they create an account, log in, find their preferences, and make a conscious decision to choose another skin - something almost no reader will do.
Editors, it seems, chose Monobook 2:1 over Vector as of this time last year, when there were many other skins; that is, they make the conscious choice. (I also think there's something really wonky about the "power user" numbers. There's no way there are less than 15,000 users active in the last six months with over 1000 edits across all of the projects.) There's a lot of value in paying attention to Monobook.
Risker/Anne _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 11 March 2014 16:02, Risker risker.wp@gmail.com wrote:
Editors, it seems, chose Monobook 2:1 over Vector as of this time last year, when there were many other skins; that is, they make the conscious choice. (I also think there's something really wonky about the "power user" numbers. There's no way there are less than 15,000 users active in the last six months with over 1000 edits across all of the projects.) There's a lot of value in paying attention to Monobook.
This is based on a mis-understanding of https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats. Specifically:
One limitation of the data is that we could only retrieve users who had
actively set a preference for their skin; users who had not done this displayed with a null value in the relevant table. Such users are not included in the numbers for Vector in the next section, which leads to an *underrepresentation of Vector*. Based on an approximate calculation,[3]https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats#cite_note-3 this may *exclude up to 85 per cent of users*.
[My emphases.]
To put it another way, all the data shows is that of active users who'd made over 1000 edits, 1/3 had actively changed their skin setting to Vector and were still there at the time.
Example scenarios:
- Users who opted-in early to Vector when Monobook was still default (there were lots of us back then). - Users who chose to go with Cologne Blue back in the day, then Simple, then later picked Vector. - Users who stayed with Classic when Monobook was introduced (*waves walking stick*), then later moved to Vector.
Making a linear assumption (which is also wrong) would suggest that there were ~100,000 active users with >= 1000 edits, of whom ~90% use Vector, ~9% use Monobook, and ~1% use Cologne Blue or Modern. 10:1 in favour of Vector is more representative than 2:1 in favour of Monobook, though some up-to-date data would be appreciated at some point, I'd suggest.
J.
On 11 March 2014 16:33, James Forrester jforrester@wikimedia.org wrote:
On 11 March 2014 16:02, Risker risker.wp@gmail.com wrote:
Editors, it seems, chose Monobook 2:1 over Vector as of this time last year, when there were many other skins; that is, they make the conscious choice. (I also think there's something really wonky about the "power
user"
numbers. There's no way there are less than 15,000 users active in the last six months with over 1000 edits across all of the projects.)
There's
a lot of value in paying attention to Monobook.
This is based on a mis-understanding of https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats. Specifically:
One limitation of the data is that we could only retrieve users who had
actively set a preference for their skin; users who had not done this displayed with a null value in the relevant table. Such users are not included in the numbers for Vector in the next section, which leads to
an *underrepresentation
of Vector*. Based on an approximate calculation,[3]<
https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats#cite_note-3... this
may *exclude up to 85 per cent of users*.
[My emphases.]
To put it another way, all the data shows is that of active users who'd made over 1000 edits, 1/3 had actively changed their skin setting to Vector and were still there at the time.
As the person who gathered the data, that's a far more accurate way of representing them, yep.
James Forrester wrote:
This is based on a mis-understanding of https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats. Specifically:
One limitation of the data is that we could only retrieve users who had actively set a preference for their skin; users who had not done this displayed with a null value in the relevant table. Such users are not included in the numbers for Vector in the next section, which leads to an underrepresentation of Vector. Based on an approximate calculation, this may exclude up to 85 per cent of users.
I'm having difficulty understanding how this point would matter for power-users. Can't we presume that every power-user has set his or her skin user preference? If so, we can see that roughly 63% of power-users use Monobook. Given the reality of who largely maintains and protects the wikis (i.e., power-users), the number of Monobook users certainly seems significant to me. Perhaps I just need to think further on this, though.
Updated stats might help here, but Trevor is probably right that the proper solution is a skins API. Whether that means it's a good idea to kill Cologne Blue and Modern in the meantime, I don't know. It'll likely just engender more ill will and not provide that much of a savings. It's not as though anyone is currently prioritizing bug fixes and feature requests related to skins other than Vector and Monobook. In other words, the de facto support standard has long been only Monobook and Vector. Shrug.
MZMcBride
On 11 March 2014 17:05, MZMcBride z@mzmcbride.com wrote:
James Forrester wrote:
This is based on a mis-understanding of https://meta.wikimedia.org/wiki/Turning_off_outdated_skins/stats. Specifically:
One limitation of the data is that we could only retrieve users who had actively set a preference for their skin; users who had not done this displayed with a null value in the relevant table. Such users are not included in the numbers for Vector in the next section, which leads to an underrepresentation of Vector. Based on an approximate calculation, this may exclude up to 85 per cent of users.
I'm having difficulty understanding how this point would matter for power-users. Can't we presume that every power-user has set his or her skin user preference? If so, we can see that roughly 63% of power-users use Monobook. Given the reality of who largely maintains and protects the wikis (i.e., power-users), the number of Monobook users certainly seems significant to me. Perhaps I just need to think further on this, though.
Not necessarily. I think we can assume power users who registered pre-Vector have totally done that, either to explicitly opt-in to vector or to stay with monobook or CB or [one of the other now-removed skins that still appear in the user_properties table]. But power users who registered after the Vector default switch - I can't see a reason for them to explicitly enable it, if they're comfortable with it. Anecdotally I know a few strange creatures who fall into that category, although personally I prefer Monobook.
I am one of these creatures. I didn’t set a preference when I joined and just rode through the default changes as it went. Later, I ended up setting a preference to Vector, but that was after I had played around with other skins, so I needed a way to get back to the one I liked.
Personally, I think talk of “power user skin” versus “reader skin” is pointless. I think we should be focusing on making the software and system better. If that means “a better skin api” or it means “a new skin” or it means “removing old skins”, let’s do that.
On Mar 11, 2014, at 5:11 PM, Oliver Keyes okeyes@wikimedia.org wrote:
But power users who registered after the Vector default switch - I can't see a reason for them to explicitly enable it, if they're comfortable with it. Anecdotally I know a few strange creatures who fall into that category, although personally I prefer Monobook.
--- Brandon Harris, Senior Designer, Wikimedia Foundation
Support Free Knowledge: http://wikimediafoundation.org/wiki/Donate
Brandon Harris wrote:
Personally, I think talk of "power user skin" versus "reader skin" is pointless.
Pointless? Would it help if we used the term "user stories"? :-)
https://www.mediawiki.org/wiki/Special:Search/user_stories
I think we should be focusing on making the software and system better. If that means "a better skin api" or it means "a new skin" or it means "removing old skins", let’s do that.
This has the feeling of a false trichotomy. I don't think anyone will be revamping the skins system anytime soon, but that isn't to say that means we should instead simply remove the skins. There's always the option of doing nothing or moving the skins into extensions or many others.
I believe the open question remains what to do with Cologne Blue and Modern. I don't think there's much objection to moving them into extensions. Then we can debate whether or not to bundle the extensions by default with core. It'll be great. ;-)
MZMcBride
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc project?
-I
Le 11/03/2014 23:55, Isarra Yos a écrit :
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc project?
I don't think so.
A few weeks of unexperimented CS student is most probably not going to solve an issue we had for years and years.
Okay to make sure stuff comes out of this thread...
So unless someone does this before me I am going to move CologneBlue out of core and into SkinCologneBlue extension and Modern out of core into SkinModern extension.
If someone could rerun the data collection for skin usage I most be most appreciative. Any takers?
Then let's decide what we are going to do with this and rethink what our attitude towards skin support is. On 12 Mar 2014 00:15, "Antoine Musso" hashar+wmf@free.fr wrote:
Le 11/03/2014 23:55, Isarra Yos a écrit :
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc project?
I don't think so.
A few weeks of unexperimented CS student is most probably not going to solve an issue we had for years and years.
-- Antoine "hashar" Musso
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Is there a particular reason for making them extensions and not (installable) skins?
Cheers, Stephan On 12 Mar 2014 16:47, "Jon Robson" jdlrobson@gmail.com wrote:
Okay to make sure stuff comes out of this thread...
So unless someone does this before me I am going to move CologneBlue out of core and into SkinCologneBlue extension and Modern out of core into SkinModern extension.
If someone could rerun the data collection for skin usage I most be most appreciative. Any takers?
Then let's decide what we are going to do with this and rethink what our attitude towards skin support is. On 12 Mar 2014 00:15, "Antoine Musso" hashar+wmf@free.fr wrote:
Le 11/03/2014 23:55, Isarra Yos a écrit :
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few
new
APIs which are needed to no longer have skin extensions be
second-class
citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc project?
I don't think so.
A few weeks of unexperimented CS student is most probably not going to solve an issue we had for years and years.
-- Antoine "hashar" Musso
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Jon, I support your proposed action and would be glad to review.
- Trevor
On Wed, Mar 12, 2014 at 9:03 AM, Stephan Gambke s7eph4n@gmail.com wrote:
Is there a particular reason for making them extensions and not (installable) skins?
Cheers, Stephan On 12 Mar 2014 16:47, "Jon Robson" jdlrobson@gmail.com wrote:
Okay to make sure stuff comes out of this thread...
So unless someone does this before me I am going to move CologneBlue out
of
core and into SkinCologneBlue extension and Modern out of core into SkinModern extension.
If someone could rerun the data collection for skin usage I most be most appreciative. Any takers?
Then let's decide what we are going to do with this and rethink what our attitude towards skin support is. On 12 Mar 2014 00:15, "Antoine Musso" hashar+wmf@free.fr wrote:
Le 11/03/2014 23:55, Isarra Yos a écrit :
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few
new
APIs which are needed to no longer have skin extensions be
second-class
citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc
project?
I don't think so.
A few weeks of unexperimented CS student is most probably not going to solve an issue we had for years and years.
-- Antoine "hashar" Musso
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 12/03/14 15:47, Jon Robson wrote:
Okay to make sure stuff comes out of this thread...
So unless someone does this before me I am going to move CologneBlue out of core and into SkinCologneBlue extension and Modern out of core into SkinModern extension.
If someone could rerun the data collection for skin usage I most be most appreciative. Any takers?
Then let's decide what we are going to do with this and rethink what our attitude towards skin support is.
That would be an excellent start. Know, however, that skin extensions belong in the skins/ repo, not the extensions/, so there is no need for a 'skin' prefix.
-I
On 03/12/2014 11:47 AM, Jon Robson wrote:
Okay to make sure stuff comes out of this thread...
So unless someone does this before me I am going to move CologneBlue out of core and into SkinCologneBlue extension and Modern out of core into SkinModern extension.
Would you mind keeping this work in a branch until 1.23 is released in May/April? That way we could announce what is coming in 1.24 and you could merge it in right after 1.23 is released.
Thanks,
Mark.
On 12/03/14 07:15, Antoine Musso wrote:
Le 11/03/2014 23:55, Isarra Yos a écrit :
On 11/03/14 22:45, Trevor Parscal wrote:
By making all skins extensions it would also force us to make a few new APIs which are needed to no longer have skin extensions be second-class citizens.
This should happen.
- Trevor
Quite so. Think hitting on some of this could be a viable gsoc project?
I don't think so.
A few weeks of unexperimented CS student is most probably not going to solve an issue we had for years and years.
I thought projects were a couple of months? Anyway, I've got someone in mind who isn't new to mw and is quite familiar with skins... just not the ones in core, so he'd need a mentor knowledgeable about that tangled mess. Would anyone here be able to help with that, or would it be pushing the scope?
-I
On 03/11/2014 05:21 PM, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them.
Isarra was referring to skins outside of core.
There's a difference between core developers forgetting non-core skins exist, and the developers of non-core skins forgetting.
If we had a proper skin API, people wouldn't explicitly need to think about Foo non-core skin, but they would need to think about managing the changes to the skin API (the same way we manage other API changes).
Matt Flaschen
Yes this is an orthogonal conversation. If it's that easy for a core change to break a skin outside core, then there are lots of fundamentally wrong things with our skin system, one being the fact that modules added with OutputPage get added to all skins even if they might not be compatible with them. If we want to talk about this I'd encourage you to start a new thread.
On Wed, Mar 12, 2014 at 11:57 AM, Matthew Flaschen mflaschen@wikimedia.org wrote:
On 03/11/2014 05:21 PM, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them.
Isarra was referring to skins outside of core.
There's a difference between core developers forgetting non-core skins exist, and the developers of non-core skins forgetting.
If we had a proper skin API, people wouldn't explicitly need to think about Foo non-core skin, but they would need to think about managing the changes to the skin API (the same way we manage other API changes).
Matt Flaschen
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
I'm going to start working on some RL modifications to make it possible for skins outside of core to add skinStyles to other modules, which will help with making non-core skins equally capable.
- Trevor
On Wed, Mar 12, 2014 at 12:00 PM, Jon Robson jdlrobson@gmail.com wrote:
Yes this is an orthogonal conversation. If it's that easy for a core change to break a skin outside core, then there are lots of fundamentally wrong things with our skin system, one being the fact that modules added with OutputPage get added to all skins even if they might not be compatible with them. If we want to talk about this I'd encourage you to start a new thread.
On Wed, Mar 12, 2014 at 11:57 AM, Matthew Flaschen mflaschen@wikimedia.org wrote:
On 03/11/2014 05:21 PM, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them.
Isarra was referring to skins outside of core.
There's a difference between core developers forgetting non-core skins exist, and the developers of non-core skins forgetting.
If we had a proper skin API, people wouldn't explicitly need to think
about
Foo non-core skin, but they would need to think about managing the
changes
to the skin API (the same way we manage other API changes).
Matt Flaschen
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- Jon Robson
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Help needed!
Okay so I made a start to this. I'm probably not going to get time to work on this during this week, but if anyone wants to pick this up and accelerate it through to completion be my guest. I will love you forever. Otherwise I'll continue the work next week.
https://gerrit.wikimedia.org/r/118345 https://gerrit.wikimedia.org/r/118347
On Wed, Mar 12, 2014 at 12:23 PM, Trevor Parscal tparscal@wikimedia.org wrote:
I'm going to start working on some RL modifications to make it possible for skins outside of core to add skinStyles to other modules, which will help with making non-core skins equally capable.
- Trevor
On Wed, Mar 12, 2014 at 12:00 PM, Jon Robson jdlrobson@gmail.com wrote:
Yes this is an orthogonal conversation. If it's that easy for a core change to break a skin outside core, then there are lots of fundamentally wrong things with our skin system, one being the fact that modules added with OutputPage get added to all skins even if they might not be compatible with them. If we want to talk about this I'd encourage you to start a new thread.
On Wed, Mar 12, 2014 at 11:57 AM, Matthew Flaschen mflaschen@wikimedia.org wrote:
On 03/11/2014 05:21 PM, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them.
Isarra was referring to skins outside of core.
There's a difference between core developers forgetting non-core skins exist, and the developers of non-core skins forgetting.
If we had a proper skin API, people wouldn't explicitly need to think
about
Foo non-core skin, but they would need to think about managing the
changes
to the skin API (the same way we manage other API changes).
Matt Flaschen
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- Jon Robson
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
This is done. Please review! Thanks to Isarra for getting the new repositories setup!
https://gerrit.wikimedia.org/r/118345 https://gerrit.wikimedia.org/r/119885 https://gerrit.wikimedia.org/r/119884
On Wed, Mar 12, 2014 at 2:13 PM, Jon Robson jdlrobson@gmail.com wrote:
Help needed!
Okay so I made a start to this. I'm probably not going to get time to work on this during this week, but if anyone wants to pick this up and accelerate it through to completion be my guest. I will love you forever. Otherwise I'll continue the work next week.
https://gerrit.wikimedia.org/r/118345 https://gerrit.wikimedia.org/r/118347
On Wed, Mar 12, 2014 at 12:23 PM, Trevor Parscal tparscal@wikimedia.org wrote:
I'm going to start working on some RL modifications to make it possible
for
skins outside of core to add skinStyles to other modules, which will help with making non-core skins equally capable.
- Trevor
On Wed, Mar 12, 2014 at 12:00 PM, Jon Robson jdlrobson@gmail.com
wrote:
Yes this is an orthogonal conversation. If it's that easy for a core change to break a skin outside core, then there are lots of fundamentally wrong things with our skin system, one being the fact that modules added with OutputPage get added to all skins even if they might not be compatible with them. If we want to talk about this I'd encourage you to start a new thread.
On Wed, Mar 12, 2014 at 11:57 AM, Matthew Flaschen mflaschen@wikimedia.org wrote:
On 03/11/2014 05:21 PM, Jon Robson wrote:
If people forget they exist I would say that equates to no one cares about them and no one maintains them.
Isarra was referring to skins outside of core.
There's a difference between core developers forgetting non-core skins exist, and the developers of non-core skins forgetting.
If we had a proper skin API, people wouldn't explicitly need to think
about
Foo non-core skin, but they would need to think about managing the
changes
to the skin API (the same way we manage other API changes).
Matt Flaschen
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- Jon Robson
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- Jon Robson
wikitech-l@lists.wikimedia.org