Conversation has been taking place on CodeReview about whether people (developer and system administrators) will find it confusing that there's code in both "extensions/Vector" and "skins/vector" that are related but not the same thing.
I personally find it simple to understand and don't expect most developers and system administrators to feel otherwise, but in the interest of upholding the community-driven decision making process that has made MediaWiki what it is, I wanted to bring a little extra attention to the matter so we can make a decision and move on with our lives.
I'm interested in views on whether the Vector extension should be named something else or remain as it is. If your view is that it should be renamed, suggestions for what it should be named would be useful.
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
- Trevor
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
On Tue, Oct 12, 2010 at 8:35 PM, Trevor Parscal tparscal@wikimedia.org wrote:
I personally find it simple to understand and don't expect most developers and system administrators to feel otherwise,
People are still confused with Wikimedia/MediaWiki to this day. You think skins/vector and extensions/Vector won't confuse people?
I'm interested in views on whether the Vector extension should be named something else or remain as it is. If your view is that it should be renamed, suggestions for what it should be named would be useful.
I think it should be merged into core. Full stop.
-Chad
On 10/12/2010 8:35 PM, Trevor Parscal wrote:
Conversation has been taking place on CodeReview about whether people (developer and system administrators) will find it confusing that there's code in both "extensions/Vector" and "skins/vector" that are related but not the same thing.
I personally find it simple to understand and don't expect most developers and system administrators to feel otherwise, but in the interest of upholding the community-driven decision making process that has made MediaWiki what it is, I wanted to bring a little extra attention to the matter so we can make a decision and move on with our lives.
I'm interested in views on whether the Vector extension should be named something else or remain as it is. If your view is that it should be renamed, suggestions for what it should be named would be useful.
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
I think its a bit disingenuous to say that the only issue is the naming of the Vector extension. My opposition is to splitting the UsabilityInitiative into separate extensions for what will very likely be only a *single release*. And as Chad noted, there's also the question of whether they should even continue to be an extension. Or to put it another way, should we intentionally hold off on adding the usability stuff to core until the next release because we're already adding ResourceLoader in 1.17? That's hardly a bike shed problem.
On 10/12/10 6:10 PM, Alex wrote:
On 10/12/2010 8:35 PM, Trevor Parscal wrote:
Conversation has been taking place on CodeReview about whether people (developer and system administrators) will find it confusing that there's code in both "extensions/Vector" and "skins/vector" that are related but not the same thing.
I personally find it simple to understand and don't expect most developers and system administrators to feel otherwise, but in the interest of upholding the community-driven decision making process that has made MediaWiki what it is, I wanted to bring a little extra attention to the matter so we can make a decision and move on with our lives.
I'm interested in views on whether the Vector extension should be named something else or remain as it is. If your view is that it should be renamed, suggestions for what it should be named would be useful.
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
I think its a bit disingenuous to say that the only issue is the naming of the Vector extension. My opposition is to splitting the UsabilityInitiative into separate extensions for what will very likely be only a *single release*. And as Chad noted, there's also the question of whether they should even continue to be an extension. Or to put it another way, should we intentionally hold off on adding the usability stuff to core until the next release because we're already adding ResourceLoader in 1.17? That's hardly a bike shed problem.
Nor is it the problem I am referring to as a bike shed problem.
- Trevor
On 10/12/2010 9:58 PM, Trevor Parscal wrote:
On 10/12/10 6:10 PM, Alex wrote:
On 10/12/2010 8:35 PM, Trevor Parscal wrote:
Conversation has been taking place on CodeReview about whether people (developer and system administrators) will find it confusing that there's code in both "extensions/Vector" and "skins/vector" that are related but not the same thing.
I personally find it simple to understand and don't expect most developers and system administrators to feel otherwise, but in the interest of upholding the community-driven decision making process that has made MediaWiki what it is, I wanted to bring a little extra attention to the matter so we can make a decision and move on with our lives.
I'm interested in views on whether the Vector extension should be named something else or remain as it is. If your view is that it should be renamed, suggestions for what it should be named would be useful.
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
I think its a bit disingenuous to say that the only issue is the naming of the Vector extension. My opposition is to splitting the UsabilityInitiative into separate extensions for what will very likely be only a *single release*. And as Chad noted, there's also the question of whether they should even continue to be an extension. Or to put it another way, should we intentionally hold off on adding the usability stuff to core until the next release because we're already adding ResourceLoader in 1.17? That's hardly a bike shed problem.
Nor is it the problem I am referring to as a bike shed problem.
Correct, that's the problem you didn't even mention.
Trevor Parscal wrote:
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
Naming and naming conventions aren't trivialities.
The issue here is not a color of the bike shed problem. The issue is that almost everyone who has commented is in favor of getting rid of the bike shed and putting the bikes in the garage. The opposition to putting this extension (these extensions) in core seems to be largely focused on a lack of written process for when and when not to merge extensions into core.
MZMcBride
On 10/12/10 7:42 PM, MZMcBride wrote:
Trevor Parscal wrote:
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
Naming and naming conventions aren't trivialities.
Show me the convention for naming extensions and I will follow it. As long as there is none, spending time going back and forth discussing the name of something that users will never see is indeed spending time on a trivial matter.
The issue here is not a color of the bike shed problem. The issue is that almost everyone who has commented is in favor of getting rid of the bike shed and putting the bikes in the garage. The opposition to putting this extension (these extensions) in core seems to be largely focused on a lack of written process for when and when not to merge extensions into core.
That's an entirely different discussion, and the results of that discussion have so far been that any such action is being deferred until after 1.17.
MZMcBride
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Tue, Oct 12, 2010 at 10:45 PM, Trevor Parscal tparscal@wikimedia.org wrote:
On 10/12/10 7:42 PM, MZMcBride wrote:
Trevor Parscal wrote:
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
Naming and naming conventions aren't trivialities.
Show me the convention for naming extensions and I will follow it. As long as there is none, spending time going back and forth discussing the name of something that users will never see is indeed spending time on a trivial matter.
There isn't one. Maybe it's because we haven't cared until now, or maybe because we haven't needed it. Policy or not, concerns were raised about the naming of *this* extension. Saying "well there's no policy so it's pointless to discuss it" is a cop-out.
-Chad
On 10/12/10 8:03 PM, Chad wrote:
On Tue, Oct 12, 2010 at 10:45 PM, Trevor Parscaltparscal@wikimedia.org wrote:
On 10/12/10 7:42 PM, MZMcBride wrote:
Trevor Parscal wrote:
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
Naming and naming conventions aren't trivialities.
Show me the convention for naming extensions and I will follow it. As long as there is none, spending time going back and forth discussing the name of something that users will never see is indeed spending time on a trivial matter.
There isn't one. Maybe it's because we haven't cared until now, or maybe because we haven't needed it. Policy or not, concerns were raised about the naming of *this* extension. Saying "well there's no policy so it's pointless to discuss it" is a cop-out.
I am explicitly requesting discussion about it so that I can take action based on the desire of the community. What cop-out are you referring to?
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
Some people are being somewhat combatant or getting tangential in their responses - fine, basically what I expect form this list - but I'm still very interested in responses to do with my original question, "should we rename it, and if so, what should we name it?".
- Trevor
-Chad
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Tue, Oct 12, 2010 at 11:10 PM, Trevor Parscal tparscal@wikimedia.org wrote:
I am explicitly requesting discussion about it so that I can take action based on the desire of the community. What cop-out are you referring to?
That would be this one:
Show me the convention for naming extensions and I will follow it. As long as there is none, spending time going back and forth discussing the name of something that users will never see is indeed spending time on a trivial matter.
This is coming across as "no policy on naming, therefore naming does not really matter."
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
Why does it have to wait until after 1.17, again?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
If it *has* to be its own extension and not a part of core of the main UsabilityInitiative, my vote would be for something like "VectorAddons" or similar.
-Chad
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
Some people are being somewhat combatant or getting tangential in their responses - fine, basically what I expect form this list - but I'm still very interested in responses to do with my original question, "should we rename it, and if so, what should we name it?".
Please, please, don't rename it again. Every time the name changes things break. They break for us, they break for third parties using SVN, and it breaks our ability to easily support the extensions when people ask for help.
Leave the name alone, and merge it into core as soon as possible.
Respectfully,
Ryan Lane
On 10/12/10 10:29 PM, Ryan Lane wrote:
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
Some people are being somewhat combatant or getting tangential in their responses - fine, basically what I expect form this list - but I'm still very interested in responses to do with my original question, "should we rename it, and if so, what should we name it?".
Please, please, don't rename it again. Every time the name changes things break. They break for us, they break for third parties using SVN, and it breaks our ability to easily support the extensions when people ask for help.
Leave the name alone, and merge it into core as soon as possible.
Respectfully,
Ryan Lane
+1... sounds very reasonable to me.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Trevor Parscal wrote:
On 10/12/10 10:29 PM, Ryan Lane wrote:
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
Some people are being somewhat combatant or getting tangential in their responses - fine, basically what I expect form this list - but I'm still very interested in responses to do with my original question, "should we rename it, and if so, what should we name it?".
Please, please, don't rename it again. Every time the name changes things break. They break for us, they break for third parties using SVN, and it breaks our ability to easily support the extensions when people ask for help.
Leave the name alone, and merge it into core as soon as possible.
+1... sounds very reasonable to me.
I contribute to MediaWiki development mostly through the #mediawiki IRC channel, answering simple questions that users have when they're trying to install or use MediaWiki. Being the most popular MediaWiki installation in the world, people often want to emulate features that they see on Wikipedia.
My frustration from this thread (and from the code review comments) is that you're creating a mess. And you're doing it for the second time.
There was already a skin named "Vector" and a sub-extension named "UsabilityInitiative/Vector". This caused a lot of confusion for a lot of people. It caused confusion when users asked questions and it caused confusion when their questions were answered.
You've now introduced a third concept, a separate extension, called "Vector". Instead of saying "just install MediaWiki and you'll have the basic tools you need," we now have to say "well, MediaWiki comes with the Vector skin, but there's also 'UsabilityInitiative/Vector' and 'Vector' (the extension just called 'Vector') that you may or may not need to install based on the version of MediaWiki you're running."
You don't answer these questions (or any support questions). That's fine, nobody's asking you to. But I do. And in my mind, you're needlessly adding a lot of confusion to what's already not the simplest installation process in the world. And even worse in my mind, you're making the _exact same mistake_ that you've already made. So when you've moved on to other projects, I'll still be trying to explain to new MediaWiki users what the difference is between all of these different concepts called "Vector".
It's bad enough to create a giant mess that everyone else is forced to live with, but please don't waste people's time with discussion for the sake of saying you had a discussion. From your responses to this mailing list thread, it's pretty clear that come hell or high water, you're not going to revert yourself. In the time you've spent discussing this, you could have _easily_ merged this into core and moved on.
I realize you don't respect me and my opinions, but what's absolutely mind-boggling is that plenty of people who you should respect have said the exact same thing as me, and you're still turning a deaf ear. I think your views and actions here are a showcase of the general current Wikimedia Foundation staff approach and attitude toward the community. I don't see how this approach and attitude is in any way acceptable, but I do see how, if it continues, it will ultimately kill the community behind these community-driven projects.
MZMcBride
My take: 1. Vector is a terrible name for the extension, because it overloads terminology. Now when somebody asks about Vector, we won't know if they're talking about the skin or the extra features. The number one thing that we engineers can do to make things simpler for users and administrators is being consistent about having a one-to-one correspondence between terminology and concepts/features. You can see how much of a mess this makes with my own mistakes in LiquidThreads' thread/post terminology confusion. 2. I agree with Ryan that yet another name change is a bad idea. It broke all of my wikis when you separated the extension, and it will break everything again if you rename it. Let's get it into core as soon as possible.
On Wed, Oct 13, 2010 at 1:29 AM, Ryan Lane rlane32@gmail.com wrote:
Please, please, don't rename it again. Every time the name changes things break. They break for us, they break for third parties using SVN, and it breaks our ability to easily support the extensions when people ask for help.
Leave the name alone, and merge it into core as soon as possible.
This. Renames are bad, period. Just look at the issues surrounding trying to standardize extension include files as Foo/Foo.php for the new installer[0]. That being said, I don't *think* we have a lot of people using the Vector extension (WMF sites, Translatewiki?), so a rename in this case shouldn't inconvenience a lot of people. If there is some compelling reason to keep this as an extension, a rename *is* in order, since Vector simply will not do.
I do see that Roan (just now) has replied to the thread indicating his support for merging this stuff to core as well, so I wonder if it's all a moot point now and the extension can go the way of the Dodo.
-Chad
[0] See r65238-40, other revs from siebrand around 17th-18th of April this year
At 2010-10-13 05:10, Trevor Parscal wrote:
On 10/12/10 8:03 PM, Chad wrote:
On Tue, Oct 12, 2010 at 10:45 PM, Trevor Parscaltparscal@wikimedia.org wrote:
On 10/12/10 7:42 PM, MZMcBride wrote:
Trevor Parscal wrote:
Apologies in advance for the sheer triviality of this matter; unfortunately these kinds of bike shed problems [2] tend to be infinitely exciting, while complex matters are more often met with general disinterest.
[1] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/73030#c10054 [2] http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality
Naming and naming conventions aren't trivialities.
Show me the convention for naming extensions and I will follow it. As long as there is none, spending time going back and forth discussing the name of something that users will never see is indeed spending time on a trivial matter.
There isn't one. Maybe it's because we haven't cared until now, or maybe because we haven't needed it. Policy or not, concerns were raised about the naming of *this* extension. Saying "well there's no policy so it's pointless to discuss it" is a cop-out.
I am explicitly requesting discussion about it so that I can take action based on the desire of the community. What cop-out are you referring to?
The only response I've gotten so far is "merge into core", which is an interesting response, but does not resolve the issue at hand, which is, until we do so (assuming we do at some point after 1.17), what should this extension be named?
I'm not defending the current name any more than stating it's my preference. I'm not only open to naming it something else, but I'm asking for input on what to name it.
Some people are being somewhat combatant or getting tangential in their responses - fine, basically what I expect form this list - but I'm still very interested in responses to do with my original question, "should we rename it, and if so, what should we name it?".
IF it should be renamed then I think it has to follow two rules: 1. Should be unique (should not collide with other people ideas) 2. Should tell you a bit about the extension.
In this case "Vector" is wrong. "Vector" might be an extension for SVG editing or something like that. To my understanding this extension is about usability of Vector skin and so it might be called VectorUsability or something more exact.
Having said that I also think Ryan Lane made a very valid point. Personally I'm still using the extension as part of "Usability Initiative" and if it is to be integrated into the core I'm not moving anywhere from that ;-).
Regards, Nux.
2010/10/13 Trevor Parscal tparscal@wikimedia.org:
That's an entirely different discussion, and the results of that discussion have so far been that any such action is being deferred until after 1.17.
With pretty much every participant in this thread (including myself, for the record) saying we should move these extensions into core ASAP (which I interpret to mean before 1.17), I think that statement is outdated at best. I realize you're hesitant to move Vector and WikiEditor into core, let alone do so before the 1.17 release, but I haven't seen anyone share that belief, and frankly I don't think any new commenters will.
Let's just move Vector and WikiEditor into core soon and be done with it.
Roan Kattouw (Catrope)
At 2010-10-13 09:12, Roan Kattouw wrote:
2010/10/13 Trevor Parscaltparscal@wikimedia.org:
That's an entirely different discussion, and the results of that discussion have so far been that any such action is being deferred until after 1.17.
With pretty much every participant in this thread (including myself, for the record) saying we should move these extensions into core ASAP (which I interpret to mean before 1.17), I think that statement is outdated at best. I realize you're hesitant to move Vector and WikiEditor into core, let alone do so before the 1.17 release, but I haven't seen anyone share that belief, and frankly I don't think any new commenters will.
Let's just move Vector and WikiEditor into core soon and be done with it.
WikiEditor is a different thing at current stage. It breaks WYSIWYG made from FCKeditor and don't give anything important in exchange. Vector extension is ready to work while WikiEditor still needs a lot of work if you ask me.
Regards, Nux.
Thank you, everyone, for responding so far (not trying to stop you here). Here's where it seems we're at.
1. Having an extension called "Vector" is neither descriptive or clear, and it is anticipated to cause confusion. 2. System administrators are not enjoying things being switched around on them, and would prefer the name either not change or for the extension to be merged into core. 3. 3 out of 5 code reviewers are voting that it should be merged into core (Brion and Tim have not weighed in yet).
My original hesitation to merge Vector (not to be lumped together with WikiEditor) into core was always that we were going to be extremely lean on CodeReview prior to 1.17 due to Tim Starling's limited availability right now. Since then we've added Roan, Chad, and me, and even brought Brion back for a bit part time to help out, thus increasing our capacity beyond what was originally expected. I think under these new circumstances, if we have Roan and Chad willing to help me integrate it into core and sign off on it, it will likely be of very little additional effort for our release manager (Tim Starling) to sign off as well, thus alleviating my apprehension about merging it prior to 1.17.
So, unless there's any objection, I will go ahead and merge the Vector extension into core as part of the Vector skin immediately, rather than waiting until after 1.17.
- Trevor
On 10/13/10 12:38 AM, Maciej Jaros wrote:
At 2010-10-13 09:12, Roan Kattouw wrote:
2010/10/13 Trevor Parscaltparscal@wikimedia.org:
That's an entirely different discussion, and the results of that discussion have so far been that any such action is being deferred until after 1.17.
With pretty much every participant in this thread (including myself, for the record) saying we should move these extensions into core ASAP (which I interpret to mean before 1.17), I think that statement is outdated at best. I realize you're hesitant to move Vector and WikiEditor into core, let alone do so before the 1.17 release, but I haven't seen anyone share that belief, and frankly I don't think any new commenters will.
Let's just move Vector and WikiEditor into core soon and be done with it.
WikiEditor is a different thing at current stage. It breaks WYSIWYG made from FCKeditor and don't give anything important in exchange. Vector extension is ready to work while WikiEditor still needs a lot of work if you ask me.
Regards, Nux.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 13 October 2010 10:54, Trevor Parscal tparscal@wikimedia.org wrote:
3. 3 out of 5 code reviewers are voting that it should be merged into core (Brion and Tim have not weighed in yet).
Does it matter if one is one of the six persons that were lately assigned to do code review or not? Did the earlier announcement suddenly give more weight to their opinions? In my opinion things like this does not help in getting other people (volunteers) to help with the code review, which is something Danese was planning to do - unless I remember incorrectly.
I've previously indicated my support of merging the working parts of the Vector enhancements to core. Sure it is more code for the new release and the merge can cause bugs, but on the other hand those really seem to be useful to lots of people. I also think that confused sysadmins are less happy than sysadmins who have to do changes but are not so confused. It takes more time to figure out what is what and why something is not working than do the actual changes.
-Niklas
On 2010-10-13, Trevor Parscal wrote:
Thank you, everyone, for responding so far (not trying to stop you here). Here's where it seems we're at.
- Having an extension called "Vector" is neither descriptive or clear, and it is anticipated to cause confusion.
- System administrators are not enjoying things being switched around on them, and would prefer the name either not change or for the extension to be merged into core.
- 3 out of 5 code reviewers are voting that it should be merged into core (Brion and Tim have not weighed in yet).
I did not realise decisions were now being voted on by code reviewers, this is very concerning.
My original hesitation to merge Vector (not to be lumped together with WikiEditor) into core was always that we were going to be extremely lean on CodeReview prior to 1.17 due to Tim Starling's limited availability right now. Since then we've added Roan, Chad, and me, and even brought Brion back for a bit part time to help out, thus increasing our capacity beyond what was originally expected. I think under these new circumstances, if we have Roan and Chad willing to help me integrate it into core and sign off on it, it will likely be of very little additional effort for our release manager (Tim Starling) to sign off as well, thus alleviating my apprehension about merging it prior to 1.17.
Since it has been deployed to the Wikimedia cluster, has it not already had more than sufficient review that we would be comfortable including it in a release?
So, unless there's any objection, I will go ahead and merge the Vector extension into core as part of the Vector skin immediately, rather than waiting until after 1.17.
That would be great.
Robert
On Wed, Oct 13, 2010 at 9:00 PM, Robert Leverington robert@rhl.me.uk wrote:
On 2010-10-13, Trevor Parscal wrote:
Thank you, everyone, for responding so far (not trying to stop you here). Here's where it seems we're at.
1. Having an extension called "Vector" is neither descriptive or clear, and it is anticipated to cause confusion. 2. System administrators are not enjoying things being switched around on them, and would prefer the name either not change or for the extension to be merged into core. 3. 3 out of 5 code reviewers are voting that it should be merged into core (Brion and Tim have not weighed in yet).
I did not realise decisions were now being voted on by code reviewers, this is very concerning.
I think it's "pick on Trevor day" :-). Can we settle down and assume that he's saying that lots of people agree with it, and is using this as an example of the weight carried by some the people who are in agreement? Nobody has made any objection, so there is no indication whatsoever that "code reviewers" are overruling anybody else.
On 10/13/10 3:00 AM, Robert Leverington wrote:
On 2010-10-13, Trevor Parscal wrote:
Thank you, everyone, for responding so far (not trying to stop you here). Here's where it seems we're at.
1. Having an extension called "Vector" is neither descriptive or clear, and it is anticipated to cause confusion. 2. System administrators are not enjoying things being switched around on them, and would prefer the name either not change or for the extension to be merged into core. 3. 3 out of 5 code reviewers are voting that it should be merged into core (Brion and Tim have not weighed in yet).I did not realise decisions were now being voted on by code reviewers, this is very concerning.
Read the next paragraph, I am simply stating that there is willingness to support the review of the integration. Perhaps "voting" was a bad choice of word, but I think if you take it in the context of the entire conversation it's clear I'm not literally asking for a few people to vote to determine the final decision.
My original hesitation to merge Vector (not to be lumped together with WikiEditor) into core was always that we were going to be extremely lean on CodeReview prior to 1.17 due to Tim Starling's limited availability right now. Since then we've added Roan, Chad, and me, and even brought Brion back for a bit part time to help out, thus increasing our capacity beyond what was originally expected. I think under these new circumstances, if we have Roan and Chad willing to help me integrate it into core and sign off on it, it will likely be of very little additional effort for our release manager (Tim Starling) to sign off as well, thus alleviating my apprehension about merging it prior to 1.17.
Since it has been deployed to the Wikimedia cluster, has it not already had more than sufficient review that we would be comfortable including it in a release?
Well, breaking it up into lots of little pieces and spreading it out all over the code base (the current reality of integrating an extension into core) is going to require review, adding complexity to the release.
- Trevor
So, unless there's any objection, I will go ahead and merge the Vector extension into core as part of the Vector skin immediately, rather than waiting until after 1.17.
That would be great.
Robert
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Op 13 okt 2010, om 09:54 heeft Trevor Parscal het volgende geschreven:
So, unless there's any objection, I will go ahead and merge the Vector extension into core as part of the Vector skin immediately, rather than waiting until after 1.17.
- Trevor
I agree aswell. From the installs of MediaWiki I did so far this year for websites, pretty much all of 'em are specifically requesting the "Wikipedia software" (because asuming) that MediaWiki is what Wikipedia runs. Which, by all means, is true.
However, installing it in such a way that the end result is comparable in terms of features is yet another. (for example the element ID change that mismatched between the UsabilityInitiative-extension and the Vector-skin (#mw-panel etc.)).
The UsabilityInitiative has worked hard on making a new look and feel for MediaWiki. So I think it makes sense that both the PHP backend, HTML-structure and CSS, aswell as the JavaScript involved are being made available as one "Vector", as supposed to seperated.
Also, for what it's worth, I personally found the name "UsabilityInitiative" as Extension partially confusing aswell. Since the UsabilityInitiative "team" made the Vector "skin" aswell, which was coded in the core from the beginning. Why some part got split up as 'extending' was unclear to me.
Looking forward to 1.17 :-)
-- Krinkle
2010/10/13 Maciej Jaros egil@wp.pl:
WikiEditor is a different thing at current stage. It breaks WYSIWYG made from FCKeditor and don't give anything important in exchange. Vector extension is ready to work while WikiEditor still needs a lot of work if you ask me.
Good point. I was gonna suggest moving the toolbar at least, but doing so without moving the rest of WikiEditor doesn't make much sense. If we make the WikiEditor default config sane so the unstable features are disabled by default (Trevor may already have done this, I don't remember) I guess telling people to install WikiEditor after which the toolbar will Just Work for them is fine.
Roan Kattouw (Catrope)
wikitech-l@lists.wikimedia.org