I have formulated a policy for Wikipedia development here:
http://meta.wikipedia.org/wiki/Development_policy
In short: * We now have a "stable" branch. All Wikipedias should henceforth be updated from that branch. It is currently identical to the main (unstable) branch, but I'll take some untested features out in the next days.
* NO MORE MANUAL OVERWRITING OF FILES ON TEST.WIKI. Sorry for shouting, but this really has to stop. If our Test-Wiki is supposed to be useful, it has to be in a consistent state. That means it should only be updated from CVS unstable.
* A basic testing procedure. Features will have to run on test.wiki for 3-10 days before they are committed to stable.
With this policy in place, it should become much more feasible to keep all the Wikipedias up to date. We still need someone to finish the conversion script, though -- Italian is still on UseMod, for example, they really should be switched over.
Regards,
Erik
Erik Moeller wrote:
I have formulated a policy for Wikipedia development here: http://meta.wikipedia.org/wiki/Development_policy
[...]
With this policy in place, it should become much more feasible to keep all the Wikipedias up to date. We still need someone to finish the conversion script, though -- Italian is still on UseMod, for example, they really should be switched over.
Sorry to ask, but where is the discussion about that policy? Have all delelopers agreed to this policy? Who count's as developer and had to agree to that policy?
Unfortunatly I sometimes feel like a developer (only if I add something to the CVS of course), but I din't feel like one as I got this mail. Are you not usually complaining that every change etc. has to be done via democtratic wiki decicion?
I think a proper poliy is ok, and it should be written down somewhere, even in more than one language IMHO. But should not be a little discussion about it, somewhere? Why not added that every change has to be documentated, especialy new features? Why there aren't any coding rules specified (at leased somethin like: no more than 78 chars/line)? Where discussions aout new features take place and where there have to be an archive about all this discussions? And I'm sure there are other things missing.
Sometimes you get the feeling that all the beautiful rules in wikipedia don't count a cent if they hit reality.
Smurf
Thomas Corell wrote:
Erik Moeller wrote:
I have formulated a policy for Wikipedia development here: http://meta.wikipedia.org/wiki/Development_policy
Sorry to ask, but where is the discussion about that policy?
Right here! :)
Have all delelopers agreed to this policy?
Me!
Who count's as developer and had to agree to that policy?
Well, if you don't, speak up.
Why there aren't any coding rules specified (at leased somethin like: no more than 78 chars/line)?
If you'd like to write up a codification of our current code style, please do. Some of the basics:
function wfGlobalFunctionName( $notethespace, $anotherparam ) { # <-- indent with tabs global $wgSomeGlobalVar; doMyStuff( $wgSomeGlobalVar, $anotherparam ); }
class SomeClass { var $mMemberVariable; # Use get/setters if possible # very very short function can be on one line like this: function getMemberVariable() { return $this->mMemberVariable; } }
I would however counsel against the sort of cryptic local variable names in much of the older code, or the unnecessary duplication of code that should be refactored into functions or classes that can be shared.
Where discussions aout new features take place and where there have to be an archive about all this discussions?
This mailing list is archived.
And I'm sure there are other things missing.
Sometimes you get the feeling that all the beautiful rules in wikipedia don't count a cent if they hit reality.
Bah!
-- brion vibber (brion @ pobox.com)
Brion Vibber wrote:
Right here! :)
Well I think I have missed some mails. I only got the 'I done it'-Mail by Erik.
Have all delelopers agreed to this policy?
Me!
You are _all_ developers? Fine, than I stop my 'developing' right now.
Who count's as developer and had to agree to that policy?
Well, if you don't, speak up.
If only you are a developer, I'm not affacted, so why speak up?
Where discussions aout new features take place and where there have to be an archive about all this discussions?
This mailing list is archived.
I know that! But why not wrote it down, for all the others? And to make it clear how the work is coordinated.
Smurf
PS: this is not meant as offence, I only wanted to point that e.g. there may be other developers around -- not only Erik and you. And some developer may read your and/or Eriks mail as a insult. I'm to old to get realy agitated about such mails, but I'm not sure of others.
On Mon, 7 Jul 2003, Thomas Corell wrote:
Brion Vibber wrote:
Right here! :)
Well I think I have missed some mails. I only got the 'I done it'-Mail by Erik.
Right, that's the beginning of the discussion.
Have all delelopers agreed to this policy?
Me!
You are _all_ developers? Fine, than I stop my 'developing' right now.
I am one of the developers, saying that I agree with Erik's suggestion.
Who count's as developer and had to agree to that policy?
Well, if you don't, speak up.
If only you are a developer, I'm not affacted, so why speak up?
Straw man.
Where discussions aout new features take place and where there have to be an archive about all this discussions?
This mailing list is archived.
I know that! But why not wrote it down, for all the others? And to make it clear how the work is coordinated.
That's what Erik's suggestion is all about!
PS: this is not meant as offence, I only wanted to point that e.g. there may be other developers around -- not only Erik and you. And some developer may read your and/or Eriks mail as a insult. I'm to old to get realy agitated about such mails, but I'm not sure of others.
Then I think there is a misunderstanding.
-- brion vibber (brion @ pobox.com)
Smurf wrote in part:
Brion Vibber wrote:
Smurf wrote:
Have all delelopers agreed to this policy?
Me!
You are _all_ developers? Fine, than I stop my 'developing' right now.
I wondered when would be the right time to jump in and try to keep people from taking offence unnecessarily. Perhaps this is that time.
Brion definitely wasn't claiming to be all developers here. He was simply listing himself as /a/ developer that had agreed. The other developers now need to give their opinions, as you did and as he has now done.
That's why Brion wrote this:
Who count's as developer and had to agree to that policy?
Well, if you don't, speak up.
And please do work these things out on [[meta:]]! Wikis are great for more than just writing encyclopaedias, you know. It's one thing to edit a policy page on the sly to win an argument elsewhere, but it's perfectly OK to hash out /new/ policies on policy pages that are being watched closely by many people. Like editing encyclopaedia articles, you must act in good faith, but you need not act always in accordance with my own desires.
-- Toby
On Mon, Jul 07, 2003 at 08:34:07AM -0700, Brion Vibber wrote:
Have all delelopers agreed to this policy?
Me!
Sure that separation of stable and unstable is nice, but if test suddenly becomes, well, real code testing place, shouldn't we make some test2 just for playing with random code not necessarily meant for usage in real Wikipedias in predictable future ?
Why there aren't any coding rules specified (at leased somethin like: no more than 78 chars/line)?
If you'd like to write up a codification of our current code style, please do. Some of the basics:
function wfGlobalFunctionName( $notethespace, $anotherparam ) { # <-- indent with tabs global $wgSomeGlobalVar; doMyStuff( $wgSomeGlobalVar, $anotherparam ); }
class SomeClass { var $mMemberVariable; # Use get/setters if possible # very very short function can be on one line like this: function getMemberVariable() { return $this->mMemberVariable; } }
I would however counsel against the sort of cryptic local variable names in much of the older code, or the unnecessary duplication of code that should be refactored into functions or classes that can be shared.
Why not just: "The code should follow whatever is considered to be good practice, unless there are reasons to do it some other way". Wikipedia software isn't any special, so it doesn't need any special coding rules. And 78 chars / line was important 10 years ago, but is seriously too small for year 2003. Come on, we're all using hardware that easily allows over 120 chars per line, not ?
Tomasz Wegrzanowski wrote:
Why not just: "The code should follow whatever is considered to be good practice, unless there are reasons to do it some other way". Wikipedia software isn't any special, so it doesn't need any special coding rules. And 78 chars / line was important 10 years ago, but is seriously too small for year 2003. Come on, we're all using hardware that easily allows over 120 chars per line, not ?
I like sometimes to print a piece of code. Without reformating it. And of course, it was only an example. Nothing we realy have to do.
I'm more interested in things like those I mailed some weeks ago: http://mail.wikipedia.org/pipermail/wikitech-l/2003-June/004376.html (still unanswered, of course ;)
Or the policy should clear the topic of a ''code freeze'', which is never lifted, but nobody seems to remember that.
All this are proposals, of course. And I certainly don't start editing the policy on meta whitout a few comments from others here.
Thomas-
I like sometimes to print a piece of code. Without reformating it.
But that's what code formatters are for. I'm no fan of hard line breaks -- I have a reasonably large screen, and with all the whitespace that's necessary for well formatted code, I need to make use of it in order to make basic constructs visible. IMHO this kind of formatting should be done dynamically as needed.
I'm more interested in things like those I mailed some weeks ago: http://mail.wikipedia.org/pipermail/wikitech-l/2003-June/004376.html (still unanswered, of course ;)
The general answer to questions like this is "just do it". Use your own best judgment.
Or the policy should clear the topic of a ''code freeze'', which is never lifted, but nobody seems to remember that.
Well, I do remember it, and what Lee originally wrote is http://marc.theaimsgroup.com/?l=wikitech-l&m=105366788222813&w=2 "I'd like to strongly suggest that we put the software into feature freeze for at least a week until all the wikis are fully functional with the latest software, and they can all be updated from the scripts in the distribution."
We all observed the code freeze for a week as Lee suggested, but nobody did the necessary updates. Making updates easier is one of the reasons behind the new development policy.
Regards,
Erik
Tomasz said:
Sure that separation of stable and unstable is nice, but if test suddenly becomes, well, real code testing place, shouldn't we make some test2 just
for
playing with random code not necessarily meant for usage in real
Wikipedias
in predictable future ?
Hey, that's a good idea, why didn't I think of that? Oh, that's right, I did:
http://test.wikipedia.org/w/tim/wiki.phtml
(actually it wasn't my idea, there was already a directory there called "magnus-version" and I took my cue from that)
Why not just: "The code should follow whatever is considered to be good practice, unless there are reasons to do it some other way". Wikipedia software isn't any special, so it doesn't need any special coding rules. And 78 chars / line was important 10 years ago, but is
seriously
too small for year 2003. Come on, we're all using hardware that easily allows over 120 chars per line, not ?
And Erik said:
Thomas-
I like sometimes to print a piece of code. Without reformating it.
But that's what code formatters are for. I'm no fan of hard line breaks -- I have a reasonably large screen, and with all the whitespace that's necessary for well formatted code, I need to make use of it in order to make basic constructs visible. IMHO this kind of formatting should be done dynamically as needed.
I was once scarred by having to convert ~5000 lines of code to an 84 column limit by hand so that it could be printed, so now I tend to use lots of line breaks. But if anyone declares that some other style is the "Wikipedia standard", I'll be happy to conform to it.
-- Tim Starling ua.ude.bleminu.scisyhp@gnilrats.t
Thomas-
Sorry to ask, but where is the discussion about that policy? Have all delelopers agreed to this policy? Who count's as developer and had to agree to that policy?
apologies for the surprise mail. Brion and I discussed this before and it was time to simply implement this in order to move forward. I used the wiki to formulate the policy so that other aspects could be easily added and agreed upon. If you object to a specific part, just mention or change it in the wiki.
I don't want a vote on everything -- never did, never will. Votes are useful for agreeing on *controversial* matters. But if a matter is uncontroversial, surely there is no need to vote on it? Even if it is controversial, it is usually advisable to seek compromise and consensus first. All our existing wiki policies were initially started like this one (with the difference that Larry would hit you over the head if you disagreed). It's an open process.
Regards,
Erik
On Mon, 7 Jul 2003, Erik Moeller wrote:
apologies for the surprise mail. Brion and I discussed this before and it was time to simply implement this in order to move forward.
To keep the cabal-watchers from worrying, this prior discussion consisted of Erik mentioning to me in IM that he thought we should have stable & unstable branches in CVS and me agreeing. We did _not_, I repeat, _not_ secretly plan to round up the other developers in a fake truce meeting where we would use poison gas released from the scale model of the development tree to kill off our competition. Nothing like that even came up. No sir, on my honor.
-- brion vibber (brion @ pobox.com)
Brion-
On Mon, 7 Jul 2003, Erik Moeller wrote:
apologies for the surprise mail. Brion and I discussed this before and it was time to simply implement this in order to move forward.
To keep the cabal-watchers from worrying, this prior discussion consisted of Erik mentioning to me in IM that he thought we should have stable & unstable branches in CVS and me agreeing. We did _not_, I repeat, _not_ secretly plan to round up the other developers in a fake truce meeting where we would use poison gas released from the scale model of the development tree to kill off our competition. Nothing like that even came up.
Of course not. Absolutely not. Never would we plan such a dastardly thing.
On a different note, we may want to discuss these issues in further detail at a kind of, shall we say, "truce meeting" to alleviate all the remaining corps^Wconcerns that there may be. I know a nice dark back alley that would be perfect for the occasion ..
Regards,
Erik
"Erik Moeller" erik_moeller@gmx.de wrote in message news:8pOypDtSpVB@erik_moeller...
I have formulated a policy for Wikipedia development here:
http://meta.wikipedia.org/wiki/Development_policy
In short:
- We now have a "stable" branch. All Wikipedias should henceforth be
updated from that branch. It is currently identical to the main (unstable) branch, but I'll take some untested features out in the next days.
- NO MORE MANUAL OVERWRITING OF FILES ON TEST.WIKI. Sorry for shouting,
but this really has to stop. If our Test-Wiki is supposed to be useful, it has to be in a consistent state. That means it should only be updated from CVS unstable.
- A basic testing procedure. Features will have to run on test.wiki for
3-10 days before they are committed to stable.
With this policy in place, it should become much more feasible to keep all the Wikipedias up to date. We still need someone to finish the conversion script, though -- Italian is still on UseMod, for example, they really should be switched over.
Regards,
Erik
This is all great. I was thinking much the same thing myself. CVS and the test server are much too chaotic at the moment. I was actually thinking of a stricter development cycle, aimed at producing a series of numbered stable releases, but I guess continuous development could be okay too, as long as there are no serious bottlenecks in getting features from conception to installation.
-- Tim Starling.
Erik Moeller wrote:
We still need someone to finish the conversion script, though -- Italian is still on UseMod, for example, they really should be switched over.
Note that while I'd like to get this done at some point, I'd be quite happy if someone else took it over and actually _got_ it done, as I'm mostly mired with general maintenance at the moment and just don't have the energy to finish this up which will just generate more maintenance needs... ;)
The half-finished script is maintenance/archives/importUseModWiki.php
-- brion vibber (brion @ pobox.com)
"Brion Vibber" brion@pobox.com wrote in message news:3F0993B7.8040706@pobox.com...
Erik Moeller wrote:
We still need someone to finish the conversion script, though -- Italian is still on UseMod, for example, they really should be switched over.
Note that while I'd like to get this done at some point, I'd be quite happy if someone else took it over and actually _got_ it done, as I'm mostly mired with general maintenance at the moment and just don't have the energy to finish this up which will just generate more maintenance needs... ;)
The half-finished script is maintenance/archives/importUseModWiki.php
I would think about doing some work on this, but I don't really know what's involved. I don't know much about the other language Wikipedias, because I don't have a pliny account.
-- Tim Starling.
Erik Moeller wrote:
- NO MORE MANUAL OVERWRITING OF FILES ON TEST.WIKI. Sorry for shouting,
but this really has to stop. If our Test-Wiki is supposed to be useful, it has to be in a consistent state. That means it should only be updated from CVS unstable.
sorry to ask dumb questions, but what is the correct procedure for putting files on the test-server?
wikitech-l@lists.wikimedia.org