Re: Kwalitee metric: Community support channels
On 7/26/06, Shlomi Fish <[EMAIL PROTECTED]> wrote: There is some software for multiplexing between a web forum, a newsgroup, a mailing list and an RSS feed, which could be useful. But we need to consider whether we also want a forum (a la Gabor's http://www.cpanforum.com/ ) as well. I wonder if there's anyway I can become automatically subscribed to all the distributions I've ever maintained on cpanforum.com? That would be cool. Gabor, can you shed some light on this issue? You can do it now: http://www.cpanforum.com/rss/author/PAUSEID is the rss feed of the recent posts related to a module under the given PAUSEID One can also subscribe to get e-mail alrerts for her modules based on PAUSEID. Gabor
Re: Kwalitee metric: Community support channels
Shlomi Fish wrote:
On Monday 24 July 2006 16:23, Salve J Nilsen wrote:
Which specific types of channels one should get points for may warrant
discussion, but if our goal is the improvement of the software, we should
at least encourage a mininmum number of ways to reach the users and
developers of a software project.
I would suggest giving a point for explicitly (and in a consistently
machine-readable manner) stating the project's...
[bugtracker]
[public mailing list]
[...and it's searchable archive]
* publically readable code repository (e.g. to a CVSWeb or SVN::Web
frontpage URL)
H... would a standard Subversion HTTP/S tree be enough?
Sure, main the purpose of such a resource (for the general community at least)
would be to allow easy access to the current source code to allow relevant bug
reporting and patch creation. The standard SVN web service is more than good
enough for this purpose, IMO. :)
"Instant" communication channels like IRC and IM can of course be useful,
but since the chat logs usually aren't stored and indexed publically,
their lon term usefulness for the community are somewhat limited.
True, but I solved many problems using IRC or at least got a lot of help.
I love using IRC too, and as a way to get "instantaneously" in touch with devs
and users, I think it's great. Therefore, I think it's very cool when the devs
of some project say they can be reached through some IRC channel. But is IRC a
community feature we feel is important/necessary enough to give a point for in
"the game"?
I think IRC is extremely convenient and definitely worth a point, but I also
think that having a code repository, a bugtracker and a mailing list with a
searchable archive is MUCH more useful for a project. Is there a way to make
such a distinction in "the game"?
The rest of us ("the CPAN/Perl community") can still get all the good
stuff, in addition to some hints on which projects one shouldn't expect
any improvements or support. :)
Yes.
I daresay that sometimes a simple forward or developer email address is
enough as a contact address. Recently I encountered some people in Israel
(relatively new to the Internet scene) who seem to dislike mailing lists and
prefer web forums and other mediums. Some of them even complained that some
relatively low volume mailing lists were high volume, while in fact they
were less than p5p and perl6-language, and much less than BugTraq or the
Linux Kernel Mailing List.
Heh... If such users manage to create a web forum for a project, all kudos to
them! Pointing out the existence of a webforum in the META.yml file would of
course be useful when that's the main channel for discussion.
There is some software for multiplexing between a web forum, a newsgroup, a
mailing list and an RSS feed, which could be useful.
Gmane is such a tool, and I've talked with Lars Magne Ingebrigtsen (the
website's main developer) about the possibility of making it into a publically
available tool. Although he appreciated the idea he said that the software
was pretty much custom made and too complex (not "packagable enough") to be
made into something anyone can install for their community. :-(
But we need to consider whether we also want a forum (a la Gabor's
http://www.cpanforum.com/ ) as well. I wonder if there's anyway I can become
automatically subscribed to all the distributions I've ever maintained on
cpanforum.com? That would be cool. Gabor, can you shed some light on this
issue?
Hehe, that would be cool. Having a META.yml field where the author can state
she's subscribing to the distribution's CPAN forum, at which some part of the
forum software can update it's subscription lists based on this information. :)
e.g. author_subscribes_to_cpanforum: yes
That would be neat! :)
- Salve, dreaming of blue skies again. :-P
--
Salve J. Nilsen / Systems Developer
Norwegian Meteorological Institute http://met.no/
Information Technology Department / Section for Development
Re: Kwalitee metric: Community support channels
Hi Salve!
See below for my comments.
On Monday 24 July 2006 16:23, Salve J Nilsen wrote:
> Shlomi Fish wrote:
> > On Wednesday 19 July 2006 17:08, Salve J Nilsen wrote:
> >> Just a wild thought...
> >>
> >> Would it be useful to check for references to community support channels
> >> like mailing lists, IRC channels, public bug trackers and official web
> >> pages?
> >
> > Interesting idea. One thing I should probably note is that ESR has this
> > recommendation in "The Cathedral and the Bazaar":
>
> [http://catb.org/esr/writings/cathedral-bazaar/cathedral-bazaar/ar01s10.htm
>l]
>
> [Shlomi's experiences using different community channels]
>
> > What did I want to say? Yes, often the scope or maturity of the module
> > does not justify a special "community" support channels. So I'm not sure
> > whether penalising CPAN distros for not having this information is a good
> > idea. But I'll have to think about it some more.
>
> I'd rather look at these metrics as a way of encouraging developers to
> think about issues around community sustainment. That way, we can use "the
> game" as a tool for software improvement in addition to improving the
> codebase.
Sure, that's probably good.
> Which specific types of channels one should get points for may
> warrant discussion, but if our goal is the improvement of the software, we
> should at least encourage a mininmum number of ways to reach the users and
> developers of a software project.
>
> I would suggest giving a point for explicitly (and in a consistently
> machine-readable manner) stating the project's...
>
> * primary public bugtracker (frontpage URL) in use by it's users and
> developers
Well, we have rt.cpan.org for free. I believe module-starter and friends can
put it in the YAML by default, while allowing you to override with something
else (in case you have your own different bug tracker, as is the case for
Catalyst with their Trac.).
+1.
> * main public mailing list (subscription URL) in use by it's
> users and developers
Well sometimes people segment between the mailing list for developers and
mailing list for users of the software. I was involved with the Subversion
development for a while, and they had one common mailing list. Then they
decided to separate it into [EMAIL PROTECTED] and
[EMAIL PROTECTED]
I should also note that there are other types of mailing lists:
1. Mailing list for Version Control Commits.
2. Wine-style Licence/flames mailing list. (Just kidding).
3. Mailing list for individual components (Mozilla style madness, where I can
never determine where to post something).
4. Etc.
> * publically searchable archive of the mailing list
> (search page URL)
Well, Google as well as mail-archive.com, yahoogroups, googlegroups etc. give
you an archive and a search for free. The archive should be publicly
accessible, and have some search functionality. I set up a htdig search for
the entire Perl Mongers domain, and it was a pretty straightforward
experience.
> * publically readable code repository (e.g. to a CVSWeb
> or SVN::Web frontpage URL)
H... would a standard Subversion HTTP/S tree be enough?
>
> "Instant" communication channels like IRC and IM can of course be useful,
> but since the chat logs usually aren't stored and indexed publically, their
> lon term usefulness for the community are somewhat limited.
True, but I solved many problems using IRC or at least got a lot of help.
>
> One could of course say distros that don't state ANY contact information or
> community support channels could be "penalized", but I'd guess these
> developers probably don't care enough about their software or "the game" to
> feel much penalty from losing those points.
Yes.
>
> The rest of us ("the CPAN/Perl community") can still get all the good
> stuff, in addition to some hints on which projects one shouldn't expect any
> improvements or support. :)
>
Yes.
I daresay that sometimes a simple forward or developer email address is enough
as a contact address. Recently I encountered some people in Israel
(relatively new to the Internet scene) who seem to dislike mailing lists and
prefer web forums and other mediums. Some of them even complained that some
relatively low volume mailing lists were high volume, while in fact they were
less than p5p and perl6-language, and much less than BugTraq or the Linux
Kernel Mailing List.
There is some software for multiplexing between a web forum, a newsgroup, a
mailing list and an RSS feed, which could be useful. But we need to consider
whether we also want a forum (a la Gabor's http://www.cpanforum.com/ ) as
well. I wonder if there's anyway I can become automatically subscribed to all
the distributions I've ever maintained on cpanforum.com? That would be cool.
Gabor, can you shed some light on this issue?
Regards,
Shlomi Fish
-
Shlomi Fish [EMAIL PROTECTED]
Homepage:http:
Re: Kwalitee metric: Community support channels
Shlomi Fish wrote:
On Wednesday 19 July 2006 17:08, Salve J Nilsen wrote:
Just a wild thought...
Would it be useful to check for references to community support channels
like mailing lists, IRC channels, public bug trackers and official web
pages?
Interesting idea. One thing I should probably note is that ESR has this
recommendation in "The Cathedral and the Bazaar":
[http://catb.org/esr/writings/cathedral-bazaar/cathedral-bazaar/ar01s10.html]
[Shlomi's experiences using different community channels]
What did I want to say? Yes, often the scope or maturity of the module does
not justify a special "community" support channels. So I'm not sure whether
penalising CPAN distros for not having this information is a good idea. But
I'll have to think about it some more.
I'd rather look at these metrics as a way of encouraging developers to think
about issues around community sustainment. That way, we can use "the game" as a
tool for software improvement in addition to improving the codebase. Which
specific types of channels one should get points for may warrant discussion,
but if our goal is the improvement of the software, we should at least
encourage a mininmum number of ways to reach the users and developers of a
software project.
I would suggest giving a point for explicitly (and in a consistently
machine-readable manner) stating the project's...
* primary public bugtracker (frontpage URL) in use by it's users and developers
* main public mailing list (subscription URL) in use by it's users and
developers
* publically searchable archive of the mailing list (search page URL)
* publically readable code repository (e.g. to a CVSWeb or SVN::Web frontpage
URL)
"Instant" communication channels like IRC and IM can of course be useful, but
since the chat logs usually aren't stored and indexed publically, their lon
term usefulness for the community are somewhat limited.
One could of course say distros that don't state ANY contact information or
community support channels could be "penalized", but I'd guess these developers
probably don't care enough about their software or "the game" to feel much
penalty from losing those points.
The rest of us ("the CPAN/Perl community") can still get all the good stuff, in
addition to some hints on which projects one shouldn't expect any improvements
or support. :)
- Salve
--
Salve J. Nilsen / Systems Developer
Norwegian Meteorological Institute http://met.no/
Information Technology Department / Section for Development
Re: Kwalitee metric: Community support channels
On Wednesday 19 July 2006 17:08, Salve J Nilsen wrote: > Just a wild thought... > > Would it be useful to check for references to community support channels > like mailing lists, IRC channels, public bug trackers and official web > pages? > Interesting idea. One thing I should probably note is that ESR has this recommendation in "The Cathedral and the Bazaar": << It's fairly clear that one cannot code from the ground up in bazaar style [IN]. One can test, debug and improve in bazaar style, but it would be very hard to originate a project in bazaar mode. Linus didn't try it. I didn't either. Your nascent developer community needs to have something runnable and testable to play with. When you start community-building, what you need to be able to present is a plausible promise. Your program doesn't have to work particularly well. It can be crude, buggy, incomplete, and poorly documented. What it must not fail to do is (a) run, and (b) convince potential co-developers that it can be evolved into something really neat in the foreseeable future. http://www.catb.org/esr/writings/cathedral-bazaar/cathedral-bazaar/ar01s10.html >> I can attest from my experience that my most successful projects started from some code I wrote for myself, for some reason or another, and then finalised and released to the public. I usually didn't start a mailing list right away for them, and instead just publicised them on Freshmeat and collected the mails I received regarding them in my inbox, until there was enough to form a mailing list. Some of them still don't have a mailing list. In regards to IRC channels: I tend to avoid starting one IRC channel for every limited-scope Perl module I put on the CPAN. There are enough channels I'm trying to participate on Freenode as it is, and usually cannot concentrate on one or two channels at a time. I feel that such modules should be discussed on channels of larger scope like #perl. (Albeit larger scale projects such as POE, Catalyst/Jifty/etc. Perl-QA, etc. may justify their own channels). For example, LeoNerd and I have coordinated the development of Error.pm on #perl. Note that #perlcafe (on Freenode) is intended as #perl's chat channel, where one can also "banish" discussions that tend to flood #perl and have too little interest on the other #perl participants and lurkers. Another note is that I'm often not on the IRC on my waking hours, because I'm finding it too distracting and addictive. People can always reach me on IM or on Email: http://www.shlomifish.org/me/contact-me/ What did I want to say? Yes, often the scope or maturity of the module does not justify a special "community" support channels. So I'm not sure whether penalising CPAN distros for not having this information is a good idea. But I'll have to think about it some more. Regards, Shlomi Fish - Shlomi Fish [EMAIL PROTECTED] Homepage:http://www.shlomifish.org/ Chuck Norris wrote a complete Perl 6 implementation in a day but then destroyed all evidence with his bare hands, so no one will know his secrets.
Re: Kwalitee metric: Community support channels
Thomas Klausner wrote: Hi! On Thu, Jul 20, 2006 at 01:56:36PM +0200, Salve J Nilsen wrote: Is there a (public) authoritative META.yml spec describing required, recommended and supported fields? http://module-build.sourceforge.net/META-spec.html Thanks, and it seems there are newer versions of this too. http://module-build.sourceforge.net/META-spec-current.html Seems some of the stuff I've been ranting about already is in the spec. :) - Salve
Re: Kwalitee metric: Community support channels
Salve J Nilsen wrote: Adam Kennedy wrote: The presence of lack thereof is more an indication of the scale and importance of the module, rather than anything you can judge all 10k modules by. I'd rather interpret the presence/lack of community pointers as an indication of how interested that community is in attracting new users, helping them use the software, or even help them take part actively in the software development. Why for $DEITY's sake wouldn't one want to make it as easy as possible for people to join in on the fun? Standardizing community Hmf. Too quick on the send button, sorry. Standardizing a way to pinpoint software community resources may have several beneficial effects, including giving users pointers on project viability (software with an active community behind it is more desirable than one without it), increase the probability for people to give feedback (comments, bug reports, feature ideas, patches) and make it more likely for a community to reach "smallest critical mass" (make it less likely for people to _not_ join a community because "there are too few people in the community".) The search.cpan.org webpages have already some community support features (link to rt.perl.org and annocpan), why not make it possible for the authors to add and improve that information? :) - Salve
Re: Kwalitee metric: Community support channels
Hi!
On Thu, Jul 20, 2006 at 01:56:36PM +0200, Salve J Nilsen wrote:
> Is there a (public) authoritative META.yml spec describing required,
> recommended and supported fields?
http://module-build.sourceforge.net/META-spec.html
--
#!/usr/bin/perl http://domm.zsi.at
for(ref bless{},just'another'perl'hacker){s-:+-$"-g&&print$_.$/}
Re: Kwalitee metric: Community support channels
Adam Kennedy wrote: Salve J Nilsen wrote: Just a wild thought... Would it be useful to check for references to community support channels like mailing lists, IRC channels, public bug trackers and official web pages? One way to do this could be to look for relevant keywords in the META.yml file or to do simple scanning of a SUPPORT section in the POD... Is this feasible? Not really. At an implementation level, META.yml would need support for that sort of things. That doesn't seem too difficult to me (sorry 'bout the pseudo-YAML). --- #YAML:1.0 support: irc: irc://irc.freenode.net/perl-qa bugtracker: http://rt.cpan.org/NoAuth/Bugs.html?Dist=Module-Starter webpage: N/A webforum: N/A blog: N/A list-subscribe: http://lists.cpan.org/showlist.cgi?name=perl-qa faq: N/A coderepository: cvs: N/A subversion: N/A mailarchive: http://www.mail-archive.com/perl-qa@perl.org/ annotated_docs: http://annocpan.org/dist/Module-Starter Is there a (public) authoritative META.yml spec describing required, recommended and supported fields? If so, what does it take to extend the spec to include community information? At a purely social level, not every module needs to have IRC channels and official web pages. Sure, not everyone needs one, and certainly not their own. But if there exists some channel where one could reliably reach the author(s), wouldn't it be "nice to know"? (Personally, I think it's more than "nice to know" - it's more like a "requirement for open source projects to prosper".) The presence of lack thereof is more an indication of the scale and importance of the module, rather than anything you can judge all 10k modules by. I'd rather interpret the presence/lack of community pointers as an indication of how interested that community is in attracting new users, helping them use the software, or even help them take part actively in the software development. Why for $DEITY's sake wouldn't one want to make it as easy as possible for people to join in on the fun? Standardizing community If at all possible, I'd like to see any new CPANTS metrics being about things that are well studied, and that are fairly universally agreed to be applicable to all 10,000 modules. I'd also love to see the metrics used as a tool for improving the community around the software... Having ambitions to improve the general code "quality" in CPAN is very fine, but since that is mostly a social "game", I think it would be a good idea to also help the game sustain itself by making it easy for new participants to join and help improve the code. - Salve
Re: Kwalitee metric: Community support channels
Salve J Nilsen wrote: Just a wild thought... Would it be useful to check for references to community support channels like mailing lists, IRC channels, public bug trackers and official web pages? One way to do this could be to look for relevant keywords in the META.yml file or to do simple scanning of a SUPPORT section in the POD... Is this feasible? Not really. At an implementation level, META.yml would need support for that sort of things. At a purely social level, not every module needs to have IRC channels and official web pages. The presence of lack thereof is more an indication of the scale and importance of the module, rather than anything you can judge all 10k modules by. If at all possible, I'd like to see any new CPANTS metrics being about things that are well studied, and that are fairly universally agreed to be applicable to all 10,000 modules. Adam K
Kwalitee metric: Community support channels
Just a wild thought... Would it be useful to check for references to community support channels like mailing lists, IRC channels, public bug trackers and official web pages? One way to do this could be to look for relevant keywords in the META.yml file or to do simple scanning of a SUPPORT section in the POD... Is this feasible? - Salve
