[isf-wifidog] Roadmap and Project Manager

[Benoit Grégoire]

On January 3, 2006 05:01 am, Max Horváth wrote:
> Hi,
>
> in the last few weeks we've all been discussing about the things
> WiFiDog lacks about.
>
> One thing I'm sure we currently miss the most is an active project
> manager and of course one of his most important responsibilities to
> create: a roadmap.

Unlike others, I had no vacation this Christmas while everyone else suddently 
seemed to have heaps of time to work on wifidog.  I even excused myself 
before christmas.  Bashing me publically is not really the best way to start 
an intelligent conversation about it.

> Just an example of my situation as a developer. I have a lot of ideas
> WiFiDog could need as improvements. Sure we have a bug and feature list.
>
> But what we really need is a roadmap so developers among us could
> concentrate on specific things to work on step by step.

As was mentionned before (several times) the roadmap is obtained by searching 
the bugs and RFEs by group:  

-For 1.0
-Post 1.0
-Long term

This would be done automatically if we had a more decent developper support 
platform.  Considering the fact that we are apparently going to go trough a 
third wiki transition anyway, and that we are having problems with 
sourceforge's CVS, I think we should take this opportunity to move to SVN and 
an integrated wiki and task manager like Trac. 
 
> So - any input and/or ideas?

-I think the lack of a unified design document hurts us much more than the 
fact that the roadmap is hard to use.  Unfortunately, I am not even sure the 
few design-like documents I posted to the mailing list over the years after 
reaching a consensus after hours of arguing on mailing lists or in person 
were even carefully read by most people.  It takes hours to agree on 
relatively simple but important things such as the Smarty integration level 
desired.  This is somewhat normal, but getting a consensus on a full design 
document that people will actually honor and respect would be weeks of pretty 
much fulltime work for the coordinator.  I don't think anyone has that kind 
of time to commit to just writing.

Right now (well, up till a month ago), I spent hours every week on wifidog 
related things, almost 100% of it being invisible project management of some 
sort.  No time left for coding and actual work, this is just no fun.  
Unfortunately, every organisational change I proposed to implement simple OSS 
best practices and make international developement easier have meet moderate 
to heavy resistance locally.  To put it bluntly, there is little point in 
doing project management if people won't either do what the manager proposes 
or come up with and defend an alternative.  I am emotionally tired of it, and 
I don't intend to attempt it again untill people seem to be more open to it.  
It may just be that the project doesn't want or need a singular project 
manager (I certainly don't think it's workable at this stage).  At least a 
lot of people seem to independently recognise the need for change.

But here are a few communication related things that are easy and that I think 
should be priorities to help the project in the short term:

-I've been very liberal with commit access to lower the barrier of entry.  
However, if you want to commit a change that isn't a bugfix or simple 
improvement, it is probably best to check on the mailing list if it has been 
discussed before.
-Someone should make sure that the doxygen documentation for the auth server 
actually builds, and should create a nightly corntab entry to build it from 
CVS and put it up on the web.  There is a ton of design-related documentation 
in there.
-Use the IRC  channel (#ilesansfil on freenode, or we can create a #wifidog 
one if it makes anyone feel better).  We used it exensively when we coded the 
gateway, and it helped smooth out cummunication massively.  We are the only 
OSS project of significance I ever heard of that actively resists this idea.  
Hopefully I don't have to explain why this is almost essential.
-It may be more on a psychological thing, but I think wifidog should have it's 
own wiki.  People are just too shy to ask for permissions on the ISF one.
-We need a better and really simple task and milestone manager like Trac (I 
don't really care which one we pick, but we should pick one).
-Now that max has proposed coding standards, and no one opposed, we shold 
respect it.
-We should make the mailing list archives searchable.


-- 
Benoit Grégoire, http://benoitg.coeus.ca/
-------------- section suivante --------------
Une pièce jointe non texte a été nettoyée...
Nom: non disponible
Type: application/pgp-signature
Taille: 189 octets
Desc: non disponible
Url: http://listes.ilesansfil.org/pipermail/wifidog/attachments/20060103/b9bd70c1/attachment-0001.pgp