[sylpheed:34057] Re: Improving Sylpheed development process.

Ricardo Mones mones at debian.org
Wed Jun 9 22:55:07 JST 2010

On Wed, Jun 09, 2010 at 12:05:22PM +0200, Antonio Ospite wrote:
> Hi,
> We all like Sylpheed as users, it's fast, clean, lightweight, but
> as developers we must admit it's not the most dynamic Open Source
> project around here.
> If I had to describe Sylpheed development model I'd say it is basically
> Hiro-centric, which is not at all bad /per se/, but I feel this
> sometimes can limit external contribution.

  It's the same model of Linux kernel in their beginning, which IMHO
  demostrates the model is not a limiting factor for contributions.

> So if we manage to refresh the development, still keeping Hiro as
> the one who makes final decisions, I think our favorite MUA can improve
> even further.

  Err... I guess that's an opinion.

> Let me list the things I feel can be improved and please let me know
> yours as developers, telling also if some change can make you contribute
> _more_ to Sylpheed.
> Have a bugtracker:
>   Bugs reports and feature requests are lost in the mailing list,
>   unorganized and often uncommented, this makes difficult to classify,
>   remember and finally address them.

  I tend to agree with this, but Hiroyuki seems to prefer the list and
  Debian already has its BTS, so... :)

> Use a modern (distributed[1]) SCM, as subversion is a little clumsy to
> my eyes:
>   - this will make people make their own experiments off-line much more
>     easily, and then propose them to the Sylpheed developer[s] in a
>     clean and consistent way, this can be done already some how, but
>     there is also the next reason:
>   - the SCM should differentiate between the author of a change and the
>     actual committer: if I contribute a change to Sylpheed I want that
>     reflected in the project history, this is not only desirable in
>     order to give credit, but also useful to assign responsibility to
>     code changes.

  Fail to see what prevents you from doing a svn checkout and experiment.
  And the second reason is just not applicable to this model, it's only one
  committer which gives credit in ChangeLog if he's not the author... even
  with multiple committers is still manageable, see Claws Mail ChangeLog.

> This can motivate external contributions IMHO, at the end of the day
> Open Source is about _people_.

  Really? Well, _anything_ is about people if you put it that way ;-)

> Let me know if you agree on what are the weak spots, so we can start to
> talk about how to address them. Of course this is under the assumption
> that there are (or there will be) people willing to contribute to
> Sylpheed :) If you are one of them, please speak up.

  That assupmtion is too optimistic! IMO Sylpheed is already good enough
  for most of its users (even the new coming ones), so they don't feel
  the need to contribute.
  It has bugs, of course, but much less than Windows, and if they can
  live with Windows why not with Sylpheed? :)
  Ricardo Mones 
  Never send a human to do a machine's job.               Agent Smith

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
Url : http://www.sraoss.jp/pipermail/sylpheed/attachments/20100609/d4d0e1b4/attachment.bin 

More information about the Sylpheed mailing list