Direction of Pidgin development

Luke Schierer lschiere at pidgin.im
Sat May 5 17:58:14 EDT 2007


On Sat, May 05, 2007 at 04:45:46PM -0400, Andrew Roeder wrote:
> 
> Personally the problem i'm seeing with the plugin system is that, Pidgin 
> comes with a load of default plugins already, if they are going to always be 
> distributed with even the source of pidgin, or packaged in its default 
> installer, they might as well not be plugins, obviously they're important 
> enough to be designed into Pidgin's preferences instead of thrown into the 
> "plugins" section.

We distribute plugins not because of their import or lack of import, but
because of their maintainer or lack of maintainer.  Additionally, we
*necessarily* have to provide some plugins, because users are always
requesting examples. 

But to return to the main thrust of this email, the prime, and really
*only* necessary element for a plugin to be in the tarball is that one
of us either have written it or be willing to maintain it.  When that
ceases to be the case, we look at removing it from the tarball.  Plugins
compile by default or not based on whether or not we think them
generally useful, or generally confusing.  But useful or confusing, they
are only there because they are *ours,* and not some other person's.

The reasoning behind this is simple.  There are tons of plugins out
there that are useful to some subset of the userbase.  However, the
project's history has shown us that 9 times out of 10, when we accept a
plugin, its author will feel less motivated to help with it, expecting
us to take on a greater and greater percentage of the work involved in
maintaining it.  Were we to accept 3rd party plugins with any
regularity, our work on Pidgin & libpurple themselves would be slowed
significantly, as we spend our time maintaining the resulting cloud of
plugins. 

For the same reason, new protocols have recently only been added as
summer of code projects, with the hope that the student would continue
on with the project after the summer ends.  After all, the summer of
code is supposed to be an introduction to open source, not a one time
experiment in it.

luke




More information about the Devel mailing list