[GLLUG] Subversion & Bug Tracking Panel

[Ed Thomson]

On Mar 1, 2007, at 10:51 PM, Richard Houser wrote:
>
> I would strongly recommend getting at least a mention of darcs in the
> presentation.  The distinguishing trait of darcs is that it is  
> built on
> formal patch theory and decentralized control.

I would suggest that a survey of all available version control  
systems sounds, at best, like it deviates from the topic.  That is -  
if I went to a discussion entitled "subversion and bug tracking", and  
I started hearing about the differences between git and darcs, I have  
little doubt that I'd get up and leave.

Decentralized version control tools are probably interesting for the  
casual subversion user - an overview of how they work and how they  
differ from subversion seems reasonable.  And if there is to be an  
example, git does seem like the obvious choice for a group primarily  
made of Linux users, since that's what the kernel's using.

> I would classify the biggest downside to darcs as limited potential  
> for
> code contribution (who here knows formal patch theory?).

I don't think that's darcs' biggest barrier to usage -- I'd say that  
the people who want to write (or contribute to) version control  
software is a pretty infinitesimal subset of the people who want to  
use version control software.

Nor do I think it's darcs' biggest barrier to contribution --  
conflict resolution (be it called "application of patch theory",  
"conflict resolution", "merging", etc) is pretty well understood in  
the field, is a comparatively small part of version control software,  
and is relatively static -- I wouldn't think that new contributors to  
the project are jumping in to that part of the code.  I would think  
the fact that it's written in Haskell would be the barrier for entry  
to development.

-Ed