On Aug 8, 2007, at 4:18 AM, Karan Malhi wrote:
Thanks a lot for committing the patches. Timely commits are such a great help to me, otherwise i have to constantly check to make sure that the current change I am making does not conflict with any pending patches , which is a big PITA.
Definitely a PITA. I did big amout of work for XStream last year and they weren't able to look at that first patch and I kept going.. then they were still too busy for that second patch.. and the third patch... and the fourth patch. It just didn't work out mostly because we weren't connecting early in the process and my work ultimately didn't line up with how they would have wanted the features implemented.
It isn't having the patches in svn in a timely manner that's the biggest help (though definitely nice), it's getting the immediate feedback and growing the code organically as a team ensuring all work is relevant that's the real magic.
On the other end of the spectrum, often times people want to hold changes till they "get them right" and weeks go by and suddenly one big code drop happens. By that point things are way beyond the opportunity for iterative (collaborative) development and it's often very difficult to absorb the changes -- usually an all or nothing situation. It can be much worse too if there was a lot of high level talk/discussion from the person doing the "in a corner" development as they feel like there's an agreement and the code will be accepted with no significant changes; "i spent all this time doing it this way cause you said that's what you wanted!" Talking with other people yet still keeping the code to yourself (either because you're afraid of judgment, don't want to bother people with too many patches, or because you don't want people messing with "your stuff") ultimately *is not* collaborative development. This issue can happen with contributors and committers alike. Whenever I know someone is working on some big changes and aren't letting them out for weeks on end, I get very nervous about the pending train wreak.
You've been utterly fantastic about submitting small, manageable patches and it's very easy to bounce ideas back and forth, try stuff out in code, see how it looks, learn from it and add more ideas on top of it. We can tweak it here and there and all the while we're working on the same code *in svn*. I'm just thrilled. Everyone who wants to contribute to open source should take a page from your book.
The shorter the time between proposal, patch, and commit the better. -David
