Process, harvesting, getting your favorite things in the image

Daniel Vainsencher danielv at netvision.net.il
Mon Mar 10 08:39:27 UTC 2003


Adam Spitz <adspitz at yahoo.ca> wrote:
> The Harvesters can't afford to be this loose. They've
> got a much bigger code base to worry about, and the
> consequences of mistakes are higher. Naturally they
> want to institute a more heavyweight process to help
> ease their load. 
This is precisely true, but there's another thing - because it's more
code, the feedback process is too long. Beyond a certain time-lapse
between change and feedback, the feedback isn't strong enough a force to
keep the code clean and fit long term. I think we're way past that point
at the current state. Bringing the feedback into a much earlier stage of
the cycle (before us harvesters have to look at the code) is meant to
mitigate this problem. This is an additional reason for the change,
separate from "consequences" of a specific change.

Of course, a single, small focused package, can be administered more
effectively with less process.

> An even better solution, though, is
> just for the rest of us to take as much of the code as
> we can and split it apart from the image and take
> responsibility for it ourselves.
Perfectly true. 
 
> Lately I decided that I'd like to use Celeste for my
> e-mail. I'm sure that I'll end up wanting to make lots
> of little changes to it, and I don't particularly like
> the idea of going through this new Harvesting process
> to get those changes approved, so I've been looking
> into making a removal change set for Celeste. After
> that, I'm hoping that someone more qualified than me
> will take stewardship of it, so that I can send my
> change requests to that person - but if not, then I'm
> willing to take it myself.
Wonderful, you don't need to write such a script, just help me out with
the one that I have made. Actually, as removal from the image of well
integrated stuff is likely to often do, it includes a bit of redesign,
not just a script. It was released initially for 3.3a, and rereleased a
couple of months ago again for 3.4. You're quite welcome to test it on a
current image. I've heard claims it doesn't work, any help tracking down
what the problem is, would be welcome.

Daniel



More information about the Squeak-dev mailing list