Re: RH and GNOME
SEGV <mlepage@cgocable.net> wrote:
> re:
> - Raster's ego
Irrelevant. Mine's bigger ;^)
> - not changing a gdkimlib #define to an enum for bogus reasons
Irrelevant. Why waste time to achieve binary compatibility for an
unstable API? [Raster lists imlib as "beta", but it's still the core
of enlightenment which he lists as "development".]
> - not integrating other changes and patches
Since he's in the process of rewriting the whole thing from scratch,
and has a lot of missing code (for example: hard coded constants
instead of currently unwritten routines), this is largely a matter
of taste.
Note that he seems very interested in having serious help, but
minor patching is down in the noise threshold at the moment.
> - not increasing the major version number even though compatibility is broken
He thought it wouldn't be. Bad call.
The software has been released at this point, there's no way to
recall it.
> All this just makes me think a little less of Raster and Mandrake.
Good for you. You going to help them out?
> I do understand that they write interesting (good?) code. I understand
> that E is in a state of high flux during development, and that they
> probably have bigger worries than "little things."
Yeah, they have mentioned that once or twice. Almost every week.
> Sure, I feel the same way when I develop. But, I always try to do
> "the Right Thing." If that means eschewing the preprocessor or making
> something more configurable, I will. Especially if others need that
> functionality to do their work.
Actually, a lot of the current practices are in reaction to
criticisms they've received on the e-develop mailing list.
> I haven't read "The Humble Programmer," but I do subscribe to its
> ideas. I can take constructive criticism of my code. I document its
> shortcomings. I improve it based on others' input. And I wonder
> sometimes why others get sore when I criticize their work, rightly.
>
> I just wish more Linux hackers were a little more "professional."
> There, that's my rant.
They've been taking a lot of criticism, and have been trying to deal with
it. 0.13 (from before the rewrite) had lots of patches supplied by
other people, and lots of stability work. 0.14 is just too early in
the design cycle for that kind of work to be meaningful.
--
Raul
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: