Re: Deficiencies in Debian
gregor herrmann <gregor+debian@comodo.priv.at> writes:
> On Fri, 18 May 2007 22:09:56 -0700, Russ Allbery wrote:
>> And one of the best things about the Linux model is that Linus
>> regularly talks about how he wants things done and what leads him to
>> take stuff or not take stuff in public on the lists, which leads others
>> to do the same. And those are interactive discussions, not just
>> writeups. I think people learn a lot from those discussions.
> - open discussions about future developments
Here, it wasn't as much future developments that I was thinking of as more
basic issues, like style and the thought processes behind why the kernel
is structured the way it is. Linus does a great job of explaining his
sense of taste, which is sort of a meta-level above future development.
I think the Debian Policy discussions, if we can kick up the level of
activity, could partly serve a similar role within Debian. There are also
some Debian developers (Steve Langasek and Manoj Srivastava come to mind)
who regularly follow up to threads on debian-devel and explain both their
aesthetic judgement and how they arrived at that conclusion.
IMO, one of the most valuable skills for someone working in IT is to have
a well-developed aesthetic sense of what a clean and supportable system
looks like. Most of the day-to-day decisions that I make are based on a
sense of aesthetics more than specific technical criteria. That's the
form that my subconscious gestalt of systems takes. My experience is that
once one has developed that sense of aesthetics and learned to look
closely at anything that feels "ugly", it becomes surprisingly effective
at pointing directly at the weak parts of any design.
>> The main obstacle that I see is that that stuff takes a lot of time. I
>> spend probably 5% of my work time on the coordination, record-keeping,
>> and reporting parts of that sort of activity, which in a full-time job
>> is quite reasonable. But it's not really a percentage; it's a quantity
>> of time that those activities take. And I couldn't take a similar two
>> hour per week cut out of my Debian volunteer work without a much
>> greater impact on how much stuff I could get done.
> Sure, mentoring/training/staff development takes time but as you point
> out at the beginning it probably "bear[s] fruit for Debian". Maybe
> Debian would be better off in the long run if some of the experienced
> DDs decided to drop one package or resign from one infrastructure task
> and to use the saved time for taking an "apprentice".
Probably true. Although two hours a week is way more than just one
typical package. That's probably the total time I spend on lintian, for
example, or about five or ten regular packages where I'm just packaging
upstream releases. (Not that I'm an experienced DD; I'm still fairly
new.)
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: