[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: "dselect" replacement team

Martin Schulze:
> On Apr 13, Ian Jackson wrote
> > If you can't understand the code as written then you shouldn't be
> > hacking it, because you're not good enough.
> > 
> > You certainly shouldn't even _think_ of replacing it.
> Ian, please don't be that conservative.  If no one should be
> allowed to change anything in the code, and this is what I read
> in your most recent mails, then Debian can't be based on dpkg
> anymore.

My apologies for being unclear.  I didn't mean to suggest that nothing
should be changed, or even that we shouldn't replace anything.

I meant to say two things in that posting:

1. You shouldn't try to modify or replace something until you
understand it.  dpkg is complicated because its job is complicated,
and attempts to hack it or a replacement by people who are unable or
unwilling to deal with that complexity will fail.

2. dpkg (as opposed to dselect) doesn't seem to me to have reached
critical mass.  It's still small enough to be understood by one
person, and modulo a few irregularities its code is still clean and in
correspondance with its current functionality.  I think the design
decisions at the heart of it still apply.  Therefore, I think that
dpkg should not yet be replaced.  dselect, on the other hand, is in
clear and desperate need of a rewrite.

Rob Browning:
> Well, the problem is that you don't seem to have the time (not that I
> can blame you) for these improvements.  It would be really nice if we
> eventually have several people capable of understanding, maintaining,
> and improving the tools so we don't have a single point of failure.

Yes, and I'm very happy to see other people getting stuck into the
code.  It's those who say `I don't understand this so I must rewrite
it' that worry me the most.  I don't think there's anything magic
about me in this respect; all it takes is a good programmer with
enough free time to get to grips with it.

Anyway, you can have me back in 3 weeks (whether you want me or

> It might be worthwhile for you to discuss your ideas (like the fixes
> you mention above) in public now and then so that people get a better
> idea of the issues involved, the complexities concerned, and won't be
> as likely to go off half-cocked, or waste time coming up with
> solutions for problems where you've already found a solution.

I do this when I find the time.  Unfortunately writing such messages
(and answering the resulting questions) takes a lot more time than
simply saying `this is complex'.  I think it's better for me to try
not to burn myself out now while I'm writing up; from May the 1st I'll
be employed with a normal day job and I'll have enough time to do more
than just say what I think ought to be done.


Reply to: