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

Re: "dselect" replacement team

> Here's what I've heard from your lead coder so far:
>  - He has said that he thinks the essential component of the Debian
> system should be rewritten from the ground up.

I have no preference nor objection to this simply because it is a
rewrite.  If they want to make use of existing code, great.  If the
existing code is a mess of "hacks", then I think a rewrite is a
good thing for future management of the code.

Keep in mind that, while most programmers like working on projects,
actual maintenence gets done because it has to (not because anyone
wants to do it).  Thus, the easier you can make future maintenence,
the better the software will hold up in the long run.

>  - He has said he thinks it should be done in a language that isn't
> guaranteed to exist on platforms to which we might want to port
> Debian.

If it gets written in G++, then it is available on any platform that can
compile gcc.

>  - He has complained that he finds the existing source hard to
> understand, when, for better or worse, it's all the documentation that
> exists.

And the relevance of this is...?

>  - He has posted wild allegations that dpkg is still written in perl,
> without apparently taking the time to look at the makefile and/or do
> `file /usr/bin/dpkg`

I wouldn't call it a "wild allegation".  I don't recall him saying how
terrible the thing was because it was written in perl.  He was simply
looking at the souce of a much-older version by mistake.  Hardly a
stoning offence.

> Things I would have hoped to have heard, but haven't:
>  - Discussion of creating a test suite for the code, so that we can
> assure ourselves that new features work as advertised (thus avoiding
> another epochs fiasco), and that changes to the code don't break
> existing behavior.

Perhaps a nice comment to the list suggesting (politely) that a test suite
would be a good idea.  You might even be told (politely) that such an
idea had already came up and was _not_ (despite what you seem to think)
ruthlessely put down.

> Tell me why I shouldn't be worried.  Show me you're being sensitive to
> the issues of bootstrapping Debian to new systems.  Show me a lead
> coder who examines the existing code before making grand statements of
> intent.  Show me an intent to produce a solid system.

What do you want to know?  I will be happy to tell you what I know and
listen to your ideas and suggestions.  My primary goal is to create
a system that makes installation and maintenence of a Debian system
an easy task for both novice and expert.

Bootstrapping new Debian systems is minor in comparison.  The existing
dpkg is sufficient for that.

                                 ( bcwhite@verisim.com )

    In theory, theory and practice are the same.  In practice, they're not.

Reply to: