Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 3 Jul 2008 20:18:41 -0400
From:      Mike Meyer <mwm-keyword-freebsdhackers2.e313df@mired.org>
To:        Michel Talon <talon@lpthe.jussieu.fr>
Cc:        freebsd-hackers@freebsd.org
Subject:   Re: Sysinstall is still inadequate after all of these years
Message-ID:  <20080703201841.2b0b955b@bhuda.mired.org>
In-Reply-To: <20080703212100.GA16598@lpthe.jussieu.fr>
References:  <20080703212100.GA16598@lpthe.jussieu.fr>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 3 Jul 2008 23:21:00 +0200
Michel Talon <talon@lpthe.jussieu.fr> wrote:
> evolve easily. The argument that there sould be no external dependency
> seems to me inspired by the NIH syndrom.

I think your seeming is wrong. I believe it's inspired by the belief
that the base system should be self-replicating: you should be able to
build the distribution with the base install and sources.

While I think that kind of self-replicating ability is a good goal,
having an X interface pretty much kills it - you have to have have the
X headers to compile against, if nothing else. I think this makes the
wanting to use a modern HLL more palatable.

However, once you tie part of the release process to some external
tool, you create extra headaches for the ports folks. You can no
longer update the port without verifying that you didn't just break
the release process. This means updating the port will take longer -
which will make users who like FreeBSD because critical software in
the ports tree tends to stay up to date unhappy. Maybe you need to
maintain two versions of the port? Not pretty if they're actually
based on the same release of the language.

Or course, you could move the interpreter into the base system, but
we've been there before, and it was really ugly (and I'm there now on
clients GNU/Linux systems, and it's *still* really ugly).  Come to
think of it, someone claimed that (some of) the problems with the
installer come from using an ancient version of dialog, so it seems
that part of the system is *still* there.

On the other hand, if you're going to do this, replacing the forth in
the boot process would seem to be a suitable second objective.

    <mike
-- 
Mike Meyer <mwm@mired.org>		http://www.mired.org/consulting.html
Independent Network/Unix/Perforce consultant, email for more information.

O< ascii ribbon campaign - stop html mail - www.asciiribbon.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080703201841.2b0b955b>