Re: vote

[Erich Schubert <erich.schubert@mucl.de>]

> All I'm trying to do is save you folks the trouble
> of wasting your time doing something, well, stupid.  Things
> that would be really stupid are:
> 
> -Using ext2fs as your filesystem
> -Using GNU libc
> 
> And, in my opinion, the following aren't wise decisions either:
> -Going through and removing every BSD tool with a GNU replacement
> 
> The first isn't an intelligent decision because a strong advantage
> of BSD over Linux is the filesystem.  You'd be throwing that away
> if you used ext2fs.

First thing is that i (but i'm no insider) just don't believe that the
difference between UFS and ext2 is so much; if it were, i believe that
Linux hacker's would already have implemented big UFS support in the
Linux kernel as well, just to get that better fs.
Instead they appear to prefer "wasting" their time with reiserfs, jfs,
xfs, ext3... (i have bad experiences with reiserfs...)
But most user's will want to run Debian/BSD as well as Debian/Linux.
So we'll need ext2 support, too.
But maybe not in the beginning... this can be put on the todo-if-stable
list. Might be the other way round as well:
As UFS support is still "experimental" (it's read-only, isn't it?)
people might prefer to use ext2 as their root fs for Debian/BSD - so
they can boot their Linux and access the data from there as well.
Sure, you could install *BSD as well to your existing Debian Linux...
It's just that people might prefer.

> The second isn't that smart because BSD libc is 1.  Already working
> with the BSD kernel, 2. Proven- major version (read: incompatible) bumps
> are much more rare (I was using Debian Linux when the libc5->glibc
> conversion occured, and it was painful), and 3. If you need glibc
> is some incantation, for whatever reason, it's available under
> linux compat.

i was using Debian as well when the libc5 -> glibc migration was done,
and i cannot complain.
But i believe that people will _expect_ to run glibc "by default".
They expect a "Debian" System, which means "as compatible to
Debian/Linux as possible" - so i will expect all my programs (unless
they are kernel-specific) to compile in the exactly same way, to behave
in the same way etc.

> I don't think the third is smart only because it seems that it is
> a lot of effort with very little results.

And this is valid for all core utilities as well.
(a lot of them are the bsd versions anyway...)
but i HATE it when a script expects the GNU variant of some tool finds
the BSD variant which has different options and thus does not work.
This might be as little as i'm used to "ps ax", and i don't want to
check what system i am on, and if i have to type "ps -ef".

> As for the worthfulness of this project in general, I do understand
> why some folks may want to see dpkg/apt (because they are comfortable
> and do not want to learn the BSD package system) - and I think
> that would be a worthwhile thing to accomplish.

Debian/BSD should NOT become FreeBSD+dpkg.
i think it's intended to become Debian with a BSD kernel.
But still 99.9% Debian.

> You do have to plan before you act...

But you can plan all your life just to see it doesn't work out ;)
And you can overdesign and over-plan a lot... ;)

Greetings,
Erich