Re: Linux: a gentle, growing approach
On Mon, 2002-10-14 at 06:24, ben wrote:
> On Monday 14 October 2002 12:33 am, David Pastern wrote:
> [big snip]
>
> dave, i'll give you a couple of points due to the fact that the subject
> header says linux, and not debian. but then, given that it's a debian list, i
> have to take back, at least, one of them. in fact, given that you'd have a
> hard time imposing on this list for any other kind of help than with debian,
> maybe i should take back the other, as well. i'm not arguing against tools,
> but against tools that never really met the goals they once aspired to
> achieve. as martin intimated, once aptitude takes precedence, we can easily
> and happily dispense with the rest. in fact, i find aptitude, even in its
> present form, to be a far more useful tool than either dselect or tasksel,
> not that any of them come near the efficiency of apt and dpkg.
>
> as far as accommodating those who want a nice easy operating system is
> concerned, debian ain't it to begin with. i'm damn near willing to bet that
> of all the users on this list, less than one percent started with and stayed
> with debian as their sole experience of linux. it's not a beginners distro.
> on top of that, don't even try to exhort the developers on the basis of
> accommodating the average user. not only will their likely response make you
> hop from your seat way down there in the antipodes, but most other users will
> be totally on their side. don't ever forget, we get the best operating system
> that ever existed, along with the best support that ever existed for
> absolutely nothing, or, failing nothing, then for, at most, $5 or $10 for the
> install cd's.
>
I'd be one of that less than one percent, but even I'm not *really* the
image of those you are concerned with - I worked with BSD 4.2 a couple
decades ago, after writing and porting computer language interpretters
and compilers for Watcom. I know Unix, I know multi-user system design,
and I know how to install an o/s on virtually virgin core metal ;)
That said, I could never get Slackware through an installation process
when I first tried. It was the first six attempts. Red Hat, I couldn't
get working boot floppies made. Debian, I must admit I chose it for
three reasons:
1) It was the only distribution left on the InfoMagic Linux Developers'
Resource 6 CD set.
2) I liked considerably what it said about being focussed so strongly on
license compliance - knowing that I'd be working at some code, I wanted
to avoid license hiccups (the code development went bye-bye with the
dot-com bust)
3) This was mistaken, but I read the reasoning for the name, thought
"Aww, that's sweet!", and erroneously reasoned that Debian must be
focussed more on being a "Core, keep from being confusing, reliable
distribution you could expect most spouses to be able to work with." -
maybe the spouse doesn't need a doctorate in software engineering, but I
think sysadmining with Debian requires a *bit* more than "I can use Word
on an MS box some tech has already installed and configured."
[***SNIP of allegory that people should have Clue #1 about what they are dealing with***]
>
> while dselect and tasksel may serve the ends of inexperienced users, it's
> nonetheless true that they still suck. i see no harm in their abandonment,
> since those who survive their absence are that much less likely to show up on
> the list asking the kind of questions that almost inevitably elicit responses
> in the fashion of rtfb. as i mentioned before, debian really isn't designed
> with a mind to being a linux starter kit, whatever about linux from scratch.
> by the time anyone gets here with a view to succeeding, they tend to know
> what they're up against. personally, i think apt and dpkg are all anyone
> should need, particularly for any of those who've busted their balls dealing
> with yast and rpm and, especially, that weird-ass halfway to nowhere
> installation aid offered by mandrake.
I'd offer here that there are different perspectives at play with this
discussion, possibly derived from some degree of a muddy-ing of what
Debian is. Debian, how can I best describe this, is something akin to a
fusion of a research project with a political movement. Debian GNU/Linux
is not so much a product, as it is a research project and political
statement combined - that software and technology can be developed and
coordinated in an bazaar environment to produce a stable, capable and
reliable system as represented by Debian GNU/Linux 3.0 (aka Woody).
Related "reports" wrt Hurd and BSD will hopefully be coming in the
nearer rather than further future, but in that regard, the impetus is
not to make Debian wea£thy, but to assert research and political points
- software in the public interest, for instance, as well as the
effectiveness of free software and open source.
There is the *nice* side effect of having such a solid, powerful,
complete, and immensely inexpensive computing platform - one that has
attracted the favour of not only us, but from repeated studies, a solid
plurality of Linux developers, and Debian does work very hard (and
remember, it is all voluntary) to address the concerns involved in
applying this project to everyday considerations.
The concern here becomes "how far does the *research* aspect need to
proceed to consistently *prove* the theory, and how far is a politically
needed to be pursued. Debian Jr., Debian Medical and the audio editing
platform distribution, as well as the Hurd and BSD efforts may lean more
to the research side at present, while I find some aspects of the
GNU/Linux efforts are migrating into more political concerns - no longer
"Can it be done?" but rather "How should it be done, and why it should
be done."
That raises the question of how far does it make sense to take Debian
GNU/Linux into being a *platform for every system* - does it get ported
to every processor that meets the basic needs of the Linux kernel? Does
it get polished to the point that people can use it without particular
training? Does it get tuned to not need sysadministration on the most
common platforms? There are apparently numerous supporters of each view
among the developers, who collective constitute the research team, as
well as political advocates who want to see all accomplished, either on
the strong belief in free software above all else, the opposition to
strictly proprietary approaches (or one strictly proprietary vendor), or
even an opposition to the commercial distribution of anything developed
under the GPL - in which case Red Hat, SuSe, Mandrake and SCO are in the
same sights as Microsoft.
>From a Debian-User list perspective, there is a gentle preference for
whatever causes fewer headaches for users, but we must keep in mind that
what constitutes success for Debian is not the same as it is for Red Hat
et al. How to define success is a moving target, as the research and
political priorities shift to reflect the changes in the environment
from each perspective. However, I would suggest that whatever happens,
as users we are in for an interesting ride from which we can only
benefit - through better resources, increased knowledge of the tools we
are using, and the increased viability of Linux, free software and open
source software in general.
>
> you've said that you're relatively new to linux, so i give you back those two
> points for being gutsy enough to persevere with debian in the early days of
> your experience, but i'm pretty sure that, one of these days, perhaps in a
> year, you'll be feeling pretty much the same about those tools as they are
> now. even given martin's observation that tasksel serves the ends of those
> requiring a quickly installed bootstrap system, i'm also pretty damn sure
> that it doesn't take much time for the administrators of those systems to
> wean themselves of that dependency.
>
> just a rant. don't be offended.
>
> ben
>
>
> --
> To UNSUBSCRIBE, email to debian-user-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
>
>
--
Mark L. Kahnt, FLMI/M, ALHC, HIA, AIAA, ACS, MHP
ML Kahnt New Markets Consulting
Tel: (613) 531-8684 / (613) 539-0935
Email: kahnt@hosehead.dyndns.org
Reply to: