gentoo - wow!! - progress

[dep]

begin  Collins's  quote:
| Also, I though moving kde and gnome away from /opt
| used to be a prime requirement of the FHS.  When did that change?

it never was the case. claims that it was were based on a spectacular 
misreading of the fhs. those who misread it took it to mean that 
anything the distributor provides must go into /usr, which is not 
true. the governing principle is the file structure, with items in 
/opt encapsulated in their own directories, while items in /usr must 
*not* be (with the *lone* exception of xfree86, which the fhs 
specifies is granted that exception because it's too late to turn 
back now).

| Since the current market for PCs is 1G+ (now pushing 2G+) with
| minimum 128Meg (often 256 Meg), it don't see this as a deal
| breaker.  Sorry, I'm not interested in linux on a 486.  I got a lot
| of mileage out of a K6/II 300Mz with 128M, but it sure was slow
| opening something like Open/Star Office, not to speak of compiling
| the kernel.  For my own tastes, 800Mz with 256M (what I'm running
| now) is the minimum entry point.  With this fairly typical
| hardware, use of bloated (but perhaps very funtional) packages like
| kde/gnome/oo is practical.  My own tastes run to xfce.  I'm more
| interested in what the system can do for me and for the clueless M$
| consumers who can be lured away from the monolith.

great. and no one seeks to deny you that which you have, want, and can 
afford. however, a current linux distribution typically requires more 
horsepower than a current windows version does. this is not a good 
thing, unless we are eager to have the microsoft monopoly spread 
worldwide. in that regard, i think that shawn gordon's kobol will do 
more to promote linux adoption in the enterprise throughout the world 
than anything unitedlinux will do. as for non-enterprise adoption, 
there are not many countries in which people can typically afford to 
upgrade their hardware with the kind of frequency that we do.

this is above and beyond the claims, now fading, that one of the 
advantages to linux is its ability to work and work well on low-end 
hardware. i'll grant you your 486, but if a fairly full-featured 
desktop system cannot be had on a p-133 w/64 megs, there's something 
wrong and that something accrues to the benefit of microsoft, in 
several ways. first, the user who tries to install a modern linux 
distro on such a machine is going to be angry -- even more so if he 
blew away his windows install, which was working, for it. second, it 
suggests that claims for linux are as phony as claims for the 
commercial stuff. third, it robs linux of the upcoming generation of 
programmers, those in countries where computers are just beginning to 
take hold and where the enthusiasm that allowed things like linux 
could easily flower anew, but won't if linux requires hardware that 
no one has. i favor a system where people in those countries can work 
their way into the information age, because the alternatives are not 
especially pleasant. and there seems to be something weird about the 
whole notion of linux becoming a new-hardware promotional tool.

| An excellent choice, especially if you run xfce.  The install is
| easy, and it seems to be pretty stable.  I wasn't able to get my
| printer working (invested 2+ days before signing off), but
| otherwise Slack was AOK.  It's still my backup distro.  My big
| problem with Slack or debian or anything that doesn't adhere to the
| Mandrake/Redhat/ELX RPM framework is maintenance, and Slack (IMO)
| just doesn't have the necessry depth in its upgrade offerings.

well, i don't mind compiling my own stuff, so if there's a source 
tarball i'm happy. an advantage to this is that i've experienced none 
of the kinds of desktop and application crashes that are often 
reported, usually by people who just slapped in the rpm. and, of 
course, source gives me far greater control -- allowing me, for 
instance, to enforce the fhs here.

| My much maligned (on
| this group, at least) gentoo system has been continuously upgraded
| for 2 years.

now, now now. that's unfair. your gentoo wasn't maligned, *you* 
were.<g>

| (I did choose to reinstall once just to verify the
| newer install procedure, but there was no other reason to do this. 
| My system (in place since about March 2000) is the full equivalent
| of a gentoo system installed yesterday.)  I'll stand corrected if
| someone knows an easier to use maintenance system.

there are several, that if not easier are as easy or nearly so. red 
hat's system does work (unless you also subscribe to red carpet, in 
which case your hard drive becomes a battleground), ximian's system 
does work (unless you subscribe to red hat's update service, in which 
case your hard drive becomes a battleground), suse's yast online 
update works well if you haven't screwed around with the system 
(which leaves me out, but i suspect that gentoo would behave in much 
the same way were someone to go in and put everything where it 
belongs instead of where gentoo puts it), and debian's system is 
legendary. but that's a little bit beside the point, because the 
ability to get software while the bytes are still hot is not 
something everyone, or even most everyone, wants to do. and the 
reasons for their reluctance are often sound.

it misses the point in another way, just as unitedlinux misses the 
point: fragmentation of linux, and by this i mean the refusal to 
stick to some standards, is a greater immediate peril to linux than 
microsoft is. so adherence to what we might call "standard linux" is 
something that i think ought to be sought above pretty much anything 
else. the unitedlinux crowd is taking a great step in that direction, 
and i think they would have seen vast support but for that last 
little sentence, "oh, and by the way, screw the user."

gentoo, for all its technical goodness, does some things that 
unfortunately toss it into the same pile as lindows -- the "novelty 
distribution" category.

now, you disagree with me. unfortunately, if i'm right, the proof of 
that won't come until it's too late.

| 3) Reasonably prompt delivery of stable new versions of important
| packages.  I don't need to be on the bleeding edge.

./configure
make
make install

| 4) Nominal but not slavish adherance to FHS standards.  Since I
| have no need to mix and match multiple distros, I could care less
| about the filesystem structure of the other distros.  I could also
| care less whether a certain RPM will work on my system, since I
| have no need for RPM.

again, see above. you may not believe that this matters, because you 
are happy with what you have. i believe that there will come a time 
when you will see the importance of strict adherence to linux 
standards, because otherwise linux will fragment into oblivion, 
leaving users, among them gentoo users, in the same boat with beos 
users. or old-line unix users.

| 5) Good documentation about how my system works and what I need to
| do to customize it for my own use.  This used to be a bugbear for
| gentoo, but they now provide excellent documentation.

this hasn't been an enormous problem for most distributions for years 
now.

| 6) A responsive support group and/or developer group.

this hasn't been an enormous problem for most distributions for years 
now.

| 7) Not mentioned before, but prompt response from the distro for
| security patches is also important.  That would be more essential
| for server users.

this has been a huge strength of all mainstream linux distros i know 
of for years and years now.

| 8) Did I mention free?

well, your case for universal adoption of debian is now complete.
-- 
dep

http://www.linuxandmain.com -- outside the box, barely within the 
envelope, and no animated paperclip anywhere.