[sf-lug] updates Re: SF-LUG Sunday March 3, 2019 meeting notice

[Rick Moen]

Quoting Bobbie Sellers (bliss-sf4ever at dslextreme.com):

>     I mention a lot of stuff from Distrowatch because it is handy
> for me but I do not mention a
> lot of stuff and I do not mention some things that I get for purely
> investigational reasons.

Sure, makes sense.  I certainly wasn't trying to get you to justify
anything.  ;->  I just happened to notice that the distros you recently
announced being added to your collection happened to be an exact match
for the latest release-news items on DistroWatch.

I'm fond of DistroWatch, too.  IMO, it's not as vital a resource to the
Linux communty as LWN.net, but still very worth following.

At the same time, it's important to note that DistroWatch covers quite a
lot of distros that are rooted in very ill-advised ideas, some that are
obviously doomed vanity projects, some that are incredibly specialised 
and that most people should stay far away from, and many other weird
variations (part of its charm, one might say).  

I've recently been re-pondering the above matter a bit because of time
I've spent re-building CABAL's distro collection, as I've mentioend
before.  (By the way, I'm hoping to drop by Sunday's SF-LUG meeting with
the collection, which is housed on an external USB hard drive formwatted
as FAT32 so that any of its contents can be copied off onto _anyone's_
computer or written to a user's flash drive, etc.).  So, you among
others are welcome to grab anything you want from the collection.)

Pondering about the CABAL library refocussed my attention on
longstanding points of pain such as ISOs omitting non-free firmware
BLOBs needed to make wireless or ethernet drivers work.  Thus, on
reflection, I ended up discarding all of the ISOs I'd downloaded of
Official Debian and replacing them with the corresponding unofficial
ISOs that are exactly the same thing except with the firmware files
merged in.

I mention that example for a reason:  In my opinion, LUGs have very
often served users poorly by not bothering to provide guidance to
newcomers in these matters.  I've seen the way it works:  Users 
default to a state of nervous avoidance of the unknown, so they assume
official must be way better than unofficial, and so walk away from a LUG
meeting with an Official Debian installer that is going to be
gratuitously difficult to install on their hardware -- solely because
nobody at the LUG took a moment to say 'No, really, you should use the
unofficial ISOs version unless you're very certain you won't have
hardware driver problems, and here's why....'

Likewise, we at the LUGs often have failed to help users who have
absolutely no idea about vital aspects of their hardware (such as, first
and foremost, total RAM) -- or, worse, think they know but 'know'
something completely wrong.  When people arrive with a Pentium IV
with 1GB of RAM, they really ought to be steered away from GNOME, KDE,
and even XFCE -- but also at the same time told they can do better than
microdistributions like Tiny Core Linux or antiX.

Basically, I have typical American-mentality problem:  I desperately
want to believe in progress.  Therefore, when I see LUGs continue to
fail to help people, and instead following the old hardcore nerd adage
of 'giving them what they asked for, good and hard', it makes me sad.


> I will not be downloading Linux from Scratch as it is now totally
> about putting a systemd
> variation on disk.

Happily, I'm pretty sure you are misreading that.  My understanding is
that version 8.4 of both LFS and BLFS are available in either SysVInit
or systemd flavours.

But, again, it's my hope that LUGs will step up to the plate in ways
they should have done for decades and have mostly failed -- in this case
carefully advising anyone expressing interest in LFS/BLFS that it's a 
very worthwhile learning project that would keep you busy for a month or
two, where you follow instructions in a long book to build up a Linux
system the hard way, by compiling each piece, BUT the end-result makes
absolutely no sense as a production Linux system, e.g., there is no
sensible way to subsequently maintain it.  Therefore, by all means run
through the book if you want to learn about Linux distributions piece by
piece through the computerist equivalent of carpentry, but please do not 
expect to use this as an ongoing computer system.

>     I don't know about 10.04 as I do not believe I have ever
> written about that.
>     My typo or yours?

Mine.  I meant 18.04 LTS (which was what you'd discussed people avoiding
and instead using the prior and increasingly antique LTS).

> 18.04.x on the other hand has plenty of mention in the connected
> equivalent of the LUG, the online Ubuntu  Users forum and those who
> could overcome a problem posted a few things to the Usenet newsgroup,
> alt.os.linux.ubuntu which I follow.  A fresh problem with a provider
> who is moving to 18.04 is mentioned today by a user.  It is because of
> updates to the libraries it uses.

In my experience, the problem with Ubuntu-specific forums, including
https://ubuntuforums.org/ and alt.os.linux.ubuntu, is that a high
percentage of what gets said there is pretty much wrong, and they 
produce a great deal of really bad advice.  

Just so I don't seem like I'm just totally dumping on the Ubuntu user
community, this problem isn't unique to its forums, just worse than in
most places.  For example, on the Devuan Project's main mailing list
Dng, recently there's been a long thread by someone claiming that the
newest version of the OpenDKIM package is now non-functional because it
now depends on systemd.  A fair amount of time got lost until wiser
heads finally spoke up and pointed out that there was absolutely no
evidence of a new dependency on system, and that the user had framed his
problem description inside of a wild and completely incorrect guess
about root cause.  The actual cause of his problem hasn't been
completely isolated, yet, although people have furnished good tips about
how to do so -- but at least more time isn't being wasted pursuing the
user's completely wrong guess that he simply assumped to be fact.

I cannot comment specifically on the 'plenty of mention' without looking
at specifics, but long experience has taught me that it's a huge, _huge_
mistake to accept at face value computer-user problem descriptions
without investigation and confirmation, _especially_ desktop users, and
_especially especially_ novice desktop users.  And, to be blunt, a
typical Ubuntu-specific forum has about it, despite best intentions from
all concerned, quite an air of the blind leading the blind.

And that is why I said (with typo) that it'd be a good idea if people
having problems with Ubuntu 18.04 LTS contacted, y'know, a Linux user
group.  

That person claiming to have a problem 'because of updates to the
libraries'?  Maybe.  Could be.  But remember the guy who was absolutely
certain that OpenDKIM broke because of a (imaginary) dependency on
systemd.

'Plenty of mention' of something on Ubuntu-specific forums, plus $2.50,
will get you a ride on Muni.  (Hurry before they raise the fares. ;-> )