Re: OpenQM vs. Everything Else


Tony,
Why did you even bother to reply?
BobJ


"Tony Gravagno" <g6q3x9lu53001@xxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:4gco02lo6vr1dg3ugbqvkbs2ggsetj019e@xxxxxxxxxx
"Tom deL" wrote:
Tony writes:
I'll add that at the moment changes in the Linux world are very
similar when one is required to completely replace the OS in order to
use any one of thousands of programs.
Which of 'thousands of programs' have forced you to completely replace
the OS? Come on, really.

The net is full of references to packages that work on kernel 2.2.a
but not 2.2.b, or 2.4 but not 2.6, etc. Kernel v2.4 is up to 2.4.32
and 2.6 is up to 2.6.15. Have a program that now uses JFS? You need
v2.4. Something else makes use of sysfs? You need v2.6. So what
happens when you update a kernel? You need to update system startup
scripts and config routines. You also need to update some core
packages - let's just take a couple basic packages which required
updates from 2.4 to 2.6: binutils, gcc, make, nfs-utils, ppp, and
procps. Anything that uses these program/packages may need to be
brought forward, and any packages that someone installs may have a
dependency on these programs and thus the updated kernel.

Can we update the kernel without updating everything else? Sure, but
how many VARs or end-users are going to do that? Maybe Glen will do a
PickSource survey on this but I'd bet that better than 90% of the
people who "support" Linux really only install stock distros. At some
point these VARs or their end-users are going to feel compelled to
update some piece of software, perhaps something as critical as their
MV DBMS, and rather than tweaking a kernel or hunting for current
dependencies they will probably opt to just reinstall the entire OS.

Before I get any deeper I'm admittedly nowhere near your level with
Linux or that of people like Martin, Doug, or other "engineer" types
who work with this stuff every day. I think I've proven my case from
the perspective of an "average" user. I hope that's enough because I
don't have the desire or ability to carry it much farther.


Several people have suggested here that you cling to a poor Linux
implementation so that you have something to knock.

Several? Now who's exaggerating?
Before you go off on that shoot the messenger crap again, take a look
at this market and note that there are very few VARs and end-user
sites who are not using RedHat as their Linux system. I've never been
asked to work with an MV system over any distro except RedHat. Is
acceptance of the dominant platform "clinging" as though everyone else
has let go?

I'm apparently not the only one clinging to this platform - many
companies are desperate for something to knock:
jBASE = RedHat only
http://www.jbase.com/support/faq.html#version
U2, RedHat and SuSE:
<http://www-306.ibm.com/software/data/u2/editions/index.html>
Revelation = RedHat among many others (Go Mike!)
<http://www.revelation.com/Revelation.nsf/byTitle/Product+Information?OpenDocument>
Reality = RedHat only(?)
<http://www.northgate-is.com/reality/v12.php>
D3 = RedHat only
<http://www.rainingdata.com/products/status***/pick.html>
Univision = RedHat only
<http://www.fastfreenet.com/support/univision/1NewInUniV/VER8XX/CurrentSupportedPlatforms.htm>
And the only Linux distro mentioned on the QM download page is RedHat:
http://openqm.org/downloads.htm

Let's say RedHat is a poor distro. Exactly who's fault is it that the
providers of one of the most popular Linux distros can't seem to get
it right after over 10 years? Are all of those people who work there
incompetent?

And BTW, I use CentOS, not RedHat.
http://www.linux-magazine.com/issue/65
Looks like Linux Magazine needs something to knock too:
April 2006 Cover story: "CentOS - This community-based enterprise
Linux delivers stability and longevity without the high price"

Yup, real poor implementation.


In reality even in
the worst Linux distro I have seen this is not true. Can you offer any
real information that supports statements like this?

This is a regular game we play - I make a statement, you ask me to
back it up, I do, you ask again as though I didn't the first or second
time. I believe I've mentioned ghostscript and imagemagick in the
past, I've forgotten others. I've provided yet more examples above,
if you don't get it this time, tough.


What's worse, changing your Linux OS often requires a hardware upgrade
as well,

I regularly restore backups from state of the art servers onto ancient
hardware (PII 350 beaters) to test my clients' MondoRescue DVD's. The
other way 'round works as well.

There are a number of VARs here who might verify that they can't
upgrade their DBMS because that requires an OS upgrade and they can't
upgrade the OS because it requires a hardware upgrade. You may be
able to restore data onto old hardware but chances are slim that
drivers for the old hardware are available for current distros. I
know you haven't tested this - what reason would you have to load old
drivers into a new system? And would you really attempt to load
drivers into an old system used just to test the backup media? I
think not. No fault there implied, I understand what you're trying to
do. I'm just saying your ability to restore bits to an old system
doesn't imply that someone can move forward to new hardware. Again,
it's easy to find great wailing among Linux users that some driver is
no longer available in their new distro for even for the new kernel.


not so with Windows.

More often than not when I have attempted the above (restoring backups
to even slightly different hardware) WinXP pukes and refuses to even
boot to the GUI making it useless until the OS is re-installed ... 98
wasn't as bad in this.

My "not so with Windows" wasn't appropriate on my part. Sorry. I've
been saying right along that I'm not trying to prove Windows is any
better than Linux, I don't think it is better or worse. My point
continues to be that Linux is simply equally as fragile, and that
while people talk about Windows/Microsoft as being so evil I still
don't see Linux as having matured into anything worthy of replacing
it.

What I wanted to say about the hardware was that we can load Windows
XP on very old hardware (link to some site recently posted here on
this very topic) because functional generic drivers for these devices
are available in the install media. "Progressive" Linux developers
have a tendency to remove support for old hardware from the core, and
most developers don't want to update old drivers to work in newer
kernels.


I never mind a difference of opinion. What I do mind is someone who
disregards the truth - regardless whether it's to sell their agenda or
trumpeting their ignorance.

Yours in continuing amazement,
-Tom

Read the above. I've supported my claims (again). Feel free to
correct my errors or state other opinions - but please get off the
constant psychoanalysis crap.

T



.

Quantcast