> Among commercial distributions, Red Hat outsold its nearest competitor
> more than 2 to 1 in February 2000, according to PC Data. In August
> 2000, IDC reported that Red Hat is outselling its next five
> competitors combined, and accounted for 48% of all sales. Granted,
> this isn't a direct measure of linux usage because even the commercial
> distributions can be downloaded for free and the CD's can be shared
> and copied in disproportionate amounts after sale, but there has never
> been any thesis put forth that one commercial distribution would
> propagate in this fashion in a way fundamentally different from any
> other commercial distribution. A logical conclusion is that, to a
> first order approximation, Red Hat is used at least twice as much as
> any other commercial distribution.
Sure, but that still doesn't offer proof that it's the most-used
distribution out there. It may be in the highest *SALES* category, but
that doesn't mean the most-used.
And like you said, it's impossible to tell, and projections like
those offered by Netcraft and IDC are simply that, projections.
> The anecdotal evidence also points to Red Hat being the most popular.
> Hardware vendors (IBM, Dell, etc.) are mostly lining up with Red Hat,
> not noncommercial distributions. And where I work, many people are
> using Red Hat, a few use Mandrake, and a few use Debian. No, I'm not
> claiming anything scientific here.
Except the fact that Gateway, Dell, IBM, and Compaq have all opted
for a custom linux distribution to be developed for their machines. I know
this because the company I work for holds the contracts for the
development of these builds.
> More important that all of the above, even if someone can come along
> and show that all of this is wrong because, say, Debian has more users
> than Red Hat, that doesn't imply that Red Hat is not also popular.
Ok, here you make a valid point. When people think Linux, they
think Redhat, and painful as that is, it's a reality. The general populus
doesn't know that there are over 83 currently shipping distributions in
the channels, with dozens more just out there for the taking.
> It is. Therefore, plucker would get more positive publicity if plucker
> worked on Red Hat. That was my main point, and I don't see anything
> particularly controversial about it.
Nothing controversial, but if Plucker works on 20 other
distributions, and breaks on Redhat, I might have to look more closely at
what Redhat is doing different, not what Plucker is doing different.
As you know, Redhat has a reputation of taking 100% compliant and
functional software, patching it, and causing all sort of havoc with their
userbase. They've done it with the kernel, with things like gcc, with a
lot of other things.
> This is the oldie but goodie defensive stand-by response of an open
> source programmer who is feeling criticized by some nobody such as
> myself complaining about their open source software. I understand the
> sentiment, but often, this kind of response can turn off people who do
> actually want to help.
This is me asking those who lodge a complaint to turn it around,
and help us improve the product, and be cooperative. It is not my attitude
about open source. I frankly don't have 10 hours a day to devote to
Plucker, and do what I can in my spare time to help. That spare time over
the past year has been very thin lately, and most of that is because the
company I work for is in a bit of lean waters. We've been through a lot of
garbage. I hope to get some of that time back soon.
> Primarily, what I have done to try to help you is to bring the plucker
> Red Hat compatibility issue to the table by reporting it six months
> ago to the plucker developer list. As I don't know python, I couldn't
> look into the problem too deeply myself, but I did look at the code
> anyway and humbly offered that perhaps Url.py contained the code that
> was not working correctly on with Red Hat 7.0. In the last six
> months, there really has been no followup to the suggestion that the
> plucker code could be modified to fix the problem.
Ok, this speaks to getting a better bug reporting mechanism linked
in on the website. Admittedly, a flaw in my time schedule. My plate is
overflowing, so I try to delegate as much as I can to get things done
lately.
> I don't know if it is, but we should all be open to the possibility.
> Mike clearly isn't, which was my complaint.
I'm open to helping, but I don't have a single Redhat system in my
fleet (of my own), but I'm sure I could slap it on a box at work and see
what floats to the top. We have some very sharp python guys there, as well
as incredible rpm packagers, etc.
I'll dig a bit more and see what I can find out.
Yes, consider this my responsibility. Flag me for details. Mike
needs to stay focused on the viewer, not the parsers. That's our realm.
/d