Hi,
> So the only real advantage of Debian is the number of packages? Mandrake
Noooo. *sigh* My only point about packages was that it wouldn't be
necessary, infact it wouldn't make sense, to port the mandrake version of
most packages (say Gaim for example), when there was already a perfectly fine
Debian package. You are reading waaaay too much into the way I say things, I
never said that the "only real advantage" was packages.
> But you seem to say that all the work that has gone into packaging the
> majority of the rpms should be dumped? I would much rather spend the
> same time packaging the additional software ....
Well, as I said, if Mandrake did move to Debian, that would be the sensable
way of doing it.
> This should be possible with rpm/urpmi (haven't tested it recently), the
> biggest issue was allowing two different major-versions of the same lib,
> which has been fixed by mandrake's lib-naming policy.
Well, I'll believe it when I see it. All I've ever heard about with any
RPM-based distro is that at least 50% of those people that try to upgrade
(i.e. from 8.1 to 8.2 for example) nuke their system.
> Ok, it wasn't Mozilla, but it was bugzilla, craft,
> crossfire-{client,server}, epic4, fvwm95, gmc, gnome-admin, intuitively,
> kdepim, moon-lander, tkdesk, wine, and xosview
They are getting ripped out of woody as woody assumes the roll of stable -
not out of the system. For now, you could get them through "unstable" (sid),
or you could wait for sid to become "testing." It's just like Cooker -
perhaps some unstable packages are removed from the release (maybe), but that
doesn't mean they are gone forever.
Also, since Debian stable/testing/unstable constantly changes, Woody will
probably get the programs back soon.
> urpmi is fast enough with synthesis. Why change over to apt to gain a
> few milliseconds? I haven't got a copy of Debian installed, but if you
apt is faster at downloading package lists, is what I meant.
> want to prove apt is faster, try install a package from 8.2beta via a
> local ftp server, and try the same package with apt. I doubt the whole
> process will differ by any more than 5%, since it is dominated by the
> downloading of the package.
Obviously, most of the speed difference is dependent on the mirrors
available. I doubt there is much difference in the efficiency of fetching
packages - installation maybe - but not fetching.
> Log in on a text console, and run them as you normally would. mcc should
> be a good start.
Okay, thank-you, I will check that out. I had "I-miss-YaST-itus," hopefully
this well help. :-)
> If Corel Linux was successful, they wouldn't have needed the cash from MS.
Well I give other reasons why I don't think they were successful. The
distro was woefully outdated (because Corel chose to make it that way), KDE
was modified so much that any generic-KDE user didn't want to have anything
to do with Corel's version, and they were slow making new releases. Other
ideas - they made many in the community angry by telling projects like KDE
what to do, rather than contributing back code, they spent way too much money
trying to make their office suite and graphics tools work under WINE - which
turned out to be a bad idea, and so forth.
Also, think of all of the other distros that have come and gone, it isn't
always easy to pass up the estabilished distros - Mandrake is an amazing
success story that most upstart distros can only dream of.
> Well, the first bit does seem to be as a result of wanting to use Debian
> and have stable packages.
Right... they should have never based their distro on Debian stable. As it
was noted in the top LWN editorial this week, most people run Debian testing
or unstable. Debian Stable is to Linux what OpenBSD is to BSD.
> I didn't ask how many urpmi ditros there are, but how many rpm-based
> distros have been around for 3 years, so you should answer with how many
> .deb-based distros there are (.deb!=apt). dpkg/.deb~=rpm, urpmi~=apt.
There aren't any that are commercially successful, that's the point. My
point was, basing future success solely on previous success isn't a good way
to decide if a product has a future. If everyone did this, DVD, CD's, Linux,
Windows, and even AOL, would no longer be around. The best way to analyze the
technology is to see if future functionality + future opportunities (i.e. a
major distro like Mandrake adopts it) = success.
Look at it this way, your way of figuring would say that since there were
no successful versions of UNIX for the desktop three years ago, Apple should
*not* have based OS X on UNIX. Sometimes a visionary concept can only be
caught by looking ahead rather than behind.
> Tell the LSB that ... I don't see an rpm-based distro changing to dpkg
> while LSB specifies rpm. Apt is of course another issue.
Obviously the LSB would not select the packaging system that is the less
popular of the two major ones. However, Debian *is* mostly LSB compliant,
after all, RPM is included - and works quite will inconjunction with alien.
> I don't think there is any value in discussing this any further, as all
> the advantages of going debian-based have been shown to be
> non-conclusive, except for the greater number of packages available.
No! I didn't say greater number of packages was the only advantage. You
said that. I only noted packages when you were worrying about all the time
that would need to be spent repackaging.
I agree though, that it's is worth ending this conversation. From what I
have seen, those that know Debian agree with me, those that don't just bash
it for unreasonable reasons. I'm certainly not saying Debian is perfect, but
I think an excellent distro like Mandrake could benefit from being based on
Debian.
-Tim
--
----------------------------------------------------------------
Timothy R. Butler [EMAIL PROTECTED]
Universal Networks http://www.uninet.info
Christian Portal and Search Tool: http://www.faithtree.com
Open Source Migration Guide: http://www.ofb.biz
============= "Christian Web Services Since 1996" ==============
