Was Mandrake Linux 9.1 released too early ?

[Guest ndeb]

I don't know. Is it just mandrake. RedHat 9 has 38 updates so far.

Take a look at https://rhn.redhat.com/errata/rh9-errata.html . There are 14 updates, 12 of which are security updates. Only one bug-fix relates to a redhat-made package, unlike in case of LM9.1 (which has about 10 updates for mandrake made tools).

 

Here's the comparison:

Total size of the updated rpms

ftp://updates.redhat.com/9/en/os/i386/ - 118.1 MB

ftp://ftp.univie.ac.at/systems/linux/Mand...pdates/9.1/RPMS - 125.1MB

 

Total number of updated rpms

RH9.0 - 39 rpms

LM9.1 - 93 items

 

Maybe the six month release cycle is just too short to do proper testing.

A one year cycle would mean less revenue.

[zero0w]

A one year cycle would mean less revenue.

 

Not necessarily, a one-year cycle could mean less support problem in release management. But again I think Red Hat's main revenue stream comes from corporate Advanced server edition, which is upgrade about once per year, or even less often.

 

The problem always is that open source package update & release is never meant to comply with business upgrade cycle. The year 2002-2003 has so many noticeable changes in desktop (Xft2, Fontconfig, Qt3.0+ & GTK2.0+ plus various apps upgrades) and kernel area (2.6 is coming summer, and glibc and gcc had numerous upgrades) and also the increasing user base that support problem becomes much a greater problem than before.

 

Once these changes become more stabilized, I think we will have less of a trouble; but I agree Mandrake's own tools are rushed, just that the rapid evolution is increasingly rapid and more update and patches would be expected, at least in the coming year or so (not to mention what the XFree86 fork project can contribute to the scene with more patches against the official tree by different Linux distributors out there...).

[Guest ndeb]

Not necessarily, a one-year cycle could mean less support problem in release management. But again I think Red Hat's main revenue stream comes from corporate Advanced server edition, which is upgrade about once per year, or even less often.

Whats true for the server market is not true for the desktop market. Desktop stuff has to be more bleeding edge because users want the latest in GUI, multimedia, graphics, office-suite etc. Also, note that a server release is based on a desktop release that has been tested by a large user base for quite some-time. Hence the desktop cycle must be much less than the server release cycle, to allow for adequate testing by the user base.

[zero0w]

Not necessarily, a one-year cycle could mean less support problem in release management. But again I think Red Hat's main revenue stream comes from corporate Advanced server edition, which is upgrade about once per year, or even less often.

Whats true for the server market is not true for the desktop market. Desktop stuff has to be more bleeding edge because users want the latest in GUI, multimedia, graphics, office-suite etc. Also, note that a server release is based on a desktop release that has been tested by a large user base for quite some-time. Hence the desktop cycle must be much less than the server release cycle, to allow for adequate testing by the user base.

 

The irony is the server revenue is subsidizing the desktop side of it. An average desktop Linux box cost much less than a server version of it. Not to mention the downloadable ISOs. However, it's only fair to say 'subsidizing' when the users don't work on ANY development or testing; in the case of Linux community, this is the opposite, the users are also the developers and also the beta testers and hence the distributors can benefit from them as kind of voluntary labor. In any case I see Linux is a different community altogether - not just different from Windows, from also very different from Mac user community altogether.

[ral]

Makes me wonder if the change in RH numbering (no moe .1's .2's .3's) is a signal that they may just be releasing a single version per year...

 

I really would not mind. After having gone from ML8.2 to ML9 to RH8 to RH9 (and trying to find a copy of ML9.1 so I have one more to go this year)... I owuld not mind sticking with one (RH9 or ML9.1) for at lest one year.

[Guest ndeb]

It appears that the fixes of

http://www.mandrakesecure.net/en/advisorie...e=MDKA-2003:004

were not proper fixes. So mandrake needed to "fix" these fixes and here are the latest updates in

http://www.mandrakesecure.net/en/advisorie...MDKA-2003:004-1

 

This may explain why some bugs did not go away inspite of updating packages that were supposed to fix the bugs.

[Crashdamage]

ral said:

 

>I would not mind sticking with one (RH9 or ML9.1) for at least one year.<

 

I agree. Like I said early in this thread, I'm still using 8.2 and will for quite a while. I run Fluxbox and some text apps like Mutt and Links anyway, so I don't much care about the latest KDE or Gnome eye-candy. And my setup is so complicated that reinstalls just to keep things on the cutting edge is just too much hassle. Besides, I need this machine for daily money-making activities too, so I'm more concerned with stability and reliablity than new goodies.

 

I am gonna put another machine together out of spare parts just for fooling around with. I had one for a while, played around with Debian, RedHat, etc., but I gave it to charity.

 

As for my main machine, I think I might upgrade it on the next version. But then again, I might stay with 8.2 'til the next hardware upgrade, probably in a another year or so. It depends on how bug-free new releases seem to be and if there's anything new I feel a real need for.