[問] Is it the best interest for Red Flag to base their system on Debian instead?
Currently Red Flag bases its system on Red Hat, which is not as free, far less versatile, much smaller software repository, more difficult to maintain and upgrade than Debian.
I just wonder whether it is the best interest for Red Hat to switch to Debian instead.
It can just use the Debian repository (without wasting its resources to make a new distribution) and just develop a custom installer, which has tasks for the installation of Chinese environment (printing, input, display etc,) and possibly a custom kernel (with extra security feature such as NE and SELinux) and a SELinux-compliant toolchain. Red Flag can just base its system on the most recent stable version (Sarge at present). This would be sufficient since the release cycle of Debian proper would be considerbly faster than before.
In addition, Red Flag can limit its supported architecture to i386 and amd64 minimize expenditure of resources.
What do you think?
Debian and the Debian logo are trademarks of Software in the Public Interest, Inc
No Way!
Red Hat (...) is not as free, far less versatile, much smaller software repository, more difficult to maintain and upgrade than Debian
Red Hat is not that hard to maintain if you are using it properly (not to use RPMs built for other systems, don't needlessly ticker with the system too much, etc.).
I just wonder whether it is the best interest for Red Hat to switch to Debian instead.
WTF? Why would Red Hat want to do that?
(...) the release cycle of Debian proper would be considerbly faster than before.
I don't believe that. Base on past experience:
hamm -> slink 9 months
slink -> potato 17 months
potato -> woody 23 months
woody -> sarge 35 months
Etch won't come out for at least another 2 years, possibly 4. In today's environment, you need at least 1 release every 12 months, and Debian just couldn't handle that.
I am sorry, I meant the one t
I am sorry, I meant the one to do the switching is RED FLAG, not red hat
If Debian continues accepting
If Debian continues accepting help from Ubuntu, and its really drastically reduce its release archiectures to 4-5, then probably it can release every 2 years.
According [url=http://lwn.net
According to a Release Manage
According to a Release Manager [1], sarge was targetted to release on Dec 1, 2003.
[1] http://lists.debian.org/debian-devel-announce/2003/08/msg00010.html
:) never trust the Debian
never trust the Debian release team. God forgive them, cause they don't know what they are doing.
Having a stable, testing and unstable stage of a project is a bad practice. But the Debian has no choice but using this scheme because the project can never keeps its schedule. And by using the scheme, the project hinted that user can use the testing stage on their 'stable' system.
It is a sad, sad situation..., but there are nothing we can do. I mean the Debian already has the largest number of developers amount all the Linux distributions. They are just not capable of managing a proper release from time to time.
All the Debian users are adviced to use the unstable source by other users sooner or later. How sad is it?
I do not agree the 3 meta-rel
I do not agree the 3 meta-release paradigm is wrong. I would rather think that it is an ingenius invention. Imagine if Debian has strict release schedule that it takes 4-6 months for a release, then "stable people" are less inclined to use unstable for the sake of new versions.
Instead, whether this release schedule is good is another problem. For a desktop system like Ubuntu, it is perfectly fine. For a server system, it can be a nightmare. Therefore, let us do not make Debian something it isn't intended to be. If you want a system with very new packages, use Ubuntu and Gentoo "stable". If you want a system with very very new packages, either go to "testing" of their distribution, or use backports or user-supplied portage.
I don't remember seeing that
I don't remember seeing that Debian was defined as either a Desktop or a Server distro. A long release cycle does not guarantee a stable release. Debian's pattern of delayed release cycle is not caused by instability but incompetent on management.
I think people should be encouraged to use 'stable' release instead of unstable release. Unstable release was intented toward developers and brave hearts. When average users are forced to use something 'unstable', they'll not trust a system as stable and they will not understand that a stable system is a important thing (as in Win3.1...WinMe dark ages).
End users should use stable releases, being they are desktop users or server users. Test users can use unstable or pre-release system. I believe Debian developers take pride on their claim on stable. Of course, now a day, end users only talk about 'out of date' about Debian. No one care about how stable Debian is because everybody is using unstable or apache1.3 and php3 a month ago.
"testing" should be "tested"?
To me, the 10-day delay to enter testing makes the package "tested". Is it really so unstable to use testing? I doubt. I don't know why nowadays people stress so much on new versions; but if this is the case, then probably we should push new versions to them, with less quality for sure.
"tested"? NO!
Testing shouldn't be run on any computer unless you are a Debian developer. It is intended for DDs only.
Testing breaks less often than Sid. But when it breaks, it takes a long time for things to get rectified. Sometimes this could be days and it could be months at times. But in unstable things get rectified within couple of days. (The reason being packages could take months to filter down from unstable, known to have happen before.) Moreover, testing is last to receive any security patches (stable first, then sid, and those patches need to make their way to testing...)
meaning of stable
Stable is taken to mean ``doesn't change'', and that's why Debian stable is stale.
Imagine when you make a relea
Imagine when you make a release in Debian, what you are looking for? Release-critical bugs. RC bugs are: (critical) affects unrelated software, (grave) unusable by all users, (serious) violates policy, and nothing else.
When (at release time) there are no RC bugs for a package, the "testing" version is accepted and will become the coming stable.
But this process can be as short as 10 days. So what process convert a package from "testing" to "tested"?
Your notion of "tested" is not well-defined
See, for example, this message in debian-user and subsequent discussions. This has been discussed to death in the debian-user list.
There is no such thing as ``tested'', especially in today's CADT development model of most F/OSS.