Re: unstable is "unstable"; stable is "outdated"
I thinking the problem here, as I mentioned before, is one of semantics as opposed to a real problem.
Options are:
1) "unstable"
pros: Very up to date,
cons: Occasion big bug that can do damage
user: Someone who knows there way around Linux pretty well and likes to say,"I use unstable 'cause I'm so cool!"
2) "testing"
pros: Pretty up to date, very stable
cons: May have the latest-greatest version of an app a week or two after your buddy using "unstable". Last to get security upgrades.
user: Someone who actually uses their computer to get work done and is less worried about being the latest greatest cool guy. Someone who laughs at their co-working trying to figure out why init keeps bombing after his last "unstable" upgrade.
3) "stable"
pros: Rock stable, quicker security updates than testing.
cons: old
user: critical server.
Most of these characterization are user standpoint. If I had a server that was super critical, I'd use stable (or *BSD). The two servers I have don't have a huge load and it's not a big deal if they go down (not to sound like a huge power user, 'cause I'm talking about a home network and a minor server at work). I have testing on them.
In short, if you're a user, it doesn't make sense to use stable. Use testing or unstable and you're system will be as "up to date", if not more, as any distro. If you're running a server, just use testing, unless security is a big issue. Then use stable. Or use testing and keep a watch on the linux security pages, and manually apply the newest patches when they come out.
IMHO, there is a level for any use inside the various debian trees. The biggest problem from a PR standpoint for debian is in the names.
Feel free to disagree with any point I made, 'cause I'm not as good as I sound.
On 2/2/02 at 6:39 AM Jason Lim wrote:
>On 2/1/02 at 4:25 PM Tim Quinlan wrote:
>
>>> kernel, etc... and as we all know, jumping from "stable" to "unstable"
>>is
>>> problem-prone and doesn't worth flawlessly every time.
>>
>>Why jump all the way to unstable, why not use testing? Testing is
>>usually stable enough for most applications plus the various software
>>packages are pretty up to date.
>>
>>
>>In my experience unstable is pretty damn stable as well. I upgraded a
>couple of boxen from stable to unstable a little over a >year ago and
>haven't been bit by any of the big bugs. I just check the mailing lists
>and debian planet to see if anything big >has popped up before doing an
>apt-get update && apt-get upgrade. Obviously these aren't servers.
>
>In my experience as well. As I said in a previous post, i've heard that
>testing is the last to get security updates, which is not acceptable if
>you're running servers.
>
>
>>I think the only problem with debian is the naming. Changing nothing but
>the name from "unstable" to "cutting edge" or
>> something and there wouldn't be close to the outcry about how 'behind'
>debian is. IMHO.
>
>Well, there more or less needs to be more frequent "stable" releases...
>something along the lines of Redhat's quick releases. Okay... Redhat
>again.. i know i know... but you've got to admit they've got the release
>aspect of their distro pretty good. They are business people over there,
>and they know how frequent business users like to have updates, and when
>critical updates should be released.
>
>I'm just wondering if it is even POSSIBLE to follow the frequent release
>schedule that Redhat follows, or if it is not possible because most/all of
>the developers for Debian are volunteers and won't "work" to such a tight
>schedule. If we can find and identify the problems not allowing up-to-date
>releases, perhaps a solution can be found?
--
Lang
Is uniformity attainable? Millions of innocent men, women, and
children, since the introduction of Christianity, have been burnt,
tortured, fined, imprisoned; yet we have not advanced one inch towards
uniformity. What has been the effect of coercion? To make one half
of the world fools, and the other half hypocrites.
-- Thomas Jefferson
Reply to: