OSNews

Yeah, Ubuntu doesn't seem to have anything in the middle. 6.10 LTS is frozen and 7.04 is cutting edge. 6.10 is good for servers and office workstations cause you know it's solid. But you can't install newer applications and they are not providing newer applications so I think on servers where you don't have to change a lot.

(Like I use 6.10 for my mail servers cause all I have on them is mail and unless my mail software has a major upgrade I almost never have to make changes to the base OS.)

It would be cool if they could figure out a middle road where they could have a stable almost frozen secure base like 6.10 that could run newer apps like 7.04!

That would be magic! LOL!

6.06 is LTS - 6.10 was a normal release.

Sorry. You are right about that. Damn point releases. LOL!

They should do away with the decimal point, and use something more distinctive.

6-6 or v6.JUN spring to mind

Edited 2007-07-30 22:35

I think he even hit on something important that's missing from even cutting edge Ubuntu, which is the ability to update to externally stabilized code bases (like Firefox and OpenOffice.org).

In Ubuntu, the version of Firefox, OO.org, Pidgin, etc. are all tied directly to the release version of those apps, and the release version of Ubuntu. When it comes time to upgrade to Firefox three, it will likely take a whole distro upgrade to get that, instead of just updating the major upstream vetted version of Firefox by itself.

It would be nice to get stable releases from external sources more easily in Ubuntu than having to wait for the next distro upgrade.

I suppose Mozilla could set up an apt-repo for that matter (others too).

Yes that's right. Some software just isn't meant to be blocked by package managers. You can get some things from http://www.getdeb.net but the quality varies. I think some programs should either be put into a special category for updating, or simply, users should get an easy option to switch sources of packages to get latest versions.

Eg: nvidia drivers or openoffice is a nice example.

But one thing that may be a good in intended side effect of not being able to install things willy nilly is that security stays at a higher level.

So you can't just rush out and go to the next released version of something. Yet they do provide security patches etc for the current released version (Of say Firefox)

Hummmm, interesting.

One could take a contrary view that rather than employ resources to develop Mepis and Ubuntu separate from the Debian "base", that instead, these same resources be employed to strengthen Debian and abolish the idea of fragmented inheritance.

Consolidate resources (time, devs, money)...
Focus effort (on the base, not the fragments)....
Reduce Market confusion (the # of distros is confusing)...

That Ubuntu is that popular says, to me, Debian is that solid, otherwise, Mepis wouldn't be going full circle. So why stop at the seeming logic of using Debian as the base. Debian could be more than the base.

I'll bet there are some Debian purists out there saying, thanks for your help now go away! But what is the goal? The Desktop or (to borrow a phrase)
"To have a lot of fun?"

Winning the Desktop would be a lot of fun. It won't happen watching major distros jockey for position like this.

That might not happen with Ubuntu but I think some of the other debian based distros should try to do this. Ubuntu is too damn big, its gotten to the point where if you see some kick ass revision of another distro come out you see like 500 different posts of people singing Ubuntu's praises. I'm an Ubuntu user and I'm probably never going to switch barring a deal with MS, but I don't have to shout my buntuphelia from every rooftop. Anyway I though it was obvious that Dapper doesn't get much more than security updates, I think that's been clear since release. Ubuntu isn't Debian, it's goal is to release a new version every six months, dapper was for those who don't need the latest and didn't want to upgrade every six months. I do think that Ubuntu should think about fixing this issue as doing a refresh every six months is getting annoying (not really I find it exciting since they usually add some pretty cool things, and I like to be surprised with the new artwork).

the Ubuntu 6 month cycle is a good trade off between the lingering "always updated, never finished" of Debian unstable and the box & bow then no more updates for years that Red Hat does. I think the 6 month cycle is starting to have the effect of pushing the maintainers to tidy up projects to a usable state in time for the Ubuntu release. Keeping the calender as much as possible helps train developers and users to work predictably, without trying to tell them what to do. Project like Debian Unstable are always held back by the slowest person to commit updates. Ubuntu is about drawing a line in the sand and dealing with it.

I dont think so at all although one easily respects the developers attempts...I see many instances of what would be considered unstable features ( and some staples of said OS ) getting into 'releases' to believe this 6 mo. cycle is all that smart longterm. I realize windows has the market share going for them and all, but they have thousands of developers working daily on their OS's and they take several years to make them. I wont accept that they are inferior coders thats obvious, so I"d love to hear views on why linux is so adept that 6 month cycles are deemed doable and wise.

Maybe if linux as a whole were less fragmented ( we have how many distros now ? ) a 6 mo cycle would make a tad more sense albeit I still doubt it.

cheers
nl

I'm pretty sure Ubuntu devs send upstream patches, so they *do* help strengthen Debian...

"""
One could take a contrary view that rather than employ resources to develop Mepis and Ubuntu separate from the Debian "base", that instead, these same resources be employed to strengthen Debian and abolish the idea of fragmented inheritance.

Consolidate resources (time, devs, money)...
"""

I've often thought the same about hemoglobin in red blood cells. Rather than fold into that complex, confusing shape that it does. Why doesn't it just stretch itself lengthwise and form the longest protein that it can? That would be better, wouldn't it?

Dude, normally I find your posts are well articulated and argued but, come on... I mean seriously, as far as ironic analogies go, don't you think this is a bit over the top?

"""
I mean seriously, as far as ironic analogies go, don't you think this is a bit over the top?
"""

I'm taking a brain holiday. ;-)

OK, I get it, I think. One size doesn't fit all purposes/needs. It's a familiar argument but at least your metaphor is unique!

"""
but at least your metaphor is unique!
"""

Well, there are actually a couple of things that I was thinking about.

1. The complexity of the shape turns out to be more useful than something simpler.

2. The shape is not, literally, taken on for any intended purpose, but simply because the protein is being what it is.

Odd analogy, I'll agree. But it is what came to mind.

The OSS community cannot be anything but what it is. A certain degree of cooperation is inherent and beneficial. But aligning everyone in one way is just against its nature.

Saying, "Hey, we should *all* team up and..." is nice, but that's just never going to happen.

The freedom to fork means that when the disagreements grow large enough, development can split off into another distro, compiler, X Server, etc.

I used to think that it would be good if we could focus our resources on a single project in each category. Over time, I came to realize that was impossible and I was sad. Later, I realized that it was not even desirable, and that embracing diversity was our best strategy.

Right or wrong, it's all we can do, really. Because the hemoglobin is not going to unfold itself, and we probably wouldn't want it to in any case.

Edited 2007-07-31 04:09

No, it wouldn't. A protien molecule's functionality is in part determined by its folding geometry.

Great analogy, in both cases one would say "it doesn't work like that"

BTW, this gave me an idea: no more manual packaging. Let's have users's computers do the job in their idle time. All they have to do is install the packaging@home program

The goal of the debian to be "the universal operating system". debian is a mouldable "substance" that can transform to anything you want.

Currently, I'm using it as my desktop system. our company uses it as servers and somebody embed it to somewhere while others distros use debian as a base.

If debian changes to desktop oriented, we lose so much things. Also debian serves as a "high profile" distro for old school people (or hardcore or you name it). I for one don't like ubuntu or whatever distro that uses major gui tools for everything because I write my config files, I read man pages and I like to struggle with my PC and the general result is a system which works in the way I like most.

If I need to polish my distro I can do it from top to toe. If I don't, I use icewm or no Xwindow system. If I want something special I can use vanilla kernel with any patchset. You cannot do it in suse for example.

So, debian is a huge underlying platform which makes development of desktop distros possible. why convert it to something lesser than itself?

I bet that MEPIS will be back to Ubuntu next summer, shortly after the next LTS release. Warren doesn't want to maintain his own distribution, he just wants to do his value-add. From this perspective, why stick with Dapper when Etch is significantly newer? Next summer, why stick with Etch when Gutsy+1 is newer?

MEPIS will keep bouncing back and forth between whichever project has the newest long-term repository.

Personally, I think that Warren's arguments are disingenuous. There should be no confusion about package updates in long-term repositories. That's the fundamental trade-off. You either get incremental package updates or long-term support, not both. As the leader of his project, he should educate his users.

By using the EXPERIMENTAL code, each and every time, the Ubuntu code tree is inherently less stable than the Debian code tree, which contains additional levels of testing and vetting and fixing of code.

I guess he conjured this "Debian EXPERIMENTAL" thing from nowhere. I always thought that Unstable was a term that erred on the side of caution, but the use of all caps really drives it home that by using Ubuntu, you are effectively playing with napalm. Debian also uses "EXPERIMENTAL" code each and every time. They just have a much longer development cycle.

So while Etch is looking rather spry right now, it could be two or more years before Lenny is stabilized. Warren's users will no doubt demand newer applications before then, so he will once again come face to face with the new packages vs. long-term stability trade-off. Since he isn't inclined to say no to his users, he will have to find another argument for ditching Debian for Ubuntu.

This was not obvious over a year ago, but it is very obvious now.

Yeah, yeah. Just rephrase that one for the next switch.

This is a wonderful thing from the point of view of improving the Debian EXPERIMENTAL code.

This is Warren's perspective on the merits of Ubuntu's development cycle. Sure, it's great for Debian, but what about my users? Well, as great as he thinks his embellishments might be, there would be no MEPIS without Debian. No, he isn't required to consider Debian's well-being, but he's hurting his users in the long-run. He's the guy sucking on the end of the still while everybody else is feeding it grain and firewood.

It's ironic that, seeking to provide stability with the minimum amount of work, his solution is to drag his users back and forth between mother distributions. Is that stable? Is this the long-term outlook for MEPIS?

Edited 2007-07-30 22:56

Debian Unstable (Sid) and Debian Experimental (Scut) are two completely different branches of the Debian GNULinux distribution.

Yes, they eventually use it. That's why the code was written... However, to go from experimental to stable - or even to testing - is a long road. A much longer road than from experimental to ubuntu.

Very true! It's a shame that many distributions act more like parasites; it appears to be a symbiotic relationship between Debian/Ubuntu and Mepis but if Warren is gaining so much just by switching bases and the world is left wondering "why does Warren switch? Which is best?", it can't be good.

Personally, I'm a bit fed up trying to explain the "differences" between Debian and Ubuntu, and people like Warren aren't making it easy. I mean, picking Etch just because it's fresh now is not a long term solution! I agree with those that have said that Ubuntu will be the base again in a short while when Etch is "old and obsolete",

Agreed.

I don't use Ubuntu, but it seems to me any distribution release that advertises 3 or 5 years of long term support is not going to have _any_ updates at all other than security fixes. Same goes with RHEL and other distros that are specifically geared for long term support. Isn't that the whole point of a stable LONG TERM release?

Besides, Warren half pointing the blame at his users is not fair to the users. It's not up to the MEPIS users to make these decisions and determinations. It's up to the developers to determine how they are going to build their distro, and if they are going to use another distro as a base, then they better know that base in and out.

The bottom line is this: either someone didn't do their homework, or someone changed their mind and is now trying to put a positive spin on things.

Edited 2007-07-30 22:54

Well, in fairness if we're going to use that anology, the enterprise distros like RH and Novell do release service pack-type upgrades that contain more than bug fixes. But AFAIK they are optional, though supported.

I agree with your sentiment though, it was a mistake basing a popular community distro on an intentionally stable non-evolving release. Perhaps Warren though that there would be "supported" backports or something to that effect for LTS, but that was wrong.

As to whether there should be supported backports to LTS might be a different point of debate, but chacun son gout. You get what you pay for, and I don't recall Ubuntu at any point claiming that there would be an effort to keep LTS stable *as well as* modern.

This may be OT, maybe I missed something, but someone was just modded up to 7?!

Sorry if I'm the only idiot on OSN that missed a news item...Or is just making a big deal out of a bug...

It's a feature. Check conversations under OSNews v4 or wait for the announcement...

"Check conversations under OSNews v4 or wait for the announcement..."

I would have expected such a statement on the primary site but what do i know.

I'm afraid OSN will regret this feature soon. If readers are able to mod up comments, they'll be probably also able to mod down in such high numbers.

For a couple of months we have a similar system on a national site (www.webwereld.nl) which uses the same system. And it's a horror!

Almost every discussion ends up in an mass flame war, because of the damn mod system.

All this time I used OSN as an example for a functional mod system. Unfortunately I lost that argument by now :-(

Hope you still have the option to turn back the clock, when it gets out of hand ;-)

And, oh this www4 styff, looks really darn cool Wow!

There are the Backports to cover this.

http://ubuntuforums.org/showthread.php?t=40291

If you want a newer version of some app, chances are, someone has packaged it. You have to find it. Every stable release of Debian, or any other distro, has this policy.

Considering the non-LTS versions are supported for 2 years, I think there is plenty of time to let a release settle out before making the very smooth upgrade between releases. When feisty came out, I upgraded about 8 systems within the first week of the release. I had zero problems. Remember, apt allows us to upgrade without having to put in a full install disc.

If you want a newer version of some app, chances are, someone has packaged it.

You haven't really checked the contents of the dapper backports repo, have you? I just took a peek and there doesn't seem to be many packages in there.

Debian etch has been out just a couple of months and it has already more stuff than dapper in its backports archive.

http://archive.ubuntu.com/ubuntu/dists/dapper-backports/main/binary...
http://www.backports.org/debian/dists/etch-backports/main/binary-i3...

non-LTS releases (that come out every 6 months) are supported for 18 months, not two years. LTS releases come out every 2 years and are supported for 3 years on the desktop (so you have a year to upgrade to the next LTS release). Basically, whenever there is a new release in a series (LTS or non-LTS), the previous version will be supported for 1 more year after that (on the desktop, longer for the server).

My mistake. The last time I looked at it, I was looking at the server support. I could have sworn that I read somewhere that it was two years for releases. Sure enough, though, the current release for desktop and server on the Ubuntu download page says support through 2008. That would certainly fit the 18 months for Feisty. Thanks!

Ubuntu breaks predictably and reliably every 6 months, that's just one of its main selling points, rather than upgrading apps on top of a stable system they rip your machine to shreds and replace it with something completely different.

Not that this is necessarily a bad thing, it seems to be a very popular approach and works fine for a lot of users, it just feels uncomfortable for the rest of us who might have expected something more elegant.

I will not criticize you if you explain yourself...

Do you know what you're talking about? What would you like to not "update" or how you put it "rip to shreds and replace"? The kernel? Then ACPI will be crap, and many drivers. The "base" eg: gcc and gnu stuff? Then security and other patches will have to be kept in line.

I don't get it, unless you can explain what you really expect I think you all have some fuzzy understanding of how things work in a distribution, and what the kernel and "base system" encompasses, and what it means to live on old versions.

If you mean the "user experience" then yes, I can agree that Ubuntu did some changes on the top, but they were relatively small, and I think are going to be smaller each release as it stabilizes and best choices become apparent.

But please clarify so we can continue I still can't figure out what you guys really expect.

I wouldn't even go that far, because deep down I think alot of people already know this. It's a simple fact as the current ubuntu releases are based off of debian unstable releases. In terms of core stability, I would say they are pretty close...at least as far as Debian Stable vs Ubuntu LTS. Both are great platforms for development.

But I wouldn't put Debian Stable and ubuntu's 6month release of the day in the same category. Debian Stable vs Ubuntu LTS is really the only fair comparison for the 2 distros, in which case I call it a draw.

"""
Call me a troll but as I said before when the hype ends
"""

You're a troll. ;-)

Debian Unstable is, indeed, a better "distro" to use as a base for other distros. It is an ongoing development process which is exactly what they need.

Yes. Debian Stable is solid as a rock. Unpolished. But solid. And less suitable for Mepis' purposes than Dapper.

It never occurred to me, when Woodford announced the switch to Ubuntu as a base, that he was planning on basing his ongoing releases on a single Ubuntu release.

Call me a troll but as I said before when the hype ends people discover that debian is more solid and bug less than ubuntu... (the windows like os!)

Who ever said that Ubuntu was more solid with less bugs than Debian? I never heard that claim anywhere.

I believe he was saying that Ubuntu was 'as buggy as Windows', and inferior to Debian.

"""
Who ever said that Ubuntu was more solid with less bugs than Debian?
"""

http://tinyurl.com/3al68m

I agree. I was a long-time Debian user who switched to Ubuntu for about a year (6.06, 6.10, and 7.04.) What I liked about Ubuntu is that it required less work to get it to my liking following installation to the hard drive. Ubuntu failed to be perfect, however; there was always something that was lacking. I grew impatient waiting for "the next big release" and I didn't want to use the "herd" or whatever Ubuntu calls its interim betas. Each new Ubuntu release whets the appetite but fails to satiate the hunger. I realize that Ubuntu simply limits choice in different ways than Windows but still limits nonetheless. With Debian, I find that isn't the case. I've recently switched back to Debian in the form of Sidux. This is a rolling distribution that's completely up-to-date as often as I care to run aptitude dist-upgrade. I highly recommend it to Debian users who've tried Ubuntu but find that it leaves them wanting more.

"I've recently switched back to Debian in the form of Sidux. "

Welcome Home

I know how you feel. I've tried many distros of Linux and even a BSD, but the stability and reliability of my Debian Etch box is too attractive.

debian IS more stable than it, but that windows like comment was uncalled for

Ditto...simple as that, just plain scary

I really liked Mepis at one point...I found it stable and easy to use.

But...I think I'm beginning to feel that MEPIS is simply living off other distros. Maybe adding a bit and tweaking a bit...but if it's just customized *buntu/debian...why bother?

A derivative you mean? That is what most distros are, and often more popular than their parent distros:
1)PCLinuxOS, a Mandriva derivative
2)Ubuntu, a Debian derivative
3)Sabayon, a Gentoo derivative

All of them are more popular than their parent distros.

"All of them are more popular than their parent distros."

Given that there is no reliable measure of popularity, I really don't know how you can make that statement.

And no, Distrowatch does not count as a reliable measure of popularity.

There isn't indeed any scientific method, but it is hard to deny, for instance, that Ubuntu is more popular than Debian.

No, not a derivative. A derivative makes major changes afaik - there's something special, unique.

Sabayon adds an easier installer and gets you set up more quickly

PCLinuxOS includes all the proprietary stuff, and uses the Mandriva tools. It's known for being up to date.

Ubuntu, is well, Ubuntu. Debian is great, but Ubuntu made it SO friendly...

Mepis though? What did they do with ubuntu? Anything special? I can't tell. Haven't used it though

PCLOS also diverges from Mandriva in some major packages (most significantly the kernel and KDE, but also some others). Most proprietary stuff for Mandriva (nvidia and ATI drivers, wireless firmware, Java (but not Flash), winmodem stuff etc) is available in the public non-free repository since 2007 Spring.

Or is that a fork?

http://www4.osnews.com/permalink?259529

Edited 2007-07-31 01:23 UTC

Interesting Adam.

So part of the code base is shared between the upcoming OSNews 4 and the present one?