• Home
  • Blog
  • Mandriva needs Debian? Come on

Mandriva needs Debian? Come on

Vincent Danen

June 23, 2005

Yesterday started some of the most laughable speculation regarding Mandriva's future that I've read in a long time. It made for quite a good chuckle at first. To recap, an interview on LinuxPlanet with Ian Murdock was published yesterday indicating, among other silly rhetoric, that "Not only can they do it, my assertion is if they don't do it, they can't survive" which is Ian speaking about Mandriva buying a commercial Debian-based distribution to use Debian technologies. Of course, this comes now that Mandriva (formerly MandrakeSoft) bought Conectiva and then bought Lycoris.

This is really amusing because, for one, Ian Murdock is the "father" of Debian and it sounds very much like he's tooting his own horn here. He also runs Progeny, which is also a commercial Debian-based distribution. Does it sound like Ian is asking for a buyout? A little too much to these tired ears. Of course, he could be more blunt by saying that Mandriva needs to buy Progeny to survive, but maybe he's a tab bit more subtle than that. Regardless, it sounds to me like he was doing the interview with his hands out asking for Mandriva to please give him a whack of cash because without it, Mandriva won't survive.

But what are his reasons? The first is that Debian has a good community. Well, so does Mandriva. Part of the reason Mandriva survived bankruptcy is due to the community that ponied up for services like MandrakeClub. This goes to show that Mandriva does have a strong community. And if you look at the development arena (cooker) at Mandriva, you'll see a lot of volunteers there doing a fair amount of work. Sure, Mandriva's community (in terms of developers) isn't as big as Debian's, but that hasn't caused it to falter at all. If this were a drowning point, Mandriva wouldn't have the assets to even consider these last two aquisitions.

His other point is that such a buyout by Mandriva would be a huge win for Debian, who just recently released the latest stable version (Sarge) after a 31-month development cycle. Ok, sure, this would be nice for Debian, but how does that affect Mandriva's future? Well, it doesn't. That directly affects Debian's future (which, in the same article, Ian has some issues with). How does it benefit Mandriva? It doesn't. So, let's be clear here. Mandriva buying a Debian-based distro benefits Debian's future. But how does it benefit Mandriva again? Hmmm... nowhere does he really say why buying a Debian-based distro will benefit Mandriva (other than the larger community pool). Oh wait, he says this:

"Debian has an enormous opportunity in front of it," Murdock stated, if it can remain the center of Debian-based development. Having such a free, non-commercial center would be a strong selling point for any company to use Debian, he added, including Mandriva. Keeping Debian's core non-commercial would prevent any temptations to specialize the core code into something only one distributor could use.

Ummm... how does this benefit Mandriva again? Oh, it's a strong selling point. This is because there is some popularity to other Debian-based distros like Xandros, Linspire and most importantly, Ubuntu. Does Ian think that Mandriva will sell more subscriptions to MandrivaClub and retail boxes because it can stamp a "based on Debian" logo on the box? I doubt it. The same people who use Debian now will continue to use Debian. A few may switch, but probably not enough to make an aquisition worth it.

I don't get it. The whole thing is to make this radical change from Mandriva's existing software base to change not only the package management tools (dpkg and apt vs rpm and urpmi), but also to roll-back to Debian-based packages that are largely out-dated by the time they're deemed stable? Yes, Mandriva is moving towards more stability, but come on... a 31-month development cycle? Debian in and of itself is still Debian, still ruled by the Debian community, and not overly influenced by outside commercial offerings. If it were, then Ubuntu would be the new Debian because at least Ubuntu is moving and moving quickly. And while I think Mandriva is moving from "bleeding edge" to stable, it's not going to want to use 2 year old apps because someone at Debian is asleep at the wheel. Which then means that Mandriva development will continue as it has, it would fork Debian's base and maintain it's own repositories which it would use, and then where's the benefit? A sticker that says "loosely based on Debian"? Yeah, that'll sell a lot of boxes.

That would be the end of my morning chuckle if I didn't get another one today. Now we get this editorial piece from Brian Masinick on Linux Today that opens with:

"I think that Ian Murdock is right on with regard to speculation about the future of Mandriva."

What?!? Has this guy been smoking the same crack Ian was? No, wait.. Ian's not smoking crack... he's begging for a buyout, probably because Ubuntu has so severely eclipsed Progeny as to make Progeny obsolete to all but it's existing client base. Brian here is smoking a whole different kind of crack... the blind-eyed zealous kind.

His arguments don't even make sense. Half his editorial is trying to explain why he has a clue. So what? So he started out with Slack in '95 and didn't start to use Linux on the desktop until 2001. So he used emacs and tar back in the '80's. That doesn't at all indicate that he has a clue.

His arguments are the same. Community. Debian's community. The theoretically bought-out company's community. Nowhere is there a single piece of technical why it should be done. Everything is based on community. So both Brian and Ian are saying that Mandriva won't survive if it can't buy Debian community. Huh? They expect Mandriva to completely overhaul how it does everything simply to buy a piece of Debian-using community? That just doesn't make sense.

He indicates that now that Mandriva and Conectiva have merged, Mandriva has "one of the assets in house that can help accomplish a move from RPM packaging to DEB packaging: the great synaptic tool". Eh? I'm pretty sure synaptic is open source and freely available and it seems pretty darn expensive to buy Conectiva just for a tool any idiot can download and use. That purchase was for entry into Latin America and the client-base that Conectiva had (read: companies and government, not so much end-users). See, that's strategy. Buying users just to have more users from disparate parts of the globe doesn't make sense. It would make more sense to buy TurboLinux and get a good foothold into Asia than it would to buy XYZ Debian-based distro just to get a bunch of users from who-knows-where.

Anyways, regarding tools, Brian also says:

Mandrake did a pretty good job of turning urpmi into a pretty decent installation tool that nearly rivals apt-get, dpkg, aptitude, and synaptic, in terms of handling application and library packaging updates.

Well, if that's the case, why on earth would they want to change? Give me a good technical reason for moving from rpm to deb and urpmi to apt-get! He just said that urpmi nearly rivals apt-get and these other tools. Granted, he calls it "decent" and that it "nearly rivals" but he never quite says what's wrong with it or why it doesn't measure up. Maybe because he doesn't think it's important, or maybe he just doesn't know. In my experience, urpmi is fast, smart, and pretty darn good. It doesn't "nearly" rival, it does rival. Granted, my experience with apt-get is limited to Fink on OS X, but heck, that's enough to give you a taste for the tool, and frankly, I'd be happier using urpmi on OS X than apt-get.

Finally, and this is the really amusing part:

I like the idea of centering that user community around a Debian approach because the Debian packaging fosters community, but even more importantly, the Debian Social Contract is all about community and freedom.

And why is this a compelling reason to buy a Debian-based distro? Because a package fosters community? The only community a package (manager) fosters is the holier-than-thou-apt-is-god mentality most Debian users have. I think it's more the idea that no one owns Debian that fosters it's community by and large, not the package management tool. Sure, some people prefer it and that's why they prefer using Debian. From a user's perspective, apt is nice. But urpmi is just as good and just as easy. So let's look at it from a technical perspective: dpkg packages suck. They are really badly put together... if there's something that needs a change, dpkg is it. Sure, there are some benefits to it like a package can recommend another package and so on, but the bottom line is working with a dpkg is a pain and RPM has nicely laid out the spec file to do all the work, separate patch files to make maintenance of patches so much easier than dpkg's one-file-fits-all mentality. When I have to pull security patches out of Ubuntu or Debian packages, I cringe because they're so painful to use once you're used to simple spec file and distinct patches. Pulling a patch out of a Red Hat or SUSE package is 10x easier and faster.

One final point before I wrap this up. Ian mentioned LSB support and whatnot for Debian as another selling point of Debian. Guess what Ian... LSB chose RPM as the defacto packaging system, not DEB. So if LSB support is such a big thing, why hasn't Debian switched over to using RPM, like the LSB recommends? For instance, in the original interview Ian says:

One such benefit could be the re-establishment of Debian as the central Linux Standard Base distribution, something that was originally intended when Bruce Perens initially proposed the LSB. The LSB, Murdock feels, has been weakened by other commercial distros' lack of full participation.

Ummm... Debian is the one lacking full participation. Has Ian not read the pertinent part of the LSB specs? I'll quote it here:

Applications shall either be packaged in the RPM packaging format as defined in this specification, or supply an installer which is LSB conforming (for example, calls LSB commands and utilities).

So, at the end of the day, Mandriva, a company committed to the LSB, is on the right track using RPM as the base, and not the non-LSB-conforming DEB packaging system. It can be argued that DEB falls under "LSB conforming", but it's still not the recommended system. RPM is.

Give me a break. I for one hope Mandriva stays as far away from Debian-based distros as possible, not only due to the inferiority of dpkg itself, but because it doesn't need this kind of clueless zealotry advocating it.

Leave a Comment

Comments use MarkDown. Need help? MarkDown Cheatsheet