Sigh. I love it when things get thrown out of context and misinformation abounds. I’ve gotten used to this with proprietary vendors and ignorant news sources, but to get this within the open source community? And from a web site that is supposed to be pro-OSS? Come on.
Yes, this is a “follow-up” to the Why I’ll never use Linux for my main desktop blog entry. Why is a follow-up required? I didn’t think it would be, but apparently a website that doesn’t seem too terribly fond of Mandriva has pounced on the comments that Fabrice and I have made regarding the state of Linux on the desktop and because this misinformation has been made, and people seem to actually pay more attention to Distrowatch than to the sources it references, I felt it to be a responsible thing to make a rebuttal to the antagonistic “journalism” of Distrowatch than to ignore it (which I would have preferred to have done).
The issue in question can be viewed on Distrowatch under the title Mandriva’s Anti-PR. Of course, the headline is there to grab attention, first and foremost, and in the most negative light possible. Mandriva anti-PR? Let’s be realistic. Since when has my personal blog become Mandriva PR? And when was that blog post anti-Mandriva? It absolutely wasn’t. Let’s see how the numbskulls at Distrowatch twisted things:
You wouldn’t expect something like this to happen at any time, let alone on the eve of Mandriva Linux 2007 release. Last week, two Mandriva developers went on record saying that “Linux is not ready for the desktop”. In a web log post entitled Why I’ll never use Linux for my main desktop, Vincent Danen concluded: “This kind of thing is precisely why, for me, Linux belongs on the server. Now if those kernel-developing idiots would get their act together and give us a 2.6 kernel that was a) stable and b) supported for more than a week, then I’d really be pleased.” Another Mandriva developer, Fabrice Facorat, agreed. In a post called Linux ready for Desktop? Of course not!, he had this to say: “For me when you have control over the environment and the hardware, Linux is ready: think servers, workstations, thin clients, kiosks. However when you have no control on the hardware, Linux is not ready because it somewhat lacks flexibility.”
There you have it, right from the horse’s mouth! From two guys who have been working for a Linux company, developing a popular desktop Linux distribution, for years! Yes, that’s right, Mandriva Linux is not ready for your desktop! Don’t waste your money buying it!
How sensationalistic is that? For one, their facts are wrong. Fabrice doesn’t work for Mandriva; he’s a volunteer (and a fantastic one at that; I value Fabrice’s opinion like few others). I am the only Mandriva employee here. Did I say “Mandriva Linux is not ready for your desktop”? No. Did I say not to buy it? No. How they interpretted my blog post to come up with this is beyond me. I can see some simple-minded idiot giggling with glee and rubbing his hands as he read, then completely missed the context of my entry. Yup, visualization is fantastic. That’s an image I’ll carry with me for a while.
One really has to wonder about the future of a company which happily asks Gaël Duval to leave, yet is quite pleased to keep developers who publicly hint that they consider Linus Torvalds, Alan Cox and other kernel developers “idiots”! A perfect pre-release PR, non?
Pre-release PR? I didn’t know I was the official voice of Mandriva. Actually, I don’t think I’ve ever been anything close to an official voice, much less a PR-spinning individual. So to call this a PR is… well… beyond stupid. To hint that Linus Torvalds, Alan Cox, and the other kernel developers are idiots? Yes, well, I’m entitled to my opinion, non? (i love that little french spin there). And yes, I don’t retract that opinion, which I can qualify in a moment (I should have qualified it before, sorry, but this is not a new opinion of mine).
Let me explain things a bit further. Distrowatch, while I appreciate the things they’re trying to do, really need to get their heads out of their collective anti-Mandriva asses and actually read what people write. Since they are patently incapable of doing so, I’ll reiterate my previous blog post here… or at least the relative parts because it seems to be required.
For instance, there’s all this talk about the 3D Desktop stuff in Mandriva. Well, I like eye candy so I want to check it out.
This was the reason why I was messing with this anyways. Why did I chose Mandriva? Well, come on, it’s the distro I’ve been using for the last 7 years, it’s the distro I use daily, it’s the distro I like the best (and I’m not ashamed to say I’ve tried many (in fact, at this moment I even have Ubuntu on my Toshiba Libretto which is more or less waiting for me to get around to sticking Mandriva 2007 on it instead)), and it’s the distro developed by the company I work for. It’s the obvious choice.
At any rate I’m less than impressed. Well, with this. Mandriva 2007 seems pretty nice so far (well, I installed RC2 and synced up with cooker right after the install). The 3d desktop is of less importance to me than getting vmware server working on it, but it would have been nice to play with since this is one of the premier features. Sadly, I need a lot more hand-holding than I’m getting.
Ok folks, I know I’m asking the impossible here, but could you actually read that paragraph? Oh, you did? I kindly invite you to read it again. No, really read it. Come on, it’s written in good english. It’s not too difficult to understand. READ IT! Oh, sorry, didn’t mean to get excited.
So what does it actually say? It says that I’m less than impressed with the experience of setting up the commercial ATI drivers (yes, I’m actually reading it in context, what a novelty!). It does not say I’m less than impressed with Mandriva. Nor does it say I’m less than impressed with any experience of Mandriva. (For what it’s worth, the day after I wrote that, I tried with SUSE on the same machine with the same results… no DRI).
Oh! Oh! How about this little nugget: Mandriva 2007 seems pretty nice so far. Umm… if I didn’t like it, would I say it? I think most people who know me, by now know that I speak my mind (which has gotten me into trouble on more than one occassion, granted). If I thought Mandriva 2007 was garbage, I probably would have said so. (Ok, well maybe not, but I wouldn’t have called it nice either). No, the fact that I called it “pretty nice” implies that I liked it. You don’t call your wife pretty nice if you plan to divorce her. Well, not if you’re honest at least.
I then go on to say that the 3d desktop is of less importance to me, which means I can live without it. It also means that I can use the distro without this feature. It didn’t say I was going to trash the install looking for something easier (which I did end up doing, but didn’t find anything easier, and went back to installing Mandriva 2007 back on that same machine and FWIW, I got the 3d desktop stuff working today anyways).
So I’m still a) not seeing any anti-Mandriva PR here nor b) seeing how it could be remotely construed as such.
I don’t think the problem with Linux on the desktop is Linux… Linux, and the associated bits and pieces make a fine desktop. It works really really well. The problem is how the desktop compares with other desktops.
Oh yes, here it is. Here is exactly where I say that Linux on the desktop is garbage. Oh… wait a second. What’s this “fine desktop” business?!? How the heck did that get in there? Oh, right, I wrote it. What? It works really really well. What the… heresy! Someone must be editing this behind my back!
If you just need a functional, no eye-candy, all business desktop… look no further than any Linux distro with an X server.
Groan. Those dang imps are at it again wrecking my perfectly sensible anti-Linux and anti-Mandriva blog entry. Damn them all! What is this nonsense? Here I’m actually endorsing Linux on the desktop? That can’t be right…
Ok, now I’m gonna qualify the only anti-anything part of my blog entry. Yes, the anti-kernel-developer statement:
Now if those kernel-developing idiots would get there act together and give us a 2.6 kernel that was a) stable and b) supported for more than a week, then I’d really be pleased.
Ok. Here’s the deal. Before the 2.6 kernel, it was easy to figure out what was what. 2.2 was stable, 2.3 was development. When it was done it became the 2.4 kernel. Then we went to 2.6. The old style of development made more sense… develop on the 2.5 branch and when it was deemed stable, start the 2.6 branch and work on 2.7 for development. But this hasn’t happened. Instead, we have a 2.6 kernel that has API changes from version to version, we have massive changes from version to version, we have things like XFS working in 2.6.16 and then causing massive filesystem corruption in 2.6.17. What the hell is this? We have a driver in one 2.6 version being changed and renamed in a later version (here I’m referring to the sk98lin driver although I don’t think it’s impossible that this hasn’t happened with other drivers).
The problem here is there is no possible way anyone can look at the 2.6 kernel as being anything other than a rapidly-moving development kernel. It has changes that are too intrusive between versions. It has inexecusable problems that crop up between versions (thinking of XFS here). It has an insane number of security vulnerabilities. This is supposed to be the kernel we want to use?
Unfortunately, it is the kernel we want to use. Why? Better support for more devices. Some really nice features. There are any number of compelling reasons to use the 2.6 kernel, and just as many not to. Unfortunately, the 2.4 kernel is in major maintenance mode… it’s not getting any new features. Unless you’re running a server (and even then), the 2.4 kernel is just plain old too old. I had resisted moving my servers to the 2.6 kernel for so long, but now I almost don’t have a choice. Better SATA support? Need the 2.6 kernel. AppArmor support? Need the 2.6 kernel. For the SELinux folks who want that support? The 2.6 kernel. Better performance? The 2.6 kernel. Newer gigabit drivers? The 2.6 kernel.
More instability? The 2.6 kernel. More insecurity? The 2.6 kernel. More problems moving from one “minor” version to the next? The 2.6 kernel. Moving from 2.4.16 to 2.4.17 was probably a no-brainer (been a while, I don’t really remember). Moving from 2.6.16 to 2.6.17? Watch out.. it’ll eat your XFS partitions and the data you choose to store on it.
So yes, I think the kernel developers are idiots. They’re so concerned with new this and new that that they’re forgetting the people that actually use it. They’re forgetting that not everyone wants to run the equivalent of Mandriva’s cooker or Red Hat’s rawhide. Not everyone wants to run a development kernel in production. What we do want is a stable kernel that we can use and know will be maintained with minimal and non-intrusive changes for some time to come. I want a kernel that I can upgrade without worrying what will break or what new userspace tools I need or any other concern besides compiling and rebooting.
But this isn’t going to happen anytime soon. The kernel developers have essentially left the onus on the Linux vendors to grab a development kernel and attempt to stabilize it themselves, which means backporting fixes because the kernel developers have moved on to the next cool thing. I understand the “geek” nature here, but they’re doing the Linux community a disservice by not simply saying “ok, the 2.6 kernel is done, now it’s time to move on to the 2.7 or 2.8 kernel”. Then the 2.4 kernel can be officially retired, and people can maintain (stabilize, provide security fixes for, etc.) a 2.6 kernel that people can feel confident in using.
But because the 2.6 kernel has been a development kernel for so long, and the 2.4 kernel is minimally maintained, we’re basically hooped. We can use something on the verge of being obsolete, or we can play with a development kernel. There is no middle ground. There is no recent stable kernel. Why is this so hard to figure out?
I need to qualify here that there was some talk of making the 2.6.16 kernel the “stable” kernel but I really wonder how well this is going to work out. I’d really like it if it did work out, and I really hope that it does, but I think that the decision of a “stable” kernel needs to be made by the entire kernel development team and/or leadership, and I sincerely believe that means that they need to stop playing with 2.6 and move on to another major version number (like 2.7) and then it won’t just be a handful of people wanting to do it, but something the entire kernel community stands behind.
Anyways, that’s enough of that. This post here is basically to a) maintain that yes, the kernel developers are, in my opinion, idiots and b) that the Distrowatch folks are idiots too. At least the kernel people can read… seems to me like that basic skill is lost on the Distrowatch folks.
Oh, and for those people leaving comments on the Distrowatch thread who are saying “that’s it for Mandriva, time to look at SUSE or Ubuntu”, if you’re basing that decision on the sensationalistic anti-journalism of Distrowatch and not on your own experiences, then shame on you. Are you sheep that you do what others tell you without any thought or questioning on your part? If so, then I pity you. If you read the Distrowatch “news” and then actually followed the link to read the blog postings by Fabrice and myself, you’d realize that Distrowatch is completely incompetent and incapable of anything even remotely resembling “journalism” or even worthy “news reporting”. And if you read Distrowatch, without reading the blog entries in question, and took it at face value, then I really feel sorry for you because you’ve allowed someone else to form an opinion for you and have given up your freedom of personal thought to a pack of irresponsible or even, dare I say it, malicious individuals (because their spin on this is either incompetent or malicious… take your pick)
And finally, this is my personal blog. This is not a vehicle of Mandriva PR. This is not even a Mandriva-hosted or Mandriva-sponsored or even remotely Mandriva-related blog. This is where I have my little corner of the web to speak my own mind. What I say here is not Mandriva’s opinion, policy, or anything else that could be construed as even remotely “official Mandriva”. This is the opinion of one simple guy who happens to work for Mandriva. And I feel bad that what I wrote was so misconstrued and interpretted in such a misleading way, but I cannot apologize for being a free-thinking individual with a personal opinion.
So shame on you, Distrowatch. I hope people read you as they read OSNews… with a handful of salt and a distinctly non-sheep mindset.