Slashdot Mirror
Linux Vendors to Standardize on Single Distribution
Jon James writes "eWeek is reporting that a number of Linux vendors will announce on Thursday that they have agreed to standardize on a single Linux distribution to try and take on Red Hat's dominance in the industry. " The vendors in question are SuSe, Caldera, Conectiva, and Turbolinux. However, as the article also points out - Red Hat has a very well established lead in the corporate market - and Sun's decision to create Yet Another Linux Distribution (Sun Linux! Now With McNealy Vision!) will make the waters even more muddy.
I've used SuSE for some time, and been happy.
However, many is the time that I wanted a newer version of software than was available from SuSE. An "uber" distribution, compatible with the assorted branded distros catches my interest because it may increase the likelihood of finding new software in rpm form that may actually work on my system.
Worth watching.
If they can pool the strengths of each distribution into the new one, that will make it stronger.
I think some major consolidation is way overdue for Linux. Of course, new distributions will always appears to fill in the empty spaces.
It should read .."cash in on redhat's dominance"
These companies came in on the wave of redhat
So they wont use rpm then ?
RedHat's success with businesses is not that their distribution is better than others; - although it's a fine distribution tailored for businesses - it's that they give manager's what they want - support contracts, courses and certificates for employees etc.
Businesses don't like to take risks, they want to see a shiny reliable company selling them a reliable product, instead of "some freeware distribution written by no good hippies in their spare time". RedHat gives them the comfort of that illusion.
When one OS or distribibution is enough, then you might opt for Windows. There is a need for standardisation, standardisation with Unix, for security, for standard Linux. Competition keeps everybody on their toes. With the combination of several distros a lot of duplicate effort will not be done any more. Less duplication of effort is good for all.
There are few details that I don't see being resolved yet..
All those companies mentioned don't give free ISO's just like RedHat (and Debian for that matter, as well as Mandrake) which kind of makes sense - a customer who downloads and uses the downloaded ISO's is one less customer who would like to pay for the distribution (not all of them - but most of them)..
I can understand RedHat point - they don't give a shit about people using Linux on the desktop - their eyes are focused only on the enterprise - thats why you won't see RedHat Advanced server available for free download, and you'll need to pay $800 for it (with the bare 30 days support - installation support) so how they're going to compete with RedHat??
This reminds me the LPI exams (which everyone but RedHat stands behind it) VS. RHCE training/exam - how many people here passed the LPI? how many passed the RHCE? somehow I got the feeling that RHCE is WAY more preffered then LPI..
Hetz (Heunique)
Linux Magazine has an article on why Debian would fill in a good role as 'arbitrator' amongst the distributions and why HP chose to use Debian as their standard distro.
A distro free from vendor squabbling and influence, that's exactly what the Linux 'standard' should be. Now all we need to do is get some LSB action going.
Why are they bothering to come up with a single uber-distro when Debian provides a solid foundation for this kind of work? If I were a Linux distributor, and was starting to realize that I can make money selling services and a name, why would I waste all this money making up yet another installer - hell, I'd hire 10 guys, slap a commercial release on top of Debian every 6 months, and let the community do the heavy lifting - all the while earning open source karma for supporting Debian.
I read about this last night, and had mixed feelings. It's certainly overdue in the market - there definitely needs to be simpler 'cross-distribution' compatibility for installing packages. Yes compiling from source is generally compatible, but not everyone wants to do that, nor should they have to. It's a waste of someone's time to do that in many cases.
I think it may be too little, too late, however. This should have been done over a year ago, and there still seems to be too little information on the specifics of the deal(s). Figure it'll take *months* before this has any impact on the installed base out there, it'll be a miracle if this actually 'saves' any of these distros from further marginalization.
Someone else mentioned Redhat feeds into an 'illusion' that businesses want - 'shiny support', etc. It's no illusion. It may cost money, but damn it - if someone in a business needs support for something (driver doesn't work, upgrade broke, whatever) having a *real person* to call who's been trained on that particular distro is invaluable. Yes, it may cost $200. Yes, you 1337 geeks out there could hang around in IRC for a few hours waiting to get an answer. *Businesses* can't afford to do that. Furthermore, they shouldn't have to put up with those channels of support (not reliable enough - quality of support is hit and miss, and they can't afford to wait for the 'hit' all the time). Whether or not they ever need it ('linux is so stable!') the fact that it's there is more than comfort enough to persuade people to go the Redhat route.
Furthermore, the Redhat certification and training and all the other secondary services simply help to bolster their lead in the mindshare of the business market. Maybe it's just that they had more cash to play with after their IPO - if so, they've put it to good use.
creation science book
So, will the SEC have to step in an approve this?
Seriously, though. Would any of us be happy if Volvo, Volkswagon, Ford, Hyundai, and Chrysler decided to "standardize" their automobiles to compete with one big vendor? I for one would say no. It would make some innovative new idea, like say a zero emmissions fuel cell car, that much more unlike the standard. New ideas will seem more outrageous if there's such a baseline from which to deviate.
Something needs to be done, because the Linux community is allowing itself to get slipped into the Microsoft mindset. With the LSB in place, there should be none of this business of "targetting a distribution" or other Microsoft-like lock-in nonsense.
1: The LSB needs to be in place.
2: All major distributions need to adhere to it, and the minor ones should too, for that matter.
3: Education is key, that LSB-compliance is the real crux of the matter, not some specific distribution.
4: Packaged software should state its requirements relative to the LSB. LSB+foolib+barlib, etc. Some distributions may choose to distinguish themselves by including foolib and/or barlib out of the box. The ISV should also have copies/pointers for foolib and barlib on their web site.
5: Distributions are good. More are better, as long as LSB can solve the interoperability and installation problems.
I'm disappointed to see LSB mentioned only once as of my writing this post.
The living have better things to do than to continue hating the dead.
Well, most die-hard linux hackers do not say they use RedHat (Notable exception: Linus Torvalds). Most have used RedHat, though. Why don't they admit it? Because there is not very much to hack on RedHat. Red Hat's strength is that the stuff they throw in actually works. More or less out of the box.
However, I don't use RedHat today. I used RedHat up to 6.2. Then I started looking at other distributions. At that time, I did not know much about how GNU/Linux _really_ works. I never made my own startup scripts. I did'nt compile programs. Heck, I didn't know where my libraries were or how I inserted a kernel module from the command line!
I started testing different distros. Mandrake. TurboLinux. SuSE. Many others. I tried at least seven or eight distros before I met Slackware. At first, the entire Slackware system seemed awkward. But after a while, I experienced a lot and learnt even more.
Now, I'm most productive on Slackware. Because I know the system so well that doing stuff from scratch is _easier_ and _faster_ than using tools like rpm and linuxconf. (overall, of course, some things are still faster with linuxconf).
RedHat is a distro for those that want a GNU/Linux that works - not for those that want to get a GNU/Linux to work. It is a good distro, but not what I want from a GNU/Linux system.
As a last addition: It is not a funny OS either. Mandrake is. Cute little penguins and round, purple install buttons. Colors and fun. RedHat is grey and red. Only a few, boring games. A corporate-type webpage. RedHat has lost the childishness of linux jokes and internal humour. It has grown up.
Grown-ups are easy to communicate with - but children are much more fun and can be adjusted and tweaked more easily.
Roses are #FF0000, violets are #0000FF, all my base are belong to you
Have you heard of these small companies called IBM or Sun Microsystems ?
It can't be ease of use, that is not the point of UNIX-like operating systems. Some distros may get close to the ease of use of Windows, but is that really the primary goal of any distribution?
It can't be the prettiness of the desktop. Window managers are not tied to the distributions, although some prefer prefer certain desktop suites. However you look at it, there is a UNIX that takes the desktop beauty pageant hands down: Mac OS X.
Maybe you judge "best" by how much control you have over the operating system... does that require working with source in all cases, or is fine-grained package management good enough? They all give you incredibly control over the operating system, differences in the layout of /etc aside.
I could keep on going, but I'm sure you get the point. How much "better" a distribution is has to be looked at very subjectively and therefore the judgement lacks meaning. As long as a distro works, installs, is reliable, and essentially does what it claims to do, you have to give it the stamp of approval as a good distribution. Past that, everything is a matter of opinion.
Perhaps, for business, Red Hat simple is the best. Personally, that is the conclusion I've come to. I love Mandrake, prefer it, more or less, to Red Hat, however I've chosen Red Hat for the servers I build (and I build servers both for personal projects and for use by the large telecommunications company I work for) and for workstations. My workstation at the office is Red Hat, while at home I use Mandrake.
Each has their benefits. I've played with Debian, Storm Linux, Progeny (i.e., Debian+), Gentoo, and so on. Every distribution has something of value, some only as learning tools.
If we want to get very Darwin about the whole thing, then Red Hat is obviously the fittest distro. Its not the first, but it is the largest, most widely used, and has all but wiped out older "species" like Slackware, IMO.
perl -e 'print $i=pack(c5, (41*2), sqrt(7056), (unpack(c,H)-2), oct(115), 10)'
Linux isn't ABOUT anyting, accept running your computer in concert with its peripherals.
Linus has said several times that he isn't some idealist promoting some grand utopia. It was a stupid learning project. His use of BitKeeper should have slapped some sense into you.
Red Hat has ONLY been about money from day one. They have been capitalising on the free marketing that is the rabid evangalism that so many beatnick youth lend to it. SuSE is no different. Caldera bought DR-Dos just to make money on a law suit, so what does that tell you about them.
Linux companies don't exist to magically whip up VC to pay the volunteers, and then go bankrupt. They exist to make a profit. And if you are going to make a profit, you need to constantly work on increasing that profit, or you slowly die.
Not to mention that this VC doesn't come from some grand benefactor but from a lot of savvy business people that have no clue what Linux is, just what a profit is. As soon as you accept their money, you have already sold your ideals down the toilet.
Linux isn't about anything.
Linux companies are about maximising profits while minimizing production (using other peoples IP) costs.
Why would RedHat be considered evil? Aren't they doing what we want them to (i.e. release a free (as in beer and speech) operating system based on the tools we love)?
How can that be considered evil? Because they take a loss on every ISO download? SuSE would probably have more market share if they gave away their YaST2 enabled distribution, but it's not in their business plan. In the copycat IT industry it's refreshing to see someone take a different approach.
Personally I think there are better distributions than RedHat (Mandrake and SuSE come to mind), but RedHat has more exposure, marketability, and history. They were in the game first, and they're on the tip of every newbie tongue. And apparently they're doing something right on the business side, or they wouldn't be viable.
If other distributions are to survive they are going to have to provide something new and different, create strategic partnerships, and establish credibility. In other words, compete. There's little intellectual property disputes, and everyone has the same resources to pull from. This is the free market, and it's far from evil.
The other aspect of Red Hat (and, perhaps, all significant distributions) is the work that goes into developing a stable combination of packages. kernel-2.4.18-4.src.rpm is a long way from the generic 2.4.18 kernel: it has over 100 patches, including a 20 MB whopper from Alan Cox. GCC 2.96 is the most visible fork, but hardly the only one.
It's all free software, the majority of which makes it back to the original project, but Red Hat is the first to take advantage of its own hard work. That's an advantage.
I agree. There seems to be some creepy groupthink going on around here.
/sanction those who've succeeded in distributing, promoting, and profiting from Linux.
capitalism = evil
profitablity = evil
success = evil
North American = evil
therefore...
Red Hat = evil
The fact of the matter is that Red Hat said, early on in this game, that Free software can be the basis for a legitimate business model. With the exception of some technical issues (LSB, gcc, blah, blah, blah), they've put their money where their mouths are. And, they haven't wavered from the "Free" part, either. In addition, they haven't shut the door on free ISOs like SuSE has ("live evalution" CD, anyone?), they don't panhandle like Mandrake (incidentally, I use Mandrake on a couple of boxes but I find their "business model" quite disconcerting -- I'd never recommend them for use in a business environment), and Red Hat isn't proprietary like Caldera, Turbo, or even (please correct me if I'm wrong) SuSE. All they've asked (although it was handled a bit heavy handed, if you ask me) is that businesses redistributing Red Hat disks make it clear that downloaded ISOs are not Red Hat supported. That doesn't seem too much to ask.
Finally, ask yourself this question: If it were SuSE or Mandrake that was the dominant distro, what would be your reaction? If you'd feel the same as you do about Red Hat's dominance, then we're fighting a losing battle. It's unreasonable, counterproductive, and just plain idiotic to punish / censure
Shoot! I happen to like the RedHat distro., mostly because that's what I'm familiar with, and recommend it to everyone else who asks me, but I sure wouldn't want to see them muscle out all the other distributions. Remember -- it's the competition that keeps these companies honest, and forces them to keep the customer at the top of the priorities stack!
Your Servant, B. Baggins
And Redhat has taken a fairly consistent free software stance for their distribution throughout most of their history. While not everyone may actually care about that, the people most likely to push Linux adoption within an enterprise are often people who do care. For those the choice falls between Redhat and Debian, with debian more adherent to the free software interests, but with redhat not too bad, and more palatable for buisness (not to mention support for all that fun enterprise software needed).
Standardizing may solve the problem with software, but it wont solve how they're percieved. Being semi-proprietary may help them get their customers to pay, but it isnt going to help their marketshare, or help them gain advocates. Redhat remains a better compromise between freedom and buisness.
The obvious solution is standardization in combination with version management that's how regular distributions pull it off.
... 3.0.1 came out more than a month after 3.0, so while I want the latest bug fixes and enhancements, I'm running an overnight compile of KDE at most once every 5 or 6 weeks. This is hardly a great burden, and the trouble saved by having packages compiled against the proper librarys, and the resultant stability, more than makes up for whatever time is spent starting the compile before going to bed (or leaving work for the day).
... I have Gentoo installed on an old Intel MMX 233, which did take 3 days (!!) to install, but as I've said before, I probably spent a total of an hour sitting in front of that box, and the rest of the time ignoring it while it churned away), in which case binary distros, with all their many faults, may be exactly what you want. Even with a slow processor, though, I suspect you would find the incremental time savings of having a more responsive and quick system well worth the initial investment of time ... my slower boxes benefit even more in some cases from the kinds of optimizations Gentoo and Source Mage provide than some of the quicker boxes.
Yes, of course, but there is a heavy price to be paid for that "standardization", namely the inability to have current software on your system, and the subsequent slowdown in providing timely feedback and bug reports to authors.
Debian, as an example (and as my favorite binary only distribution), had one of its developers respond to a question by a curious user as to when XFree 4.2 would be included in debian with the curt answer: "Leave me alone. It will be months." Source Mage ("Sorcery" at the time) had X 4.2 available within a day, Gentoo very shortly thereafter. Those of us who needed the bug fixes and additional hardware support didn't have to wait "months" for its inclusion into a binary distro, or alternatively have to compile it ourselves (by hand) and then watch as various distribution-provided binaries start to break because of X 4.2's differences from X 4.1.
On the contrary, we had clean, solid, good support from day one, which meant we got the bug fixes in a much more timely manner, and were able to deploy configurations not even possible with other distriutions. And we didn't have to sacrifice stability in order to do it.
As for Mozilla, it may have nightly cvs builds, but Gentoo and Source Mage both reference release builds (e.g Mozilla RC2, RC3, etc.), not nightly builds as a rule. So while those wishing to have the very latest may find themselves compiling mozilla once every two or three weeks, it certainly isn't a nightly affair. Ditto for KDE
It also makes keeping up with security fixes much, much easier than with Red Hat, Mandrake, Suse, and Debian (stable excepted, but debian stable makes the Jurassic appear contemporary).
You may be personally allergic to compiling large packages (or have a processor that is painfully slow to do so
In any event, for those of us working with this stuff every day, who have reasonably modern CPUs and who are required to troubleshoot live systems whenever anything goes wrong, having quick and painless access to the current bug fixes and features is an invaluable asset. In short, for many people (I would venture to say most), once they've tried a source-based distro, binary distros feel klunky by comparison and a person will never want to go back. The benefits are simply too great, naysayers notwithstanding.
The Future of Human Evolution: Autonomy
It's nice to see Linux evolving into a healthy competitive ecosystem. An ecosystem of true competition vying for customers by offering better products at better prices.
Contrast with Microsoft's vision of an ecosystem where they are the big predator and everyone kow-tows to them and their whims. Nothing really happens in this ecosystem without Microsoft doing it first.
So far from being a disturbing development in Linux's history, I consider this a good sign that, contrary to Bill's opinion, the Marketplace works!
The bitter lessons of a veteran coder: http://bitterprogrammer.blogspot.com
Let me speculate a little here...
These 3 companies have large market share in different regions of the world:
SuSE == Europe (Sans France => Mandrake)
TurboLinux == Asia (Well, more SE)
Conectiva == South America
I don't know where Caldera fits into the equation... maybe their IP is worth something or they have lots of support/consultancy staff with Linux expertise...
IBM have enough presence in North America, and also Australia (which probably services New Zealand as well...)
So... lets say IBM is behind this "merging" of distros... which in reality is a conglomorate of Linux services/consultancy companies which spans the majority of the world (Sans Africa/Middle East -- there's probably enough IBM staff there to cover the demand, although maybe SuSE services this region as well...)
Genius? I think so.
I would not be surprised if these companies were receiving funding from IBM... or if they were bought out by IBM in the future...
Time to invest in IBM, I think.
There seems to be a lot of meta-creepy groupthink going on around here - seems some people think that "Slashdot" is some unified, collective entity and not an aggregation of individuals, many of whom do *not* have the view that "____ is evil" where ____ is any entity or concept.
Well, it isn't.
Also, many individuals who might call Microsoft "evil" do not base this evaluation on Microsoft's profitability, or its large market share, or the fact that it is based in North America, or because they believe that capitalism is evil. Rather, some people believe that Microsoft's predatory behavior within the capitalist system adversely impacts things that said individuals might value - e.g., the benefits of improved products at lower prices that have traditionally result from competitive markets.
I personally believe that where a market is not competitive, and as a result, the products are not getting cheaper (consider the static absolute cost of Windows) or are increasing in price (consider the cost of Windows relative to the cost of a computer today, vs. even 2 years ago), then either the structure or operation of the market is flawed (e.g. lax or non-existent antitrust enforcement), not that the monopolist is "evil."
Of course, this belief is predicated on the concept of Adam Smith's "invisible hand" - i.e. markets should be structured and regulated such that participating entities (producers and consumers) operating in a totally avaricious way will advance a given end. The question is, ultimately, "is the given end good or bad," not "is Microsoft good or evil." (The question of who gets to chose the given end is a good one - I believe some recent attempts to regulate the system of legalized bribery known as our current campaign finance system are germane to this issue.)
If our society believes that an appropriate social policy goal is that Bill Gates should get richer, then we have a system today that works well. If others believe, as I do, that the benefits of competition (e.g. lower costs and improved products) should accrue to consumers (not necessarily end users, though end users would ultimately benefit), then the market for desktop operating systems is not optimally structured or regulated.
This has nothing to do with whether Microsoft is good or evil, or the question as to whether Red Hat would behave as Microsoft does, were it in a similar position (IMO - Red Hat management would be rational to do so).