Slashdot Mirror
An Open Letter From Bob Young
Subject: Freedom & personal responsibility good, serfdom & tyrannical control bad.
From: bob@redhat.com
Dear Slashdotter,
The wild and heated debate about Red Hat 7 in recent days has been interesting to follow. It demonstrates the strength of the open source model. By comparison (I'm not sure if anyone noticed this) Computerworld had a front page story a couple of weeks ago about how there were problems with Solaris on Sun's Enterprise systems, but that these bugs were not well known because Sun was making their customers sign NDA's (non-disclosure agreements) before helping them fix the problem.
Consider the contrast between a proprietary vendor's unwillingness to debate the merits of their technology with the open debate that Red Hat Linux enjoys.
This discussion is of such value to the users of Red Hat products that we feel little need to even attempt to comment. Informed readers can read all sides of the debate, download our products, test them, and decide for themselves whether our critics or supporters are correct. Of course the readers who post things like "well I haven't tried RH7 but I've heard..." aren't very helpful, but I trust most Slashdot readers to see through that kind of stuff.
There is one recurring comment that I could not resist addressing. Namely the regular habit of our critics of comparing Red Hat to Microsoft. I just don't get it.
There are many things for which we should be justifiably criticised (I have no idea what these might be, but I'm certain they exist ;-) but trying to act like Microsoft is not one of them. Red Hat's business is built on solving the problem thatMicrosoft's business model has imposed on the software user since Bill Gates disagreed with the members of the Homebrew computing club back in 1980.
The software industry that Microsoft has been the role model for is built on the premise that customers are not to be trusted with the technology that they are building their organizations on. The legacy software industry is built on the proprietary binary-only model where not only does the user not get the source code he needs to make changes, but worse he receives the product under a license that essentially says that if you make any improvements to the technology you are using, if you solve a bug that is causing your systems to crash, or add a feature that your users or customers desperately need the vendor can have you thrown in jail. (If you don't believe me, just read any shrinkwrapped software license). This kind of business model, where the customer is completely beholden to his supplier exists in no other industry in any free market that I know of. It harks back to the old feudal systems of 12th century Europe.
Red Hat's business success is owed to one simple benefit our products and services offer that our larger binary-only OS competitors do not. Namely that our commitment to publish the code that we write and distribute under open source licenses enable us to give our customers control over the technology they are using to build their systems. We cannot promise to deliver perfection. All we can promise is to acknowledge the problems immediately and work with you to fix them publicly and in real time. With control over their systems our users can simply build more stable and reliable systems than the binary-only model allows.
This is why the fear that Red Hat is somehow going to wake up one morning and abandon our commitment to open source is so mis-placed. Open source provides us with -the- competitive advantage that enables us to compete effectively against much larger competitors. To abandon open source is simply not in our customers interest and hence not in Red Hat's financial interest.
So if you want to criticise us for shipping gcc 2.96, you have every right to do so - you'd be wrong, but it is at least a legitimate debate and I'd respect your opinion. But to compare Red Hat to Microsoft indicates an ignorance of what is driving our success.
Remember that this debate was begun by someone going to Red Hat's public site and trying to add up all the registered bugs in Red Hat 7. When was the last time Microsoft (or any other legacy software vendor for that matter) gave you access to their complete bug registration system? Which software model do you really want to see succeed? One where you have to trust your vendor (who can and frequently restrict access to information you does need) or one where you are in control of the technology you are using?
We may be making mistakes - that up to you to decide. Some of them may be important to you and while I have no doubt you will point them out to us, you have control over the technology you are using. We work hard to build products that please most of our users most of the time. But if you don't like something about Red Hat Linux you don't have to use that feature or function. We simply are not pursing a business model that bears any resemblance to Microsoft's, so just quit it.
The next slashdotter who compares anything Red Hat does to Microsoft will be punished. The punishment will be to find the nearest blackboard and write "freedom & personal responsibility good, serfdom & tyrannical control bad" seven hundred times.
Cheers, Bob.
I noticed that you used the term "Slashdotter" in your letter several times. This is incorrect.
The correct term is "Slashbot".
Please take note of this in your future correspondence.
Regards,
AC
However I think that there is something in the worries expressed by many about Red Hat's decision to include a gcc that the gcc people have publicly stated is not up to the standard of an official release due to incompatibilities.
Consider that Red Hat do engage is a policy of encouraging proprietary vendors to release their software "for Red Hat Linux" offering binary products only.
Now consider what happens should one of those binary only products is compiled against Red Hat version 7.0 and the product makes use of C++ in it's development.
That product will then only work on either Red Hat 7.0 or another platform that chooses to discard the views of the GCC steering committee.
Red Hat is the biggest player in the Linux market. They control the biggest share of the biggest market for commercial Linux use and they know it. By including "GCC 2.96" in Red Hat Linux 7.0 they are creating a situation of vendor lock in for the users of any proprietary software that is released "for Red Hat Linux" and compiled on the latest version.
I am sure that this is not a deliberate policy at Red Hat - were it deliberate it would certainly suggest a certain leaning towards the position of companies like Microsoft - however the effects remain the same: vendor lock in.
For Bob to reply to such critisisms with "You're wrong" without offering any explanation suggests arrogance. That is the kind of thing one expects of Microsoft. Red Hat need to offer a better explanation than "you are wrong".
Dear Bob,
I admire your company greatly and have had very productive relationships with Red Hat and Cygnus engineers going back many years. I am a member of the GCC steering commitee. I wish you nothing but success.
However, you do have a problem with openness that you are not acknowledging. There is one sense in which your practices do resemble those of Microsoft: your practice of keeping outsiders in the dark about upcoming plans that will affect them. To be specific: your management ordered its employees, including those who were members of the GCC Steering Committee, not to discuss anything about your plans for Red Hat 7.0 with other members of the committee. Advance discussion could have led to improved quality in 7.0, better relations with the outside developers you depend on, better planning by your customers and a whole lot less anger against you.
Joe Buck
They're shipping 2 compilers because their userland compiler doesn't work
Untrue. It works almost perfectly. There is, so far, one known problem that fails to compile code that is ok (and that's currently being fixed).
It can't compile 2.2.x kernels because of bugs in the kernel code (ask Alan), that's why the two compilers are needed.
This message is provided under the terms outlined at http://www.bero.org/terms.html
Please send bug reports on these, including sample code.
FWIW, all of Red Hat Linux 7, including Powertools and the Extra CDs for the European edition have been compiled with 2.96 using -O2 -mcpu=i686 -march=i386 without problems.
KDE 2.0 is completely c++, doesn't cause any problems with 2.96 -O2 -mcpu=i686 -march=i386.
This message is provided under the terms outlined at http://www.bero.org/terms.html
Quoted from Bob's article:
But if you don't like something about Red Hat Linux you don't have to use that feature or function.Hi Bob. My name is Lawrence Wade, and I have root on four Red Hat Linux boxes.
I'm still a relative Linux newbie. Though I was on the Internet (Arpanet) in 1988 and am therefore very familiar with UNIX heirarchy and commands, my first Red Hat Linux system was my first attempt at administering a UNIX(-like) box.
So, looking around, everyone told me how wonderful Linux is, and how light and easy on resources it is.
At the time, Red Hat 6.0 was current, and I installed it on a 486DX2-66 that I had kicking around. (Linux doesn't need much hardware, right? One would expect that the geniuses at Red Hat wouldn't expect that knowledgeable computer users are going to give a new OS a spin on their main computer, right? One would *think* then that a 486 would be a reasonable place to try out Linux.)
My first problem came when the installer detected that I had a monochrome VGA monitor, and set my text to the same color as my background. I plugged in a color monitor, and still couldn't read it. A reboot with a color monitor on, then a swap back to the monochrome monitor after booting, and the installer was still legible.
Next thing was, RedHat 6.0's installer asked me if I had PCMCIA card slots. This was a VESA-bus 486. I indicated NO.
The installation continued, and then finished. I restarted the machine:
"LI"
Stuck in the rescue disk, booted off that. Eventually found out about the LILO >1024 cylinder BIOS issue. Oops. Not Red Hat's fault, sure, but new users don't understand enough to distinguish that.
After I finally got the machine to start up, the machine hung at "Bringing up PCMCIA services". Still being a complete neophyte at the time, I had no idea how to go and kill that from the machine's startup. You'll note again that when the installer asked me if I wanted to install support for PCMCIA services, the response was probably sent to /dev/null; the installer apparently did what it wanted to, independent of my input. It took instructions about as well as my cat. With a cat, that's cute. With a computer, it's not.
Frustrated to all hell by this point, having wasted a weekend farting around, I took my RH6 disk and threw it across my home office. It landed behind a desk. Windows 95B went back onto that old, occasionally-used 486.
A few months later, I decided to get DSL internet service, and there was no way that I was going to use Windows for my gateway/firewall. So, I hit the Linux websites and got a list of supported network cards.
I rooted around under the desk, and found my RH6 disk under a dustbunny and an empty coffee cup. This time I knew that the LILO bug really only affected 486-vintage machines. I had a Pentium 133 ready to go. With 2 identical and supported ISA ethernet cards installed.
I'm not an idiot. First off, these network cards were set properly so that they didn't have any conflicts. Even Windows 95 was happy with them. Yet, every time RH6 tried to use the second card (eth1), a kernel panic happened. This continued until, in frustration, I replaced eth1 with a PCI NE-2000 card. Immediately, things worked.
After RH6.2 came out, I put it onto that machine, and then swapped back to my original matched-pair of ethernet cards. The machine worked like a million bucks, and has ever since.
So, Bob, what is this, a rant without a purpose? No.
I run Red Hat because, from the perspective of a newbie, Red Hat makes the most sense. Information is readily available. And 6.2 has been very good to me.
And while I realize that you have shareholders who will lose interest if you don't frequently bring out new releases, I'm afraid that most people, upon going through what I went through with RH6, would decide that Linux wasn't worth the trouble.
All the advocacy in the world isn't worth squat if the impressionable Windows user goes out to the local software store and buy a copy of Red Hat 7 to discover that nothing works as it should.
It undermines the Linux movement as an alternative to the scourge of Windows. And while I'm very sensitive to the fact that you have to keep your shareholders happy, discouraging new users by releasing buggy software in colorful boxes only serves to hurt Linux.
We simply are not pursing a business model that bears any resemblance to Microsoft's, so just quit it.New users don't distinguish between distributions - they don't know the difference between SuSE, Caldera, Debian, Slackware and Red Hat. To them, Linux is Linux. If Red Hat is flaky ("and golly, they're a big name!"), then *all* Linux must be flaky.
So, don't be surprised when someone falls asleep at the switch or buckles to shareholder pressure, releases a shitty version of your operating system, and Linux users all around the world start to compare you to Microsoft.
No one asks for perfection. That's simply impossible. But even Windows 95 Upgrade was more stable, reliable and functional than RH6. If RH7 is anywhere near as bad as I'm led to believe it is from here, let the comparisons begin.
I run RH 6.2. I like RH 6.2. I look forward to when your firm releases 7.2; until then, I won't be upgrading any systems. And I'll keep on burning CD copies of 6.2 for my friends, with the explanation that Red Hat bowed to shareholder pressure and released 7.0 before getting it to work properly. <sigh> Sounds just like any product Microsoft has ever released.
Fire and Meat. Yummy.
I understand your frustration with this scenario - many of these bugs are not with your code but with others, but there appears to be frustration all around. I applaud you for publishing your errors and fixing them, as well. However, there are certain things I can't abide by.
I'll give you a personal example of what's griping me. I installed the Inti class library because I was looking for a decent object oriented framework under which to develop Linux/GNOME apps. I can't even compile a basic tutorial program with it, because the gtk library that got installed with RedHat 7.0 is a devel one that's prone to crashing, deprecation problems, etc. The error message advised me to go and get a different version of the GTK in order to compile a basic "Hello World" style program.
I'd understand it if I found a complex bug. Really, I would. But what I found was something anyone could have found by taking a blank machine, installing RH7.0 and then trying to compile a program that RedHat has on its website as a tutorial.
A packaged, paid-for distribution is NOT the place for devel code, for alpha code, use at your own risk code, strange new worlds of compilers that can't even build the kernel, etc. Make that an option if you wish.
The best way to continue in your business is to get out of the CD world entirely - except to publish one on ISO and at nominal cost as a baseline, and then have a subscription-based service whereby when I log on to my RedHat machine I'm advised of new developments and offers to upgrade given packages.
--- Jump!! Fire!! Bullet time!! - Lego version of the Matrix
Ok, This Is gonna get a lot of people pissed off. In My opinion Red Hat Is doing the right thing by taking charge. They might end up looking like the new microsoft of the linux world, but someone needs to. The standards that the dominant linux distro makes should be followed. I've made comments before about netscape giving in to IE's standards. Why Not? This would make everything a lot easier for the rest of the world. But Oh Well, Just My opinion.
Btw, Make A Reply to this, i really want to know other people's opinions.
I have found that those who bash Red Hat the most are often newcomers to Linux who have maybe been part of the community for a year or perhaps even less. I am a bit confused about why they do this, perhaps they think Red Hat bashing is cool. I have been using Linux since 1994 and I think Red Hat is a great company! They pay people like Havoc Pennington to work fulltime on GTK 2.0, Alan Cox to hack on the kernel and many many other. Also, since when does Slashdot represent the community? Most people here don't contribute anything, the core community has always and still consists of the various mailing lists (the kernel, debian, mandrake, redhat and many other lists). Just MHO.
Welcome to leap before looking school. You have successfully leaped to a conclusion without looking at the surrounding facts. To help you better understand your position, here are some facts that you may have missed:
This is software intended to help online merchants interoperate with existing credit card clearing houses.
Red Hat (or more likely, the company they bought CCVS from) probably had to sign a lot of NDAs in order to get access to the specifications for those proprietary protocols.
Thus Red Hat has released the software with a license that complies with the legal requirements, but still allows merchants to build Linux solutions. This is a GOOD thing. For details see this.
Are you moderating this down because you disagree with it,
We call it art because we have names for the things we understand.
It's always nice to see an open letter, and even nice to see that Red Hat are actually watching the flow of conversation on Slashdot.
When something goes wrong, it's so usual for someone to sit around and say 'not my fault', that it's nice for someone to sit down and honestly say 'We may have made a mistake, perhaps.. But we made decisions we believed to be right, and you're quite at liberty to disagree..', and actually take the time to read the arguments against, and weigh them up.
I guess there's no progression without controversy and dissent over the paths to follow, and somebody has to spark controversy.
Maybe I speak just for myself, maybe for many, but thanks Bob, for paying attention to what we have to say, and for letting us know that you are, indeed listening to our many voices.
Cheers,
Malk
Well, there is a lot here to digest, both in terms of what Mr. Young wrote, and in terms of the larger debate about RedHat. Personally though, I feel that there is, in fact, a lot of paranoia in the Open Source Community - and rightly so, for we effectively LOST our community back in the early 80's. So here we are, in fear of RedHat becoming another Microsoft, or at the very least, a Sun (which I find to be an infinitely more likely situation, and a totally different debate altogether). Well, I can say in all honesty that I have used RedHat's products before, and I have kept up with all the various claims against them - and as a Debian user, I can say that I believe RedHat has done NOTHING to deserve the kind of claims against them that people have levied against them.
Now, this issue is moot in my mind - Young, in effect, seems to be feeding the trolls more than anything else here. No one who had actually researched this debate would claim RedHat to be acting like Microsoft. Again, as a Debian user, I feel that there is a hell of a lot WRONG with RedHat's distro - for ME. This more of a personal taste though - and don't anyone dare tell me that Debian hasn't had it's share of crippling bugs.
In the end, this is all just incredibly silly for me, and I have finally made my decision about RedHat. RedHat is a company, like SuSE and the Turbolinux guys... they are selling a product. RedHat has supported it's product admirably, and though it isn't in my decision the product that I want to use, that doesn't mean that when something goes wrong we should all point our fingers and scream "REDHAT IS BECOMING MICROSOFT!!!" RedHat, as a company, had done a great deal of work for the Linux community and for the acceptance of Linux in places that in turn have made Linux a better OS (IBM, Dell, etc...). So I think we should back off and turn down our flamethrowers here - if you want to criticize RedHat, fine, but at least find a good reason.
Know ye not that ye are Gods???
Most of the criticism of RedHat seems to be coming from college students with a selective dislike for what they see as large, successful businesses. I say "selective" because these same critics also fawn all over corporately produced mass market products, like Mountain Dew, The Simpsons, and anything Star Wars related.
One great thing about Linux distributions is competition. Don't like RedHat? Buy SuSE or Turbo or Corel or Stampede or Debian or one of several others. If you want to jump on corporations, then there are some pharmaceutical companies and food processing companies that really need public vehement public criticism of their practices. But don't bother pointing the same kind of fire at RedHat.
You are missing that any large software development effort, if they are honest and have high standards, must admit that they have zillions of "bugs". Most don't make the product unusable or even bad; instead they are things that could be done better, or that some person doesn't like based on taste, or are just notes worth keeping around to think about. (This isn't even counting the many bogus, incomplete, and duplicate bug reports.) There is in fact good reason that most organizations call it an "issue" database instead of a "bug" database.
The 2000 (or whatever) "bugs" in RedHat's bugzilla is a meaningless number, and by itself is no cause for alarm (by users), shame (by RedHat), or criticism (by sanctimonious know-nothings). Go count the bug's in Debian's database. (Don't bother telling me when you get finished, because you won't.)
I think Bob was trying to say that counting the bugs is a silly response to the availability of RedHat's bug database. A more "enlightened" response is to appreciate the value that the mere existence of a public database gives you.
The evaluation of an action as 'practical' . . . depends on what it is that one wishes to practice.
From: bob@redhat.com
Dear Slashdotter,
"it appears that you're trying to write an Open Letter. Would you like some help in choosing defensive words and phrases?"
--
--
"It is now safe to switch off your computer."