Slashdot Mirror
Alan Cox on The Risks of Closed Source Computing
cd-w writes "Alan Cox has written a column for osOpinion regarding the risks of using closed source computing.
A sample quote:
'No company now would commit to a closed hardware strategy. It would cost them more than using commodity components. Just as importantly, it would commit them to a single source for support and parts. Why then do they commit to a single software supplier? A closed source strategy exposes the company to serious business risk. As many telephony companies have discovered, your OS supplier might suddenly decide to be your competitor.'
" As always, Alan writes well and explains why the commodization of information has critical bearing on the success of the open source movement.
The theory behind a $300 computer is that it's going to be more feasible to buy another $300 computer thats twice as fast in 18 months than it is to pay a $50/hour technican to upgrade the existing thing. (Maybe you wanted to play quake on that computer, but from a business's standpoint, a $300 "closed" computer should be just fine for most tasks.)
So is this a "closed" hardware strategy? Not really in the classic IBM Mainframe sense. Any investment that you can throw into the trash and forget about without a second thought seems pretty 'open' to me.
Business. Numbers. Money. People. Computer World.
I'd think that the difference is that any company can pick up a competitor's product.
Still, the innovator or developer of any product has three advantages when it comes to selling.
They are the most familiar with the existing code base: this is very important when it comes to porting a given piece of software. RMS, for example, charges usurious rates to port GCC. Any complex piece of software is going to take a considerable amount of time and thus expense to familiarize oneself with.
They hold the copyright, and can offer alternative licensing terms. This may be important to some companies, who don't want an "open" piece of software or who may want to redistribute their modifications without including source and who will pay for source access.
Finally, they gain name recognition, credibility and thus consumer preference through the product's use. Even though Jim Joe Bob can put togeter a linux distribution based on Debian, they certainly aren't Debian and don't have the credibility of Debian. Red Hat is an even better example of this.
--
--
There is no premature anti-fascism. -Ernest Hemingway
I was a little disappointed to see Alan jump into this debate. I believe his strengths as a technician are unassailable, and that his greatest value to the community is to make Linux the best technical solution, period - which simply makes all the other questions about open vs. close source, NT vs. UNIX, Microsoft etc. moot.
Aside from the fact that the computer-car analogy is one of the most overused analogies in existence, he does not add much value to the open source side of the debate.
There are a few exceptions. For example, the comment about NT-Alpha users is dead-on! I'll bet *their* opinion on open source is changing dramatically right now.
But he has trivialized the other questions, which are primarily business issues that are far more subtle and complex than he presents. For example, he questions why anyone (i.e. Windows users) would commit to a single source for support and parts. Well, the fact is, businesses do *not* usually make their commitment to Microsoft, they make it to Microsoft's partners, for which there is a profusion of choice! If the Windows NT support from Dell or my reseller is insufficient, I can go to Unisys. If that doesn't work out, I can go to IBM or Compaq. The same goes for buying a Windows system. Granted, Microsoft lies *behind* all of these entry points and is ultimately responsible for resolving issues, but that fact rarely becomes visible to the user. From the user's standpoint, there are enough choices that he can worry about what matters - running his business. Now, of course, Linux provides far more flexibility for the *suppliers*, which is why the various OEM agreements being put into place are the most significant evidence of Linux's progress. However, that has little to do with decisions made by users - most will accept what the system vendors provide, as long as there is choice at the levels that matter: hardware and support.
Alan should stick with what he does best: making Linux real. Everything else just adds to the noise.
By the time you get a company-wide database that meets your needs, you have 5 or 6 figures tied up in custom code, specific to one vendor. So to switch, you have to redo all the custom code, convert your priceless data, and retrain everyone.
Are you saying that people in your daily life are constantly extolling the virtues of open source software? If so and they're getting on your nerves I suggest you talk to them about it.
Or are you saying that you make a point of reading sites where supporters of open source software post and then get angry because erm.. there are posts supporting open source software?
The "closed hardware" companies who expect to receive direct return for their development have a very definate need to be different for exactly the reason that their gear costs more. This leads to innovation in the areas that their customers require, thereby supplying a better valued product. Various companies such as IBM, HP, Sun, DEC.. er Compaq etc. supply *LOTS* of closed hardware systems to customers that are happy to pay the cost, because the hardware meets their needs.
In the "open hardware" area (notably the PC industry) we have hardware that is quite frankly rubbish. This hardware has the advantage that because of it's cost it has made computing more accessable and allowed small hardware designers experimentation in various areas. However because of the entrenched compatibility requirments (I still have an XT ethernet card in 1 linux box) we have hardware that has lots of realy broken features. (10 IRQs !!!!).
Closed hardware manufacturers have the ability to throw away computer models that are no longer relevant, and thereby cleanly advance hardware architecture in a way that the PC industry can't.
A lot of what's new and great in PC architecture has been hanging about in proprietry systems for a long time (15 years or so difference for memory managment when the 386 arrived - another 3-4 years before software that could use it caught on).
OSS gives me the best of both worlds - I can run my open software on "closed hardware" :-)
You really seem to have trouble understanding the basic concept. You find a bug in the software, you can pay someone to fix it, either a one off payment for that fix or as an employee, or a contract to support the software for the specified purpose including bug fixes. No hoping that Microsoft prioritise it, as the source is available you can ensure that it is fixed.
And not just bugs, but extending or customising the software to meet your purposes.
It's with closed source software that your only option is to hope that someone addresses your problem, because you can't do anything about it yourself except complain. With open source you can ensure that the problem is resolved.
>Sigh. When will Slashdot newbies learn to write HTML for clickable >links? ;-)
Most likely never,since they have far better things to do than fuck with HTML or XML or whatever the new fad of the week happens to be....
I think your suggestion makes my point about commodity hardware in PC space.
External highly available RAID box
= even more proprietary than Compaq
= Better because it gets the job done better
Business. Numbers. Money. People. Computer World.
The whole article made complete, complete sense.
Until the final footnote.
[1] As an aside to the main discussion, one reason that the Beowulf clusters have been so dramatically successful is that it is a set of open source software that allows the replacement of proprietary closed mainframe hardware with standard PC components. The Beowulf thus delivers a double blow against proprietary competitors. It breaks the existing mainframe lock-in and it breaks software lock-ins.
Beowulf clusters replacing mainframes?!
Beowulf clusters replacing mainframes?!?!
Beowulf clusters replacing mainframes?!?!?!
What company/school idiot would use a mainframe for handling PVM/MPI technical computing jobs, even pre-Beowulf? Talk about the wrong tool for the wrong job! Mainframes today are used for backwards compatibility with old COBOL/REXX/JCL/etc. code and for certain larger commercial batch and transaction processing. Beowulf replacing RISC/UNIX technical servers, I could understand, but mainframes?
Name one company that has replaced a mainframe with Beowulf.
Am I missing something here?
Weirded out,
--LP
Unfortunately, I can't get to the article, so I can only respond to your post. Yes, what you say is true. Vendors are always trying to find ways to trap you into being dependent on their HW or SW platform, and then milk you for every penny they can. The major commercial UNIX vendors always charge 2x markup on commodity hardware items like memory, disks, tape drives, printers, etc. They also require you to spend anywhere from $2k-$20k per year, per machine for service contracts. And if you don't have a service contract, you don't get software patches/updates and you have to pay out your ass for other software upgrades.
However, there are a lot of companies who need systems that are damn near 100% reliable and are willing to put up with price gouging to get what they need. Banks and financial service companies are two such cases. Paying a company like Sun a million dollars a year for support contracts and overpriced memory and peripherals is acceptable to a company who can loose millions of dollars in a matter of hours if a system goes down. Such companies honestly don't care whether the hardware and software is open as long as it does it's job and meets their reliability and scalability requirements.
Similarly, a lot of companies need the kind of processing architecture only a large SMP closed hardware solution currently provides. You can get better theoretical performance/price out of a Beowulf cluster than a large SGI Origin, but if your application isn't trivially parallel you will quickly saturate the communications bandwidth of the Beowulf machine. Also, it seems that a large percentage of current software systems fit well in a shared memory machine of 8-32 processors, which is too much for Linux. It's also too little to be worth building a Beowulf type cluster for, especially considering you have to abandon the convenient shared memory model. Another consideration is that the company needs to spend a fair amount of time (a precious commodity in today's development schedules) engineering, building, and deploying a Beowulf style cluster. In contrast, a company like SGI or Sun can have a big multiprocessor machine installed and operational much more quickly.
If all else were equal, I don't think anybody would prefer a closed HW & SW solution over an open one. However, there is no open solution that can currently take the place of the big commercial UNIX (and VMS & AS/400) vendors when it comes to reliability, scalability, and relative ease of deployment. Linux has come a long, long way since I started using it in 1995. However, it has only penetrated a few commercial niche markets so far (although web servers are a pretty large niche). FreeBSD is at about the same level, albeit with less commercial SW and HW support. WinNT is really the first OS project that was ambitious enough to try to take on the whole market at once and be the single OS for everything. It certainly hasn't lived up to expectations.
Linux appears to be headed the same way. Many people are trying to push the widespread adoption of Linux for all sorts of applications that it isn't suitable for (at least not yet). The point that many advocates miss is that aside from the desktop (email & office application) segment of the computing market, the larger market is a collection of smaller niche markets. Unless something truly revolutionary happens, no single OS or HW platform can satisfy the needs of every niche market. That is why (at least for now), companies need to consider the suitability of each platform to their needs first and the openness of the platform second.
And I thought he liked programming in perl! For shame!
--
Man is most nearly himself when he achieves the seriousness of a child at play.
****Gfx Scrollbar Special case hit!!*****
The biggest problem with OSS is that there are always people willing to fix your software for you or upgrade it for free. That's a hell of a big assumption. Right now, the economy is good, so there are lots of college kids and uber-geeks with enough free time to write code for free, great. What if the economy goes in the toilet? What if OSS isn't "cool" anymore (I don't happen to think that it is)? OSS relies on people's good nature to get software written. That's just plain crazy! No company, with any kind of clue, would base something as important as their computers' OS on people's good nature. Anybody who does, deserves to, and more than likely will get hit in the face for it. You simply don't get shit for free. You're kidding yourself if you think that you do.
You make an erroneous assumption. The affects of 'Good Will' on the OSS industry are secondary to the driving force of Self Interest. The people who hack on linux are people who want their OS to do more than the commercial alternatives. No one does work on OSS just out of the goodness of their heart, they do it because they want to see a certain feature develop for their own use. They don't care if anyone else uses it, as long as they get to use it. Good will is secondary.
Kintanon
Check out JoshJitsu.info for Brazilian Ji
No company now would commit to a closed hardware strategy. It would cost them more than using commodity components. Just as importantly, it would commit them to a single source for support and parts. Why then do they commit to a single software supplier ?
:-)
This one statement is just so unbelievably wrong..
Now before you get all upset, I love open source as much as any of you. OSS is a great, good, noble thing.. So keep that in mind..
It's fairly obvious that Alan Cox really doesn't fully understand the business mind, which is understandable, since nobody but a business mind can grasp it.
Companies DO commit to closed hardware solutions. Companies WILL put pretty much their entire revenue in the hands of their suppliers. This happens often. Daily. Hourly.
Why? Companies are run by people in management who don't care about "the right way" to do something. Heck, most of the time they don't bother to analyse the long-term effects of their actions. A company wants to make money, and they want to make it NOW. Sure, they'd like to make it later too, but if we can make more NOW, that's the important thing. I have to deal with this mentality all the time.
Similarly, in a production environment, downtime is unacceptable. Companies need guaranteed, strong, support. But they need something else: they need multiple sources of support.
Of COURSE a company would rather buy a closed source software system, and trust them for support. Why? Shifting of blame. Remember that a company usually does not make the decision to have a software system designed for them. No, a person in management who works for the company makes that decision. If it doesn't work, he gets the blame. By going with a closed-source app custom made for them, they have someone to blame when it breaks. They have someone to push around to fix it. The guy who made the call to get this system probably will keep his job when it fails (and all systems will fail eventually).
However, let's say he's got brains. Let's say he knows that the open-source app is the better way to go. In terms of support, you can't beat it. Thousands of people working on it? GREAT! But, the downside is this: How do you push around people who work for essentially nothing in their spare time? Quick answer: you don't. You have no leverage. If your system breaks, and they can't fix it quickly, you're more or less screwed. Even if you KNOW this will never happen, you can't prove it to a board of directors, can you? The simple possiblity that it could happen is enough to make you reject the open source concept. Better for it to fail fairly often and have someone who you can blame than to have it fail rarely and have all the blame yourself.
Now I don't say this is right. But this is how the management in the companies I work for (read: program custom apps for) think... This is based on my limited experience, your mileage may vary, type of thing...
One more thing:
In many ways the motor car is a very good example of the fact that the open source model is not something revolutionary, as Bob Young is so keen to point out - it is the model we use in almost all serious grown up industry.
HAHAHAHAHAH! One of those things I program custom apps for is several motor companies. Many programs I have written are on the line at several plants now. I have to go to these places to install and maintain these apps I have written. This stuff is a closed-source as it gets, my friend. Automobile manufacturers are one of THE MOST custom application, closed-source, driven industries in the entire world. WHOLE SYSTEMS that exist on these production lines are not only custom software, but custom HARDWARE from ONE COMPANY. One company makes the hardware AND the software. If that system goes down in any large way, SO DOES THE LINE. Neat, huh? Get a tour sometime of your nearest car production plant. Watch the systems hanging from the line very very closely. None are open-source. Not one.
Just my $0.02...
---
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
Yes, open source is better for the end user. It's better for the developer (typically). :)
It's very bad for companies who don't have a business model designed around open sourcing their hardware and software.
Sun, Microsoft, Novell...these companies aren't evil for not open sourcing everything. Open source is a new model, and if it's going to become "the way" of doing things, it's going to take a LONG time. These companies aren't going to shift paradigms on a whim...that's not profitable. That's right...it's all about making money...
I've placed a state-side copy of Alan's text-only piece up for those who are finding the trans-Atlantic link a little slow. Hope he doesn't mind.
--
No company now would commit to a closed hardware strategy.
Companies continue to commit to close hardware strategies every day. Look no further than your nearest IBM shop -- Do you know that AS/400s are selling like hot cakes? In that context, Sun hardware looks pretty open, because you could add commodity memory and disk if you weren't worried about the support.
Even PC Hardware has lock-in. There's a reason our server room is completely filled with Compaq Proliants (it's the proprietary hotswap disk subsystem), and not three from each company. "Commodity" hardware is really only well suited for commodity tasks - companies will happily pay through the nose for proprietary stuff if it gets the job done better.
Business. Numbers. Money. People. Computer World.
However, he's addressing managers and techies at large here, and notice how he doesn't try to sell Open Source, but rather outlines flaws in the Closed Source model. That's the best way to go about it: point out a loss opportunity for profit, then propose an intelligent alternative.
Is it FUD? Well, no. FUD spreads ideas that are barely grounded in fact. This is advocacy: it is a clear argument with a definitive target audience, and it exposes a flaw of Closed Source quite simply.
I think all OS advocates should take note. Selling advantages of OS may not be the way to do it, because managers believe they already have a system that fulfills all their needs, and will be wont to change for the simple promise of more features. Managers won't switch from NT to Linux for the same reason they take forever to upgrade, say, from Solaris 2.6 to 7: they want to play it safe, and new features won't come into play unless they know they have a definite disadvantage.
However, when you present an OS advantage as a shortcoming in CS, you're speaking their language. You're showing them where they're not making money, and to a manager or CEO, not making money is worst than losing money.
The problem is that we geeks and them managers speak a different language. Contrast the following statements:
"Linux offers greater stability than NT."
"NT suffers from outages that cost money to the enterprise. One solution is to use a more stable platform, such as Linux."
Sure, we'd love to sell OS's merit on its own. Someday, when OS has gained mainstream recognition, we'll be able to. But for now, I think Cox's strategy is very sound.
"There is no surer way to ruin a good discussion than to contaminate it with the facts."
I know there are some people with a deep seated and sometimes inexplicable opposition to the GPL (Brett Glass mysteriously leaps to mind) but I'm sure there are those who don't actually realize that you *can* sell your stuff under the GPL - in fact, I had this discussion only yesterday with a guy on a private news server. He was genuinely under the impression that the GPL forbade you from making commercial gain from your code. He had no objection to giving away the source to his product but he wanted to sell it and services based on it. And why not?
Maybe the OSS community needs to make some education type noises, as well as pointing out the benefits of the development model.
--
Peter
We're still back to good will. Who's to assume that those people will continue to make their code available to the public? I write apps all of the time for my personal use, but I certainly won't make them public. They took me a lot of time and effort, and if somebody else wants to use them, they can pay me. Under OSS, you're still ASSUMING that people want to make their code free to the public. Not a good assumption, considering there are many people, such as myself, who would never do that.
Many of the people I deal with on a daily basis (I work for a small software company) will proudly proclaim that they are "A Microsoft Windows Shop" and that they only run NT...like it's a badge of courage. I've never understood this, I'll sell them an overpriced NT version of our software and take advantage of them, but I'll never understand them. Why would you pledge allegiance to a company, any company. Why not be more concerned with getting the proper tool at a fair price ? That's not to say that a MS NT solution might not be the best tool sometimes, I just think it's ridiculous to go through the day with blinders on.
The power of marketing is frightening to me.
I think you meant to say "Granted, Microsoft's lies are *behind* all of these entry points".
;>
You seem to have forgotten that the reason competition is good is that it allows you to determine what needs to be done. If GNOME/KDE were to have merged back when people started bitching about it, they would be nowhere near where they are now.
Not only would they have spent enormous resources talking about what to do, they would have missed out on the opportunity to do a little 'my feature is better than yours' nagging which is always productive. If you don't have someone in mind to beat, you slack off. If you are convinced of your superiority, you end up losing (just like what is happening to M$).
I guess my point is that you have to allow code forking in order to get a better end product. Keeping the 'one true source' concept around discourages people from innovating. GNOME wouldn't be as good if KDE didn't exist, and vice versa.
Personally, I can't stand KDE's WM. Am I stuck with it? Not since GNOME came along and encouraged the development of WM's like Enlightenment which will run both. We'd all be stuck with KDE's ugly, designed by committee, CDE clone interface if we had to agree on one standard at the beginning. Without that WM, KDE is cool again. Thanks GNOME!
Tony
There is no assumption that they will continue to code as much as they do for no monetary gain. Or that they will give code to the public.
:)
But this doesn't matter! The source is out there, they can never retroactively revoke it. (at least with the GPL and other licenses)
So, even if the origional authors give it up, as has happened on so many programs, someone else, a company, a user, even yourself can pick up the code and run with it.
If you write the code and want to only sell it for money, that's your choice. Others make different choices.
Besides, I spent a lot of time and effort on this post. If you take my advice you can pay me for it.
but what the hell are you going to do with a cpu schematic? build one in your fucking basement, i think not you ignorant piece of shit you should die
Problem with URL? Server down?
Error 503 Service Unavailable
When reading a totally unrelated article after my last post, it occurred to me that perhaps Alan Cox was confusing mainframes with supercomputers? Beowulf is replacing supercomputers.
In any case, I find it very odd and mildly disturbing that Alan Cox appears to have so little clue what a mainframe is, even the most basic minimal knowledge.
Hate to accuse someone I admire of gross ignorance, but that's sure what it looks like. Is there any other explanation [for his statement]?
--LP
Alan has the mutual self-interest or public benefit of Open Source down pat & well said. His piece falls however: it rests on the assumption that Software is "Mature", that the basics are known. The basics are NOT known. Look at user interfaces. For comparison, a car driver need know very little, compared to a car mechanic or even a service station attendant, and still get great use from the car. You don't have to be a software engineer to use a computer, but you do have to become 'computer literate' - way closer to the technical side.
The zillions concentrated by Microsoft, Oracle, SAP et al are deserved, as long as they are the main innovative force on the basic, what can it do & how little need I know to use it level.
But let Linux seruiously invade the desktop market (I hate phrases like 'desktop space', space is four dimensional at least...) and things may finally change.
IBM is no longer the bad guy. WebSphere is just a solution to the reliability problem when you have 200-300 or more servers scattered, and the support and logistics problems (no matter what os - even ibms). And if you don't like IBM, you can stop and convert back down again. Free Software is fine, you have all the choices to do what you want. MS does not agree, it would seem. I think IBM wants the service, management and support contract in the long run. Of course big harware helps. And if they take the 'headache' away, both parties win. MS, on the other hand, looses big time on server consolidation. Like a car, software has become easier to use, but more complex and expensive to administer and repair. Say a Huge initial expense, or lots of small expenses that add up to a fortune, and get lost. Bad bean counting means MS looks good Commodity pricing has reduced the prices IBM can charge, and it went open to market pressure. Not the same with MS, and the price of their software is NOT at comodity levels by a long shot, plus is not open. Management is stuck with a propriety OS, whose price and supply are monopolistic, and forces a distributed system. I think we can trust MS not to bend folks over the barrel any more than present. ie known situation. Funny thing - that was where ibm was 15 years ago, yet here is mgt locking into a propriety solution. The circle of life
Hint to OS opinion: Stop having good content. Your server can't handle the load! :)
After all, electrical service in many areas started to suck after it was made public domain and became regulated as such.
Please illustrate this theory of yours with a real-world example.
I can't think of any.
Dude?
Are you calling Alan Cox a newbie?
Hewlett-Packard is a big offender. We had a rather large number of old model 730 RISC boxes. During servicing by HP under warranty, they discovered we had non-HP memory. Now, five years ago, replacing somewhere in the vicinity of 8 GB of memory was expensive enough, but we ended up having to replace it all with HP certified and supplied memory. Our new HP Netservers are about the same--only HP parts. Hell, I'm surprised we didn't have to by HP patch cables, hubs, routers, and terminals!
This is indeed the part that I find incomprehensible about businesses: that they would allow a vendor to tie up the business' own data in proprietary secret file formats. If my data are stored in an open format, then there can be competitors in the particular application market. If storing documents, spreadsheets and presentations in a standard open format were a requirement, Office (and all of its competitors) would be priced much lower because of competition. I personally think that this is at least as important as open source. I may eventually choose an open source application to manipulate the data for the reasons that Alan gives, but I also have the option of choosing between multiple closed apps that may have their own set of advantages.
I actually think the automotive analogy is closer to this (open data standards) than it is to open source. All of the apps (cars) can use the same data (roads). Open source seems more like saying "I can replace the piston rods with something different if I want to."
Redhat spends money putting together their distribution, right? It cost money, because someone has to quality test it.
Quality test it? Are we talking about the same Redhat?
The one that should have a "broken package in every release" notice similar to the "prize in every box" notice on Crackerjacks?
But strange things are afoot.
I work in an extreemly beurocratic pro-Microsoft environment. When it comes to IT, decisions are made here that can match any PHB anywhere else. But there are strange undercurrents.
The other day I was happily doing my own thing. From the office behind me comes "Damn this Windows - damn Microsoft". Another admin. doing some trendy Microsoft-bashing? Nope. This was uttered from a decisively non-tech-oriented budget analyst. A year ago, it probably would have been "damn computers". Now Microsoft is getting the blame for her system crashing. The claim is made that the "masses" accept buggy software and computers crashing... that they don't understand the problem. I think this is an example of a change in understanding. The backlash towards Microsoft is coming.
Our organization has had an interesting history. One IT director kicked a hornets nest when he tried to route out Apple from the environment. The Mac fans faught back. It was messy. In the end, the Apple contingent took some hits... but apples are still on site. With this kind of history, what hope does Linux have? One organizations has built their own "standard load" consisting of a dual-boot site-standard Windows and Linux (not sure which distro). Meanwhile, the contractor for the majority of our IT resources has gotten requests to design a linux-based "standard load".
This doesn't mean that life here has dramatically changed for the IT worker. Its still tightly controlled by PHBs with Microsoft sales brochures firmly grasped in their fists. But even here... small cracks are evident. One can't help but muse whether these are signs that the damn is about to burst. Maybe the industry is ready to change how it does business.
We'll all fall asleep first.
AirSick is a pretty boring writer once you understand his style.
Notice : pushed by a vendor. They want to lock you
into their higher ram prices, their higher scsi
disk prices.
Who is getting annoyed - you the customer. Its up
to you (or more likely your boss) to spot the problem.
Alan
Software is a tool. If you like the way Black and Decker's power tools work better than brand X, it may still be worth your while to spend the extra money, even if they come in factory sealed units that need to be serviced by the manufacturer. If you later find that you are unhappy with the quality of Black and Decker, then you use Brand X drills and modify them to suit your needs...
Free thought, free beer, and free love are all great things, but let's not leave out free competition!
--
E2 IN2 IE?
Totally correct. Thats why they go to Red Hat,
to Linuxcare and want contracts. But they can go
to multiple people for that support.
A large closed source vendor can do what it likes,
so if you think about it they are offering 'good will' support - for a fee.
Alan wrote; In the Linux world you have things like the Red Hat Certified Engineer and you have approved support partners.
Certainly this is true, however I would point out that this is not something that I feel currently applies better to the Open Source world than the Closed Source world. Seriously, there are more people out there who have heard of a Microsoft Certified Professional than have heard of a Red Hat Certified Engineer. (Unfortunate, but true).
Here is a place where the closed source people currently have the jump, though in time that could be changed.
How could it be changed? Simple - by being better. While many more people have heard of Microsoft Certified Professionals, they also know that the certification is relatively easy. It doesn't *actually* say much about a person's ability. I have met MCP's who have been worth their weight in gold but I have also met those to whom I would like to attach their weight in lead and drop them in the Thames!
I personally work for a small company, but we get great clients because of word of mouth that we deliver - and deliver reliably and well. This is what companies need to remember as their number one priority if their goal is to be in the support business.
With luck, the Open Source support guys can build up a way better reputation than the closed sourec guys and with time become better know - for quality.
A little planning goes a long way...
What if the economy goes in the toilet? What if OSS isn't "cool" anymore?
OK, What if?
What if Redmond got hit by a Meteor tomorrow, and every MS programmer got killed... What happens to all the bug-fixes and tech support? Can you say "Well, that's no problem, we'll just take what we have and WE'LL fix it!" - No, because you don't have the source.
OR: What if every Sendmail programmer got hit by a bus simultaneously tomorrow instead? What happens to every major mail server on the planet? They just say "That's no problem, we'll fix whatever bug occurs ourselves."
See the difference?
In your examples, if a company who runs OSS stops getting "free" support, then they can HIRE PEOPLE WHO KNOW, because there are lots of them (and any decent programmer can learn, because they have the source.)
What happens if MS goes bankrupt during your (rhetorical) recession? How do you get support, upgrades, bug-fixes, etc? Answer: You don't.
No company, with any kind of clue, would base something as important as their computers' OS on people's good nature.
Nobody is doing that, or asking that people do that - you're kidding yourself if you think that ANYONE is doing this.
You simply don't get shit for free.
Nobody is saying that you do. There is always a price to using any software; with OSS, you hire people who know the system (and you can _always_ find them, as the source is available.) This is the point the article was making.
No, they have some little startups to go to for support. But let's say they need a bug fixed with the code. There's no way a multi-million dollar company is going to stand aside, cross their fingers, and wait for some hacker to fix it for them. Companies need to BUY something, so that if there's a bug, or some kind of problem, they can always go back to the company. And of course, they can always use the court system, if need be. With OSS, they have to just hope that somebody helps them. No smart business would ever run that way. If they have a major problem with closed source, either it gets fixed, or they get paid for damages. With OSS, it either gets fixed, or it doesn't. Now, assuming you were in charge of millions of dollars and thousands of jobs, which would you choose?
How's this for a twist...(Please excuse the fact that not only have I NOT read the article, I also haven't read ANY of the replies!)
When you have closed source software, you can leverage off of open source software. Now before people send me hate replies, let me point out two ways of going about this:
1.) Steal source code. -RED FLAG- You'll probably say that this is ILLEGAL. They could be sued. That's EXACTLY right! They COULD be sued...but ONLY if someone found out about it!
2.) Free ideas. Unless the ideas are PATENTED and a person can prove patent infringement...again, no problem.
Well, this is my premise. Rip me apart if you must.
In fact there are at least two websites set up to match developers with people who will pay them to write OSS - CoSource and SourceXchange.
Tom Swiss | the infamous tms | my blog
You cannot wash away blood with blood
No company now would commit to a closed hardware strategy.
Isn't Apple and HP already like this? You can only buy Apple parts through Apple, to my knowledge. HP is open on the PC end of things, but aren't their servers and mainframes pretty much proprietary? (Don't hold me to the Apple thing, I'm not to experienced in the ways of Apple).
Just a thought.
I think the hardware analogy doesn't quite translate right to open source... The specification for manufacturing the boards are open, so interoperability for plug-in components know how to communicate with the system. BUT, the implementation details of, say, the chipset are very proprietary and not for the hardware manufacturer to look at.
:)
But, if you want, you're free to implement your own chipset, because you can see the specification.
So, in the end, this is probably closer to the Java license than open source, or maybe implementations of different CORBA services...
Sorry to nitpick...
Primary Risk: You get that Richard Stallman - Jon "maddog" Hall - Alan Cox look going.
For the companies like Sun Microsystems that are struggling with their decision whether or not to suppport Open Source, there's still hope. With Halloween fast approaching, a number of stores will be carrying those fake ZZ Top beards, giving everyone from the CEO to the boys in the mailroom the opportunity to get their feet wet with the Open Source way of life.
:)
Cheers,
ZicoKnows@hotmail.com
Gawd, I suck. *thwap*
Cheers,
ZicoKnows@hotmail.com
Are you calling Alan Cox a newbie? Hence the ;^)
Well, well, well; three holes in the ground...
> There is a great support mechanism (the community itself). However, that mechanism, while typically very responsive, isn't truly "responsible", in that no-one is required to support anything
While, for the most part, I have had good success getting help with various problems I have encountered using Linux, it hasn't been easy. I've been working with computers for many years, and I'm pretty good at hunting down information and just hammering at a problem until I come up with a solution. However, there are a lot of people who don't have the time or patience or experience or inclination to do things the way I do. And, even at that, sometimes I just can't find a solution. There have been times when I have posted a request to a newsgroup and either found someone else with the same problem or got a reply within hours. However, there have been times when I've come up empty.
My point is, if no one is responsible for providing an answer, then there is a high probability that noone will answer your cry for help. And, as the size and complexity of the community grows, I have a feeling that the "signal to noise ratio" is going to get so bad that those who are out there trying to help people are going to become overwhelmed to the point where they will retreat back into their Hacker communities where the hordes of newbies won't be able to bother them.
Obviously, the one hope for this situation is organizations like RedHat and others who will set up support contracts, but these will not be for free. While these are great, they lose some of the "openly available" spirit of sources like UseNet and various User Groups. A lot of people try out Linux simply because it's free (as in beer), and if the free supply of support dries up, it will put a serious dent in the growth pressure for Open Source software.
(In other words, we're doing well, but the battle isn't over yet.)
Your Servant, B. Baggins
Again, it doesn't make any business sense. Then, granted, you're benefiting from the new code, but so is everyone else, including your competitors. If a business were to pay someone to write OSS, for these very reasons, they would not want to pay very much, if anything for their efforts.
Is this a troll, or are you really missing the point over and over ?
If a company wants guaranteed support, they can BUY it. But they can BUY it from more than one vendor, so they get the support they require AND competitive tendering. Win-win.
osOpinion recently switched to a new hosting company... (You know... one of those... "unlimited everything for $8.95 per month? places) Well, as I learned after as little as 2 weeks on their servers and 2 weeks worth of bad service, I learned that you get what you pay for. I have since terminated my service, and sighned on with a new company who appears to be more reliable. (Lets hope)... I won't know tfor sure until the transfer goes through.... lets all cross our fingers...
I have been on the phone to icoms (the current host)'s tech support personel every single day for the last two weeks, and If you can believe it, they keep saying that they don't see any problems! As a matter of fact, I just got off the phone right now after speaking to one of their technitians... He said, "I don't see any problems. It came right up for me" Then he sarcastically said, "Perhaps we have some magical computers that allow us to see it while yours don't?" So, I then told him to go to slashdot and read about all the comments everybody posted...
Him: "huh, slashdoot?"
Me: Slashdot!
long pause...
Me: Its a tech news site!...
long pause...
Me: www dot S-L-A-S-H-D-O-T dot O-R-G
Him: Slasdot.org?
Me: Slashdot! with an "h"!
Him: oh... ok... um, it came right up for me, I don't know what the problem is...
Me: Sigh.... No, this is the site that will tell you the error messages. Scroll down a bit, till you see the headline, "Alan Cox on The Risks of Closed Source Computing" and the penguin...?
Him: ok. I'm there
Me: ok, click on the "read more" link
Him: ok.... (pause) It looks as if they're saying that a site is reporting 503 errors.
Me: YES! Mine! On your Server!
Him: Ok, I'll tell the techs over here...
I've had a nearly identical conversation like this every single day this week with another phone rep... I am so pissed I could scream. Lets hope to God this new host can handle the load my site generates...
Companies, of course, do commit to closed hardware strategies.
Do I have the ability to find out what my hardware schematics are? Do I have access to Intel's microcode and chip design?
No, of course not.
The same thing applies to software. I think the OSS community does not differentiate between open SOURCE and open STANDARDS. Companies should not depend on closed standards. Closed standards are bad, but closed source is not (as long as it is based on open standards). I have pointed this out the ESR many times but he refuses to answer this.
A good argument for open source is security, since you can theoretically look in the source and find trojans, backdoors, etc.
Ok. I'm a little confused. What's so interesting about this comment? Oh. It's an AC comment. Glad to see the sycophantic moderators are busy today.
Hates people who have stupid little sigs
OSS relies on people's good nature to get software written.
I've got news for you, my friend. If you've ever worked in a commercial software house you know that commercial software relies on people's good nature to get written. Why is this? It's human nature. Software development in a team, in a company is all about pissing contests, it's about who's ideas are going to get used, and seen to be used, who's going to get promoted, and thus not have to do real work like coding any more, and so on, and so forth, ad nauseum until you barf. That's the truth, can anybody tell me it isn't? So you see, in commercial development, there always has to be that one guy who doesn't give a sh*t, and just sits down and codes the thing until it's done, while everybody else plays musical chairs trying to be the guy that gets to move out of cubicle-land to the corner office.
In short, commercial software development seems to bring out the worst in people... now, if you have to rely on good nature, would you rather rely on open source, where the natural state of things is for people to be good natured because they're doing something satisfying, or... what?
Life's a bitch but somebody's gotta do it.
There's some (I should know, I spent a year indentured in a parts plant), but it's not like it used to be. I will never get used to "Jeep" being a German brand ... :^/
In the US as well -- MAP started with GM, although all of the "Big 3" eventually backed away from it. I don't know about penetration of fieldbus in US automotive -- although I've worked with it in non-automotive factories, and I think it's a neat technology. :^)
As for your gratitous slam on the fine OSI protocols, (disclaimer: my day job is OSI protocols), somebody ought to write up the victory of TCP/IP over OSI as a victory for Open Source (plus, of course, "rough consensus and running code" over comittees and "profiles"). Even though OSI's CLNP has technical advantages that are only now being brought to the Internet by IPv6, I suspect the OSI suite was doomed from the start by trying to "compete" against the freedom of both the BSD TCP/IP implementation and the RFC documents (and the fact that every OS vendor in the world started bundling TCP/IP). (And yes, I know about ISODE being free as well.)
and so is your mother
The Open Source model works only for apps that are commonly used. Things like operating systems, text editers, compilers, etc. Things that everyone needs on their PC.
I work in the ASIC industry. I use EDA (Electronic Design Automation) software every day from companies like Synopsys and Cadence. The licenses for these tools are very very expensive (I'm talking > $100k in some cases). The reason that these licenses are so expensive is because they are needed for a very narrow application - designing ASICs. Not everyone needs to do that, so not everyone needs a copy of the software. The EDA companies must recoupe their investment in the software develpement, so the prices are very high.
I don't see EDA companies (or any other speacialized company) distributing source. It is just not possible. If they where using a licenseing modle like GPL, then the software would have to be free (as in beer) and they would have no way to recoupe costs (and the support model would not work in this case. I very rarely go to Synopsys for support. I normally use newgroups, in house experts, or web searches before going to them).
The stupidest moderation category is Redundant, IMO. How the hell can anything be redundant in a forum? If someone wants to express their opinion, who cares if 50 people said it better? Maybe the 51st guy will say it the best.
/. is broken. Hopefully it doesnt' degenerate much further.
Moderation of
Hates people who have stupid little sigs
The thrust of the article was really an argument in favour of component based software architectures rather than GPL'd software.
[by "component model", I mean a software architecture which would allow the user to build or buy components which would plug in to an application.]
You don't want MS Bob in your word processor? fine - dont buy the plug in. You want a grammar checker which understands Finnish? Fine buy it and plug it in, or write it yourself.
Binary distributions will protect property while R&D expenditure is recouped.
It is, however, a good point in the article that customers are given no opportunity to fix y2k bugs without buying the latest complete product. This is resolved either by just upgrading the date related plug ins, or (better) if the source is available, simply fixing this.
Both osopinion and upsidetoday are screwed...
sheesh
It's 10 PM. Do you know if you're un-American?
I'm not talking about support! I'm assuming that there are bugs and other problems with some piece of OSS. THEN what do they do? If there's a problem with NT, you go to MS's page, and find the fix. Or, you call them and bitch and they fix it. If Linux has a bug, for example, you have to hope that somebody fixes it. Again, not a way to run a business. Not smart at all. Notice there are no large companies that run a significant amount of OSS? Yes, I've heard the same rumors, but the fact is that OSS is NOT being used my large companies for critical stuff. It never will.
Sure you can sell it, but you have to compete against the guy who downloaded your software for free (or bought the one copy with source), and is giving it away.
How much are you going to charge for it, if someone else is giving it away?
Of course, you could sell services, but why would anyone buy the service from you with only support for one app, when you can buy a package deal from RedHat, LinuxCare or IBM and get support for their whole environment (and in IBM's case, even their hardware).
Those are the issues that seem to be ignored by the OpenSource people (not Free). Where is the viable business model?
Je ne parle pas francais.
Actually, uou can. Just mail Rob with the offending comment ID. Not that I can easily find the comment ID any more. Either way, a URL to the comment should do.
That said, although I'm not happy about it (it lowers my karma :-), it was offtopic. It's interesting to note that the only two posts that were marked offtopic were mine and _Sprocket_'s -- both of us have high enough karma to have a default score of 2 for all our postings. I wonder if it's because our postings at score 2 fell above the moderator's threshold, so were shown in full rather than just as titles? That's why I'm explicitly posting this without the +1 bonus...
"The invisible and the non-existent look very much alike." -- Delos B. McKown
However, I can't agree with his conclusion. He argues that running a proprietary OS is untenable. This isn't really realistic.
I must respectfully disagree. I can confirm Alan Cox's point about proprietary OSes being untanable from a strategic business perspective from personal experience.
The things that made running a proprietary OS untenable for the company I work for (a trading company with millions of dollars at stake every minute of the day) don't just include stability, reliability, and speed, although these ARE very important, and few proprietary OSes have managed to achieve Linux and FreeBSD's levels of performance in any of these three regards. The problem with proprietary OSes is that they can and do become orphaned, often much sooner than the customer can cope with. Either the company goes out of business, drops the line altogether, or forces the customer into a series of upgrades which are not compatible with their business plan or objectives. Examples include the demise of NextOS, the forced migration from Windows 3.1 to Windows 95 (which actually drove some companies out of business), not to mention variouse versions of NT, and Sun's deprication of SunOS in favor of Solaris (which in its early incarnations was vastly inferior in most respects). At best it cost their customers vast amounts of money in lost productivity, time, and licensing fees, at worst it led to the companies actually going out of business.
With Open Source operating systems (and software in general) this is not an issue. We have Linux 1.2.x boxes which have been running for years. If we needed to, we could patch the source code of the os itself and never upgrade to 2.0 or 2.2. When we do upgrade, it will be on our schedule, as we require and see fit. And we always have the option of hiring programmers to maintain the OS and any other open source software we require indefinitely, should the Linux community somehow vanish in the next few years (a meteor impact destroying all life is more likely). The publicly available support insures that, even if all the systems engineers were killed in a car wreck, newly hired people could ask for and get help on the net and elsewhere, even for someting as old as Linux 1.2.13.
However, that mechanism, while typically very responsive, isn't truly "responsible", in that no-one is required to support anything.
The only way someone is "responsible" for supporting you is if you purchase a suport contract, in which case the contract itself clearly spells out legal responsibilities on both sides. Such a contract is just as binding on a new startup as it is an established, entrenched firm. In fact, you'll generally find the new startup working harder to provide good service that the more entrenched entity (I have vast experience with this phenominon and can provide specific examples too numerous for this forum if you wish).
In addition, if you check the licensing terms of any of the proprietary operating systems available today, you will see that they not only do not guaruntee support, they don't guaruntee the thing will even boot or, if it does, that it won't destroy all of your hard work, and your company along with it. Any sense of security management might have by believing their lawsuites will bring them anything is illusionary, and for that precious illusion you have sacrificed many of the real, tangible protections an open source solution would have given you.
But closed-source is, for now, the way some organizations and individuals will go, and I have trouble saying to them, "You're wrong."
I would submit that the problem is with you or the situation you are in, that prevents you from making recommendations which you imply you feel are better, because of political concerns. (This is not intended as a flame. I've been in that situation myself and can relate all too well in how uncomfortable it is.) Yes, by recommending open source you are sticking your neck out and taking responsibility that the system will work. This isn't as comfortable as being able to say "yeah, another Microsoft product that's flakey, it is Bill Gates fault." But don't forget, there was also once a time when "no one got fired for buying IBM," but after that time had passed, ironically, many people did in fact get fired specifically for buying IBM, at inflated prices, while their competitors bought competing products with more capabilities at a fraction of the cost. Those on the tail end of that curve paid dearly, with their jobs. I would submit the same will be true of those on the tail end of the "no one got fired for buying [insert proprietry software/os of your choice here]," and that the tail end of that particular curve is very close indeed.
Organizations that lock themselves into inferior solutions because managers are more interested in protecting their own assess through a readilly available chain of blame, at the expense of the company's ability to compete and protect itself against the vagaries of a volitile and fickle marketplace, will in the end find themselves at a terrible disadvantage in their respective industries when going up against firms that do have more solid technical foundations. If nothing else, it is that very disadvantage which will force change, by either requiring the entities in question to adapt, or to cease to exist. This has already worked very much to our advantage in our particular industry, much to my personal delight (and benefit).
The Future of Human Evolution: Autonomy
Intel is closed hardware. I cannot see either their schematics OR their microcode.
Perhaps you are confusing open STANDARDS with open SOURCE.
he's very right about the advantage of open source being the right to modify. but his opening examples don't go along with that argument.
No company now would commit to a closed hardware strategy. It would cost them more than using commodity components
first of all, what does he mean by "closed hardware" strategy? sun sparcs? ibm mainframes? i know there are plenty of companies that have committed themselves to both. and there still will be for some time. besides, who really has the "right to modify" when it comes to hardware? i don't. at least not with any hardware i know of. which actually means there are no "open hardware" strategies. at least not in the sense that he talks about with software. with hardware, the only choice you have is to go with something else if your hardware vendor isn't meeting your demands.
well, surprise! you also have that option with software. if sybase isn't living up to your expectations, you can switch to oracle. or vice versa. if you have a beef with solaris for your oracle servers, use irix. or nt. or linux. or.... oracle will run on any of these.
while the right to modify is a big advantage for open ource, it is not the only criteria for choosing software. oracle has a huge market right now for databases, and no one has the right to modify it. and as long as they continue to offer other advantages that outweigh the right to modifiy, they will continue to have a huge market share. if people decide that mySQL does everything that they need, and the right to modify is important to them, they will use it. but there are several commercial databases that have huge advantages in both performance and power over mySQL that, for most companies, greatly outweigh the disadvantage of not being able to modify them to fix their problems.
If I don't put anything here, will anyone recognize me anymore?
Has Solaris ever made money? No. But Sun's hardware makes money hand over fist. IBM makes money from hardware and support. Why bother writing a web server when they can use Apache for free?
The big companies (IBM and Sun) are starting to catch on, e.g. IBM's adoption of Apache. By adopting open source they gain several benefits:
- The adoration of
/. - Cheaper development (free volunteers!)
- No revenue loss -- IBM's customers will buy WebSphere whether parts are open sourced or not.
- The crushing of small software competitors. Microsoft, Sun and IBM have powerful brands and huge advertising. They can embrace open source and still make pots of money. Small software companies can't.
Microsoft, Sun, IBM have a lot to gain by open source. It's the smaller software companies that have a lot to lose.Sun's internal disks indeed only work with Sun SCSI cards. However, Sun SCSI cards will work with other disks, and Sun's external disks work with other SCSI cards. Sorry if I wasn't clear.
Why don't you throw SCSI cards into these Proliants and hook then up to an external highly available RAID box (EMC, IBM, and others)
Use the Compaq stuff only for the OS and swap.
The article presents a well thought-out view of the necessary commoditization of the underlying technology infrastructure, both in terms of hardware and software. An interesting analogy is that of the actual manufacture of electronic systems. Historically electronics OEMs (original equipment manufacturers) manufactured the entire system, only purchasing the components required to assemble a complete system. As a result most systems were proprietary in nature until their component suppliers developed standard commodity components that they could leverage in selling across a wide variety of customers. But even then the OEM used proprietary design and manufacturing methodologies. Over the last several years the electronics manufacturing paradigm has taken a radical shift whereby the OEMs are increasingly outsourcing the manufacture of their products to electronic contract manufacturers. This is commonly referred to as the Dell Model. (Do you realize that Dell manufactures practically none of its products, other than final box-build assembly, test and ship? And that for every dollar of net fixed assets Dell produces $35 of revenue versus Hewlett Packard which generates about $9.65 of sales for every dollar of net fixed assets.) The difference is that Dell outsources nearly all of its manufacturing to the likes of Solectron, Jabil, etc. Now electronic contract manufacturers serve a very wide base of OEMs and as a result their component content and the manufacturing processes employed need to be somewhat uniform in order to fit a myriad of projects into the contract manufacturers manufacturing model. The manufacturing process and all the requisite design parameters have been commoditized. The point is that this new manufacturing model allows the OEM to focus on what it does best; R&D, design, marketing, and customer service. Those OEMs that continue to adhere to the proprietary design and manufacturing model run the very real risk of being unable to effectively utilize their own scarce capital resources in the most effective and time sensitive manner, they are saddled with an overhead structure that doesn't allow them to do what should be their core competencies. In the end the electronic contract manufacturer must differentiate itself on service/quality and price and as a result they are forcing their component suppliers to deliver more and more standard components (proprietary components can cause a major hiccup in these large and complex manufacturing processes.) If an OEM finds that its electronic contract manufacturing solution is not meeting its needs in terms of service and price, its switching costs are not nearly as high as if it employed a proprietary model. Remember, electronic contract manufacturers provide essentially the same solution, they can only differentiate themselves in service/quality and price. It is only natural, and now very main-stream, that the same type of thinking be applied to software. Organizations that adhere to proprietary software models cannot effectively use their capital resources to do what they should be doing best. This is why I see the open source model as analgous to the electronic contract manufacturer: 1) It provides a highly specialized service that produces a high quality solution with an open and commodity-like nature that can be leveraged across a very wide customer/user base. 2) It allows customers/users to more wisely use their own scarce capital resources do what they should be doing best and in a much quicker "cycle-time". 3) It tends to minimize the risk associated with the proprietary model in that customers/users are not dependent upon any one flavor of solution, switching costs are minimized. You need only point your senior managers to what has happened in the electronic contract manufacturing sector when arguing for open source solutions. They'll understand the analogy as this sector is now very sexy on Wall-Street and in the board-rooms of corporations seeking to better use their own resources. By the way all of this outsourcing was pioneered by the auto makers, you'd be surprised how little say a Chrysler actually manufactures (don't confuse manufacturing with final assembly.)
I have great faith in fools. Self confidence, my friends call it.
Actually, in the automobile industry's case, it's an oligopoly with some limited price fixing as per the current years price leader. They may innovate some, but their goal is just to keep a stable market share. Ford would get completely burned if they tried that one year. There are other elements to that industry (yay labour unions), but that basically summarizes it.
Therefore, it is a pretty good metaphorical comparison. In the future, as more companies move to the open source paradigm, the more the ones that don't will feel its pull and the consequences of not following it. The advantages are just too good to completely miss out on. However, this market is *HIGHLY* competitive -- unlike the auto industry -- and anything can happen (most likely a large number of mid-sized competitors; unless someone comes up with a better business plan to direct customers to themselves instead of competitors).
As for COTS, I'm really not sure. Companies like products guaranteed to work -- and support to match. They also like solutions. Open Source alternatives can potentially and essentially provide the same thing; though, I don't think they will overtake closed solutions any time soon. Their business plan isn't in a standstill either -- it's just a matter of momentum which no one can predict though..
----------
yup. Once linux moves up to enterprise scalability, reliability and performance at a level comparable to sun at prices a hundred times cheaper; there goes their market share.
IBM is trying to do the same thing with their "solutions" and linux as well though. Competition and more market information will eventually soften those prices up however.
----------
The Microsoft's, the IBMs, the Novells of this world have a lot to lose, in terms of customers who are no longer tied to their existing products
The number one reason for a large company like this to go Open Source is customer demand (I'm guessing IBM took on Apache because they realised that Apache was what the customer wanted).
Moral? -- keep evangelising. We have advertising budgets up against us.
--
The best thing about this article is that first posters can mention Beowulf clusters and actually be on topic. Thanks, Alan!
It's an interesting sound bite piece, but I think he fails to address the main area of resistance to Open Source, the loss of intellectual property. If a software company has a long R&D before receiving any income on a product, then they immediately release that product as Open Source, competitors can use the source code at no charge.
Of course, the competitors could not possibly understand the product as well as the inventors, so no PHB in his right mind would choose a cheaper solution over a better solution, right?
A middle-ground solution could be providing the source code to the company with the provision that it can't be disclosed to a competitor. That is, use a very limited source code liscense rather than Open Source.
This doesn't apply to the use of the Linux operating system in the workplace, which is, of course, a win-win solution. The initial investment has been recouped by the inventor(s) and only goodness can follow in the wake of this ship of code. (insert birds chirping happily)
Posted by Mike@ABC:
I wish I could read this piece, because I can't connect to it right now. Still, I'd like to add to the musings posted above. I wonder that if we start treating operating systems more like public domain electrical power, will the open source OS still be able to grow? After all, electrical service in many areas started to suck after it was made public domain and became regulated as such.
And secondly, as I anticipate the response to that last question, what happens if this current generation of right-minded hackers falls by the wayside and is replaced by people with less imagination, drive and fairness?
I probably know the answers to those, but figured why not ask anyway? Discussion is healthy, after all.
This is a really good article - how come, whenever someone comes up with a view like this you look at it and say: "That was really obvious, why didn't I think of that?!" I guess that's what separates us into those who write this stuff, and those (like me) who merely read.
We see alot of these kind of articles, and I thnk that those would be good to quote (or at least draw from) for those internal justification type documents one is called upon to write from time to time. Could we get together a repository of this type of article as a reference for such situations? or is there one already? Naturally there would be copyright issues to consider, but I can't see anyone having major objections [not that I've asked]. Perhaps that's something for my 15Mb of free webspace...
etc
I've put a copy (not as nicely formatted tho) on
http://www.linux.org.uk/FEATURE/risk.html to help
spread the load a bit.
Alan
I'm writing because, while Alan's analogy above makes sense (by comparing close hardware to closed software), it's inherently flawed.
Companies DO commit to closed hardware solutions. And for the same reason they commit to closed hardware solutions, I believe they'll continue (in some cases) to commit to closed software solutions. It's sad, but true. On the other hand, it's not a lost battle, and it's not an omnipotent problem. It doesn't have to be this way everywhere....but it will continue to be this way in some places.
My example: Just a few weeks ago, the call went out that we needed some new workstations. Not TONS of them, just 2. The matter was discussed and I said almost exactly what Alan said: "Buy commodity parts for the computers, and they'll last you longer. Total cost of ownership will be less. We'll all be happy." They all looked at me and said "You're right, but we want cheaper machines and we can get these barebones systems for $300." SO in the end, I lost, and we bought systems where everything was hardwired to the motherboard...which is pretty damn closed.
This attitude, namely the "It's cheaper, and we don't have to worry about it once we've bought it" really attracts companies to closed solutions. They're wrong and they will have to worry about it once they've bought it....but they still, in many places, equate "closed" to mean "finished" and/or "solid". Yuck.
Werd.
This of course differs from automobiles. The blueprints for a Dodge Intrepid are not made public, and even if they were, you couldn't necessarily build one. So while I think that the analogy is good one, the real key of Open Source is something that is not stressed in this article... and that is, of course, value-addition.
By now, we're all familiar with the concept. You give a product away, but you support it for a fee. Or you let everybody download an ISO of your CD and give it to whomever they'd like, but printed manuals cost extra. I think this is the only model that really makes sense when you consider the unique aspect of digital commodities: they're non-physical and infinitely replicatable.
A thing which can be duplicated ten billion times at no cost has very little intrinsic value. It is no longer scarce in any sense of the word... and as even the most uneducated high school economics student will tell you, when you remove scarcity from the economics equation, everything else falls apart. Value is derived from scarcity... at least in traditional economics. So what do companies wishing to offer non-scarce products and also make money do? Tie in some scarce things! Support is scarce. Printed manuals are scarce. So people pay for them.
I wonder how long this sham is going to hold up? :) As we move increasingly toward an economic system based heavily on products that are non-scarce, I think traditional capitalism is going to collapse more and more. And it's not just software, either. Music can be made non-scarce. Art can be made non-scarce. Anything which can be digitalised can be made non-scarce. When we look at this more, the key emerges: just because it is non-scarce does not mean it isn't unique. The best song in the world duplicated 10,000 times is still the best song.
What does this mean? It's hard to say. Perhaps we're approaching a day where creativity and the actual utility of a product will be valued paramount, rather than its price or availability. I can't fathom an economic system based on non-scarcity, but then I'm hardly an expert. I'm sure we'll think of something. :)
When I looked at the post I'm responding to, it had been rated, ``Offtopic", & probably downrated because instead of discussing the subject of Alan Cox's article, it comments on the server software at OSOpinion. And does so thoughtfully.
/. has gotten, there's been actually very few trolls or flamers here lately.
I sense an abuse in moderation here. And I'm a little ticked that I can't find an easy way to tattle-tale about it to the guys running Slashdot.
And also the fact that the other day while meta-moderating I voted five ``troll" or ``offtopic" posts as being unfair ratings. Somebody was having a bad day & took it out with his 5 moderation points.
Look, people, all discussions wander a little from the main topic, & explore the subject or its context a little more fully. Sometimes the interest in a given item posted here is the context. And that context includes the software of the site.
In other words, lighten up on the moderation. Despite the reputation
Geoff
I think I see a trend here. Maybe for them it really would be easier to muzzle the entire internet than to produce p
I think the post I'm replying to mentions something very important: companies do commit to closed hardware solutions.
The reason is simple: its not all black and white. Essentially this whole discussion always circles around two terms (closed & open). Generally the shades of gray are ignored (i.e. partly closed/open).
With any product (including software) there are generally two parties: the consumer and the producer. Those parties have different interests. The consumer wants bang for buck and as much control as he can get. The producer has a different interest: maximum profit, prevent that competition is succesful, keep customers happy.
If we look at software we can identify different levels of opennes:
I - not open: you receive a blackbox piece of software. If you are lucky a manual is provided. Basically the license only allows you (and you alone) to use the software. This level of openness maximizes the benefits for the producing party.
II - in the next level of openness the software is still handed over as a blackbox. However, it's external behavior is fully described in API documentation. Because of this the blackbox can be used by third parties but they still have to buy it and the producing company fully controls what happens to it.
III - in the next level of openness the source code is provided along with the product. The right to edit the code still lies at the producing company, however. This gives some flexibility to the user since it is now possible to see what the software does and how to use it. Also bugfixes can be suggested to the producing party.
IV - the user now also has the right to edit the software code. The producing party still owns the software though and can restrict the right of the user to distribute the changes.
V - The producing party no longer owns the software. The user can edit the software any way they think is useful. They cannot restrict the license on the existing software but can license the changes in any way they think is appropriate
VI - nobody owns the software and changes have to be put under the same license.
Maybe more levels of openness can be identified, maybe they should be described in more detail but I think this covers most common situations. I think that for each level of openness you can find examples where it is the appropriate level of openness and examples where it is the worst possible level of openness.
I liked the car analogy in the article, I liked the production company analogy too but I don't think these analogies always apply.
Software can come in many forms ranging from critical (an OS, a database) to non critical (games, desktop image switchers); from valuable to non valuable; from comples to simple; etc.
Also it differs from user to user in which categories a certain piece of software falls. So any piece of software can be placed in a multidimensional space of quality attributes. And on top of that the placement is subjective and subject to change in time.
Generally the level of openness of software changes over time. When it is first released it is valuable, perhaps very critical and only available from one company. The company will try to maximize profit and go for a closed license to prevent the competition from catching up. After a few years similar software is produced by other companies so smart companies make their product more open to attract more customers. Over time the software becomes more open until the point it is no longer profitable to produce it. At this stage it becomes fully open.
As I see it operating systems and client side software is moving towards more openness on this scale. If you have a PC you can choose between a dozen or so different operating systems. Some of those are fully open (linux), others are nearly open (BSD), some are semi open (solaris) and some are fully closed (windows).
The closed operating system are losing their advantages to the open operating systems. I think the point in time where closed operating systems have no advantages over open operating systems is nearing. Depending on the type of user this may already be the case.
In my opinion there is not much profit to be made in low level system software so you might as well make it open. Device drivers and the lower layers of the operating system fall in this category. The situation is different for the higher layers of the operating system (configuration and management tools)
So you can either make the OS open (linux) and concentrate on selling what is running on top of the OS or you can bundle the OS with a lot of software and sell the whole bunch in a single package (windows).
In the first approach the producing company can concentrate on the portion of the software that is actually bringing in the money. In the last situation the company has to concentrate on the whole bunch of software, including the part that is no longer bringing in any money.
If we look at NT we see that a lot of software is provide with it. It is this software that adds the value to the whole package not the OS kernel.
If we look at development environments we see that they come in different editions these days: a cheap edition with only the very basic features and an enterprise edition which is the same product with some added value in the form of class libraries, extra software, manuals etc, support, etc.
From both examples we can see a similar pattern: the common, often central part of the software doesn't give you a competetive edge since it is available from many sources. You might as well make it open since this gives the rest of your software a competetive edge.
I think software production is the same as any other production. To make it profitable you take cheap ingredients, you add value to it and you sell it for a higher price then you spend on the ingredients. It doesn't make sense to give it all away. You only give away the stuff that doesn't provide you with (enough) profit.
Linux is something you can't make much money on. The same applies to much of the software that runs on top of it. The software industry is realizing this (except for a certain company in Redmond) and is shifting their attention to adding value. Companies like SUN and IBM have realized that they won't get rich of selling software so they are shifting their attention to what they are good at. For SUN this is selling hardware, for IBM this is both selling hardware and providing support. Neither company would benefit much from completely open software so they generally don't put GNU licenses on their software.
Let me explain why I wrote this lengthy posting. I frequently read threads about OSS on slashdot. I see lots of people bashing SUN for not giving away all the software they have. I see lots of idealistic and religious crap.
I think those people generally miss the point about open sourcing your software. The article featured in this thread maybe us an eye opener to these people. Hopefully my posting contributes to it in a modest way.
Jilles
You are wrong. "doc" is not a secret - you can get the specs from MS site.
Suppose I or another consultant find a bug in Excel that prevents our customers from doing their normal work. Suppose we contact Microsoft about this bug, and the engineer there tells us that we will have to upgrade to Office 2000 in six months to fix the bug. (True story from February 1999). It's a really dumb bug, too.
Suppose Microsoft generously provides us a beta-test CD and allows us to find other bugs in Office 2000. Suppose we do. (True... ah, you get the picture.)
Granted, we benefit from the bug fixes, but so does everyone else, including our customer's competitors, if they use Office 2000.
I suppose you could claim that Microsoft didn't have to give us a beta of Office 2000, which is true. Still, how much money are we (actually, our customers) paying to Microsoft for the privilege of talking to a second-tier phone monkey? Much more than it would cost for me to e-mail someone at Abiword or KOffice or even StarOffice, that's for sure. If it were up to me, I wouldn't want to pay that much. Different business sense, I guess.
--
QDMerge 0.4 just released!
how to invest, a novice's guide
It also gives them someone they can sue.
One might think that companies are in the business of [whatever the company ostensibly does]. In fact, as Otto said, companies are in the business of making money. If a problem with a supplier's product interferes with their ability to make money doing whatever it is they do, they gladly will substitute compensation obtained from the supplier.
(Although, as Otto noted, they'd rather not have to wait.)
This makes the assumption that you're ever going to make any money off of those apps. In an enviroment where many classes of software are seen to be free already, this is a pretty dubious position.
"Sun internal SCSI disks only work with Sun SCSI cards" That is *not* true. I'm using a seagate drive on my sparc2 right now, it works fine with the sbus scsi card. They make you buy their own brands because its easier for their techs to weed out problems. If you use generic cheap ram in a critical server then call Sun for tech support, who is to blame? You for using cheap parts, not thoroughly tested with their hardware.
I missed that emoticon too. Glad to see no one bothering AC. The man has done so much to increase my computing joy, I don't think I could bear to see him get abused.
Ok, time to stop protecting AC and get back to work.
You are right, this statement is so unbelievably wrong. But companies don't commit to closed hardware simply because of a desire to have now Now NOW! And it has nothing to do with the "right" way to do things.
I make carrier class telco switches. Sure there are some ISO standards and some Bellcore standards and some ATM Forum and Frame Relay Forum standards. But those don't cover everything in our system. We have a lot of closed hardware in our system. We use closed hardware because:
We have closed hardware from several vendors and in every instance it is because of performance considerations or because THERE ARE NO COMMODITY COMPONENTS for us to use. In no case has it ever had anything to do with cost or Pointed Headed Bossisms.
If you wait until there are commodity components available for everything to build your systems you aren't going to have much in the way of high performance, cutting edge systems. Maybe home users have that luxury but Level Three and Qwest don't.
That single OS provider is the only one offering a solution in many areas. Try doing video on Linux. It's trivial for a large company like Microsoft to make it work. Take away large corporate funding and no-one has the money to get that basic feature to actually work. The only reply you ever get from someone who actually knows how to fix it is "See what you can code yourself. I'm out of time for this project."
For instance, you say, "Gee, I'd really like a GPL'd foobar client with a built-in spam filter." You put up an RFP, I see it and say, "I'll write one when there's a commitment of $1000". You guarantee $200. Your competitor sees the RFP, thinks, "Hey, I could use a foobar client like that!" and puts up $150. Someone else pledges $20, and so on until there's a total commitment of $1000. Then I write the program, it goes through the approval mechanism, everybody pays up their pledges, I get paid, the code gets released, and everyone's happy.
There are other projects where it's difficult for your competition to benefit, at least immeditately, from the new code. If I do an open source port of something to the AIX architecture for IBM, Sun won't benefit.
Tom Swiss | the infamous tms | my blog
You cannot wash away blood with blood
You mean, like the recent buyout of Daimler-Benz by Chrysler? Oops, "strike that -- reverse it." :^/
Of course, I don't know if there's a UK automaker that hasn't been bought out by an American or Continental company yet ...
You can get the specs, but the last time I looked you have to agree that you won't use them to build a product that does the same functions -- that is, you can't use the specs for .doc to build a word processor. I still claim it's not an open standard -- MS can change it arbitrarily, without notice, and has done so repeatedly in the past.
I think having a single (if forked) open-source implementation is inferior to having multiple implementations of a frozen, agreed upon, standard API.
For instance, I would much rather have multiple HTML browsers rather than a single FooBarML browser, even if it is open-source and "forked"
The whole Linux desktop situation would be better if the GNOME/KDE guys hashed out some standards, agreed to them, and both implemented them. XDnd is one, but almost everything else is divergent.
Protocols are more important than source. Having source to all the implementations is nice too, but if they are divergent in protocol, they add to the cost of software development.
Component based programming and standardized APIs/protocols solves most of the danger of committing to a single vendor.
For instance, for databases, we have SQL/ODBC/JDBC. For distributed programming we have CORBA/EJB. For file formats, we have XML, etc.
I'd much rather target SQL/ODBC than target "MySQL" only, even if there were 50 open-source forked implementations.
The biggest problem with OSS is that there are always people willing to fix your software for you or upgrade it for free. That's a hell of a big assumption.
So fix it yourself, or hire someone to do so. Software vendors generally do end up charging you when they provide a bug fix, since it tends to come in the next upgrade. Moreover, you can get those bug fixes done in a hurry if you need a fix in a hurry, rather than waiting for the next bug fix release.
No, the biggest blindspot of OSSers isn't this. It's missing the fact that for many purposes, effective open source isn't a practical alternative due to its non-existence. 8 years ago, what open source OS would you have used? Linux? Linux was still Linus's pet project then. Even today, with the growth of Linux, there's plenty of Linux users who still use Windows a lot. Why? Because of the apps, ones which have a better interface or simply don't exist under Linux. If I want a really good flight simulator, FlightGear just doesn't compute with MS Flight Simulator 2000 yet. There's no open source project comparable to Age of Empires/Kings, Starcraft, Quake, Quicken, Diablo, etc. Maybe this will change, maybe the flood of work being done on open source code and its inherent advantages (open source code almost never gets orphaned or neglected) will be enough to fill in the large gaps. But I wouldn't expect to see Bill Gates redeeming food stamps any time soon...
Ooh, a sarcasm detector. Oh, that's a real useful invention.
I put a copy of the text of the article here
http://ubasics.com/adam/commodity.shtml
Enjoy!
-Adam
Generally, I agree with Alan Cox. The advantages of proprietary, closed-source software (e.g., having a corporate entity to call or sue) are dubious at best, and the advantages of open-source software (e.g. community and multi-vendor support) are typically not understood by corporations.
However, I can't agree with his conclusion. He argues that running a proprietary OS is untenable. This isn't really realistic. At my last company, we had problems when we had to bounce important machines because they had been up so long (~2 years), that no-one had been around or could remember what to do when they came down (this was a problem of poor planning). The point is that, despite the fact that we were running a proprietary, closed-source OS (Solaris) on proprietary hardware (from the same company as the OS, no less!) we had as much stability as anyone has the right to expect.
The advantages of open-source are both obvious and important. There is a great support mechanism (the community itself). However, that mechanism, while typically very responsive, isn't truly "responsible", in that no-one is required to support anything. The formalized, "responsible" support mechanism for open-source software, of which Alan writes, is new, and I have yet to see a comprehensive review of the OSS support companies. When these companies reach a level of maturity (both experience + time) comparable to that of very good proprietary software company, they will be able to add much more to the community than simply saying, "Hey, we exist."
The best thing we can do is, of course, OSS advocacy. I selected my most recent job in part because it leans to OSS instead of the proprietary. But closed-source is, for now, the way some organizations and individuals will go, and I have trouble saying to them, "You're wrong."
I know many will disagree with me, so if you must, flame on.
-Josh
I don't mean to sound flamey, but I don't think that's what he really meant. Sure, it's inconvenient for upgrades, repair, etc. but I think he meant more architecture wise. Say Mac vs. PC. With a PC you have hundreds of choices for motherboards, video cards, hard drives, etc. But with a Mac, you get what they give you. Sure, you can swap some things occasionally. Video cards and such. But you might accidently void your warranty by doing that. Even some PC manufacturers, such as HP and Packard Bell make proprietary parts, so that if you spill mountain dew on your keyboard, you can't just run to the store and buy a cheap one, you have to shell out $60 to the manufacturer, so that you can get a keyboard that is compatable with their motherboard. Many people don't like that sort of thing, so they would avoid it if they can.
I put a lean copy of the text here
http://ubasics.com/adam/commodity.shtml
Enjoy!
-Adam
This is a trend I like. When I think closed source, I think MS - I actually have run into a lot of closed source stuff that works great, but like I said, when I think closed source, I think MS. But most software manufacturers still behave the in the same elitist manner: "we know what you want better than you do - and if something doesn't meet your needs, tough. Doesn't work, tough." Nobody would buy a toaster with a 3 page EULA
"This toaster could cause your house to burn down unexpectedly, but you agree that you use this toaster at your own risk. GE shall not be held accountable for damages or loss deriving from the use or misuse of this product, even if this product is being used according to established guidelines. You agree to indemnify and hold harmless GE and all its subsidiaries, and you agree not to file suit against GE for damages caused by the use of this product. You agree that if any problems arise in this toaster, you will not unscrew the casing and attempt to fix it. Failure to abide by any of these guidelines will result in the revocation of your license to use this toaster."
Hmm, actually getting to control my entire computer environment like I want?... Yeah, I like that idea. (sorry, had to rant - its sad what NT does to you)
I liked the article. It uses (and acknowledges) lots of the ideas expressed by Bob Young.
I've used the analogy of the automobile with Open Source as well. But I wonder, with today's new techno cars, are we going back to "closed sourced" vendors? If, let's say, Ford comes with some new high tech fuel injection, or even a new engine design all together, and patents it. Will the normal mechanic be able to fix it? Dealers make a lot of money off of maintenance, does Ford get some of that too? If not, then Ford doesn't have a problem with sharing the "innards" of the design. Just licensing it. Food for thought.
Second, I like to mention another aspect of Open Source. Several years ago, a buzz word called "COTS" (Commercial Off The Shelf) was used by management. I was one to suffer under this. I had to make closed sourced tools do things that they were not designed to because it was a COTS product, and was considered "cheaper". Well, in the long run it was not. I wrote about the same amount of "glue" code to get it working that the time and effort was almost the same as doing the desired tool from scratch. But now we have licensing fees and we don't have a product nearly as good.
I've shared a term before called Modifiable COTS. This is what Open Source is. The whole COTS idea was not to reinvent the wheel. To let someone else do it and use a commodity tool that would be less expensive than a customized one. But this failed because commodity tools where always general. Good for the basic purpose, but not good for all specific needs. But with Open Sourced tools, you have a tool that is "off the shelf" and good for the basic purpose. But where Open Source leaves COTS is that you can tweak it to your needs. Thus the term "Modifiable COTS".
Third point.
A competitor of ours has a board that runs Linux. We use another "closed" source OS for our board. The few of use that realize the benefits of Linux has brought this to the attention of our Upper management. But they still don't understand the technical details. But we were able to (after lots of complaints to the management) get a unfunded OK to do a work study on porting to Linux. If this does well, then it will be a BIG boost for the Linux community. This article may have helped in the effort! Thanks Alan.
Steven Rostedt
Steven Rostedt
-- Nevermind
Alan Cox -- and most of the open source software movement -- doesn't seem to understand that the vast majority of hardware aren't "commodities". How many companies make processors that understand the proprietary x86 instruction set? What about 3dnow or MMX? There are, what, three or four companies making such processors? That's an oligopoly. Processors are cheap and widespread but they are not commodities. Every business on the planet uses non-commodity processors. Even Linux is based around non-commodity, proprietary, closed hardware. They are built around PowerPC, Sparc, Alpha, MIPS, x86 -- every single one a closed non-commodity.
I'm a hardware engineer. I work in the
;-)
ASIC world. I spend a great amount of my
time building CUSTOM hardware for different
companies. There is a HUGE amount of private
IP that the world has no access to beyond buying
the little black piece of plastic with legs on
it and using it in a larger board design.
Alan is correct when we talk about "commodity
computer systems" but incorrect when we talk
at the component level of those systems. For
that matter - do you imagine Sony gives you the
schematics for their walkman? It plays a
tape recorded in a standard manner, or a CD
recorded in a standard manner - but the
hardware implementation to accomplish that
playback is highly proprietary.
(Okay - so Alan and I can get schematics to
our ham rigs..but that is about it..
Thus - the closed nature of hardware depends on
where you draw the viewing line. There is LOTS
of IP under the hood that isn't open.
Have you compiled your kernel today??
Yes but let's not forget the immense pleasure we've all experienced from using Microsoft and other wonderful commercial products. We all know that in the end we'll have had our fun and come back to safe reality of a Microsoft/Sun/blah world. Won't we?
when it comes to support
The only group that gets hurt are the companies that only know how to sell products based on Intellectual Property or trade secrets.
99 little bugs in the code, 99 bugs in the code,
fix one bug, compile it again...
I'm a leaf on the wind. Watch how I soar.
To paraphrase someone's tagline here, "The difference between theory and practice is that in theory, they are the same, in practice, they are very different."
-=-=-=-=-
-=-=-=-=-
My mom's going to kick you in the face!
The biggest problem with OSS is that there are always people willing to fix your software for you or upgrade it for free. That's a hell of a big assumption. Right now, the economy is good, so there are lots of college kids and uber-geeks with enough free time to write code for free, great. What if the economy goes in the toilet? What if OSS isn't "cool" anymore (I don't happen to think that it is)? OSS relies on people's good nature to get software written. That's just plain crazy! No company, with any kind of clue, would base something as important as their computers' OS on people's good nature. Anybody who does, deserves to, and more than likely will get hit in the face for it. You simply don't get shit for free. You're kidding yourself if you think that you do.
Sigh. When will Slashdot newbies learn to write HTML for clickable links? ;-)
-------
Warning: Slashdot may contain traces of nuts.
No company now would commit to a closed hardware strategy. It would cost them more than using commodity components. Just as importantly, it would commit them to a single source for support and parts.
As much as I'd like to believe that people don't buy closed hardware any more, I've actually seen a resurgence in closed hardware lately. It has been pushed by none other than our old friend, Sun Microsystems.
Examples:
I'm sure there are other examples, but the point is the same-- they want to lock you into a single hardware platform. This leaves very little leverage for negotiation when they drop the ball.
I totally agree that open hardware with well-defined, standard interfaces is the way to go from a customer's perspective. One of the reasons I have such high hopes for Linux is that the feature set is virtually the same across a variety of otherwise completely different hardware platforms.
What other examples of common closed hardware can you folks think of?
Just yesterday I experienced a good examples of the kind of craziness Alan Cox is describing. A colleague asked me a question about LaTeX, which he said he was using on one of our Suns. My boss, listening in, said he should quit using LaTeX and start using a "real" document format. "In this country," he said (we're in Germany), "there is a de facto document format, and it's called .doc!"
Now my boss is not your everyday PHB -- he's pretty clueful on technical matters, in fact, and didn't mean that this is a good situation, but just a fact of life, albeit an unfortunate one. I can see what he means, but nevertheless I've been muttering to myself with anger ever since that conversation. This so-called document standard (he's talking about MS Word, in case you didn't get it), is a secret! And not only that, the secret gets changed, secretly, at least once a year, so you have to shell out hundreds year after year, just to keep using the "standard"!
I think this exemplifies just what Alan is saying. It's simply crazy that businesses put up with such a situation. In any traditional manufacturing business, no vendor could ever get away with it, and yet millions of businesses tolerate this kind of treatment with their office software, paying billions for nothing and all the while insisting that they're being economically rational.
Always keep a sapphire in your mind