Slashdot Mirror
Motley Fool on Microsoft vs. Linux
Simon Janes writes "In a two part article series, Rob Landly (aka TMF_Oak) discusses Network Effects and how they make a product more valuable to users. Rob continues his series the next day with a discussion of Microsoft vs. Linux. This is excellent 'outside of the industry' press for Linux because 'Fools' are people like nurses, teachers, accountatnts and retirees who are not normally exposed to this."
- stupid
- commercial
- server
- business server
- goes wrong
I could ask similar questions about the part I didn't quote, too, as in "win", "apps", "win apps", and "non-commericial use". (Note that "win" as you used it is a nice example of a single-word oxymoron.)The fear of taking responsibility for one's own choices, and for solving one's own problems, has reached epidemic status in our litigious society. It's a wonder anything ever gets done at all with so many cowards and crybabies running to their licences and lawyers every day.
Then you go talking about this figmentational "desktop" bogosity again. I won't waste time talking about it again. I've done that a bunch lately. Check other threads.
Virtually no one writes anything for the Linux kernel. And very few people write anything for Linux-based operating systems. You're just seeing people work on Unix stuff, and you see it running on some vendor-supplied, Linux-based operating system and figure it's something specifically for the Linux OSes. In 99.98% of the cases, you're wrong in this.
And as for business computing... huh? What does that have to do with anything? Now you're coming very close to talking about the marketing world of lies and avarice. Geeks have no trek with that crap without losing their souls.
But quibbling aside, you're right about a good bit of this. For example, the need to have highly competent people managing the computers and the computing environments of the highly incompetent ones. You can never make computing easy enough for the idiots. The stupid shall be with us always. So you have to establish set-ups so that they have professional caretakers. System adminstration is as important as it ever was, if not more so. Show me a system that even an idiot can adminstrate, and I'll show you one that only idiots would ever use.
System adminstrators shouldn't have to be in the same hemisphere as their user unless new hardware requires installation. Anything else sould be location independent. Also, the number of sysadmins should be related more to the number of users than the number of machines. If you find that merely adding another host to your net incurs significantly more admin overhead, something's wrong. A professional sysadmin automates everything, so one more machine shouldn't matter very much. One more user, however, does, because it's the routine human interaction that takes all the non-exceptional-event time.
You know we keep hearing about how such and such a side effect of open sourcing something makes it better: well you can put some energy into answering phones instead of coding all day and increase its value. You can conceptually increase its value by expanding its user base. You can sell t-shirts with your product's name on it and increase its value. It seems as people get more experienced with software they find engineering to amount to less and less of a product's value. When do we finally tell engineers to shove it and focus on marketing instead?
> 2. Most people who are not STUPID want their business servers to run an OS that is COMMERICAL so someone is liable if something ever goes wrong!
So, you think you can actually sue Microsoft, Sun, or ANY other software vendor when their software crashes and screws up your business? Have you ever bothered to read the licences for this software? They specifically state the manufacturer is NOT liable for any losses that might occur when the software fails. Have you ever heard of such a case? They simply do not happen, because of the licences. If I had someone working under me who proposed using any software because we could sue if something went wrong, I would consider firing them, because they would clearly be ignorant and incapable.
Eric Geyer
corduroy@sfo.com
Microsoft has done so much for us, like create the internet, bring down the price of PCs and software, make an easy to use operating system. Without microsoft we would still be using DOS applications.
They have done so much for the world, brought PCs to a vast majority of people. Made it easy to connect to internet, all I have to do is load internet explorer, and internet is there.
Also, they have made programming easier, with Visual Basic. Even I can program applications, and I'm not particularly good with computers. Alot of applications look the same, so I don't have to go learning new interfaces all the time.
Many people complain about MS bad tackits, but I think what they have given us, the world, is easier use of computing. I think, other companies could not do this, because each would be inventing there own standards - there would be no one common interface, common look and feel of applications.
We should be thanking microsoft, thank you very much Bill Gates, your wonderful!
It's nice to see sites such as fool.com actually recgonizing that Linux is really attacking Microsoft. Not to mention the entire open source model as a viable option to the current software development model. I love hearing things such as "212% annual growth" and "will eventually surpass windows about three years". There are only two places I actually deal with Windows one is a terminal at work (the one I write this comment from) and the other an NT server at another ISP I help maintain -- just fyi Linux has all but replaced NT since I got there. Microsofts garbage about TCO and crap is summarized by the "Windows NT Hique"
..."Free plus free
equals more than
NT"
Interesting note: www.fool.com is running Microsoft-IIS/4.0 on NT4 or Windows 98
As a developer I must say that getting good feedback from users and getting resect from users is far more important than earning (a lot) of money with the software.
So Open Source software benefits the author probably more than traditionally selling the product. (It always nice if you can get a living outof it, but that's never been the main issue I believe.)
Wouldn't you be doing the same stuff if you were not paid for it? (developers view)
a-quite-happy-soft-dev
StarTrek.org Free Webmail
So he says, but he's pretty much ignoring everything that happened in the anti-trust trial. I don't want to go over the trial analysis that everyone here is sick to death of, but MS do have the resourses -- both in terms of money and influence -- to damage whomever they wish, including Linux.
However,
Microsoft's high end is the desktop, and that's the last market Linux will take over.
This is a telling point, and one that is rarely made in analyses of Linux. We'll get there in the end, but only after all the other markets have fallen.
Do a poll: if network admins could fix bugs as they came up, would they? The people who did the 3D water effects for the Titanic movie used a huge Linux cluster for processing. The kernel had limitations on the Alpha platform that prevented it from working right, but considering the money they'd saved over an NT based solution, they just put some ressources into making it work properly (and better than NT). Everyone now has a more stable Alpha platform.
As for documentation and some of the other comments you've made, the man pages and other Linux documentation are often more complete and more TRUE than MS' documentation on any given issue.
- Michael T. Babcock <homepage>
- Michael T. Babcock (Yes, I blog)
- Ease of use
- Rapid Prototyping
- Power
- Speed
:+)
Personally, I think you should go for the middle ground under Windoze - use Delphi, which has the RAD and Drag/Drool frontend of VB, but with almost as much power as C - not to mention that for many years, Borland^WInprise^WBorland had the fastest and most optimising compiler in existance for the Intel platform - I don't have any current stats, but imagine it must still be pretty goodVB is VASTLY easier to use than VC - so much is just click-drag-and-drool that in VC takes a week of headscratching to find the obscure MS-Specific call that does $FOO
With VB, you can have a "look and feel" prototype, will all screens in their final configuration but "dummy" fixed-text data, in about ten minutes. you can single-step in interpreted mode without having to recompile, you can add and remove forms, modules and so forth quickly and easily without having to recode half your calls.
With VC++, you can do anything the OS is willing to let you; in return for not having a pretty button for everything, you get a list a screenfull of buttons couldn't cover.
Compiled C is just faster and smaller than VB - and less files. there isn't a why, it just is
--
-=DaveHowe=-
Check the stock price...MS is going where that stock is going so until it heads south they are not in trouble. MS is a corporation where general public opinion doesn't matter that much and technical opinion matters even less. What really effects their bottom line is corporate and financial market optinion. They don't seem to even see the cracks in the empire yet. Heck, I bet that all the pundits predicting their demise are buying stock right now.
All things do come to an end but I don't think that MS is going anywhere soon. There is still one tactic that they have not tried yet: making high quality, inexpensive software. MS could do it. They do have a lot of smart people working there. Their products lack quality often due to unreasonable deadlines and impossible requirements. I don't blame the programmers. I blame the managers and higher ups.
What if MS decided to give all those softies a chance to do things right?
-- soldack
Linux is based on Minux. Minux was written in part to give students a version of Unix simple enough to study completely. In that sense Linux was "reverse engineered" in that it was based on Minux and Minux implemented many of the features of Unix. This is similar to how Compaq reverse engineered IBM's PC bios. They studied what the code did, not the code. Then they tried to write code that did the same thing. Linux falls into this catagory.
-- soldack
The implication would be that there are no Open Source projects that have "died out" for lack of continued development. This is just plain silly.
Open Source project will continue as long as people are interested in the effects, but the network effect that he describes is important here. If another alternative is available and better, then the development will stop or never start.
This is a demonsration that Linux (and Free Software in general) is unFUDable. If a project dies it is because of a REAL problem, not because the competing company says that it is dead technology.
Do you know of any Free Software project that dies because his competitors where saying it is dead technology, or it is "hobbyist" software...? Personnaly I can't think of any (but maybe you can).
Would GIMP have started if Photoshop was free (or even dirt cheap) and available on many platforms?
Yep, because Photoshop still wouldn't be free (speech), but the project most probably would have begun at a later date, because there wouldn't have been such a need for it (because of the proprietary alternative).
"The obvious mathematical breakthrough would be development of an easy way to factor large prime numbers." Bill Gates,
As long as Linux remains an open source project, its success will be limited to only a niche.
Linux has already gone beyond niche status in the server market, and is currently attracting a great deal of attention in the desktop, point-of-sale, embedded systems, and supercomputing markets. And its presence seems to be growing in all these areas (one of which, BTW, does not have any MS presence at all).
Oh, yeah, about your "as long as" -- Linux will always be an open source project, due to the nature of its license.
Linux lacks organized and official testing, and it lacks paid engineers who make their living improving it.
Microsoft has both, and look at the two decades worth of crap that they have released on the public.
As far as "make their living" goes, you must be thinking about some other species if you don't think more quality goes into a labor of love than what goes into a labor for wages.
But this is the real world, and people only do so much bona fide work.
They've already done enough to create a robust and multifeatured operating system and a whole pile of utilities and applications to run on it. In fact, they had already done so long before Linux came over the public's horizon. And now the coder base seems to be expanding faster than before.
they will fail due to lack of integration bonuses that you get with Windows.
Windows only "integrates" with other Windows systems. Linux integrates with almost anything. The biggest problem is the deliberate incompatibilities Redmond introduces with each new release of a product.
Windows will have the stability and reliability that Linux has
Sorry, but stability and reliability are not optional add-on features. The only way Windows will ever become S&R is when Redmond decides to re-write it from scratch with S&R in mind.
Of course they might. They're talking about S&R now that people are citing that as a reason for ditching their NT servers. But they started talking about "ease of use" the day they heard about the Mac, and it took them 12 years to get an approximation of what the original Mac offered. I think there is extremely small risk that Windows is going to undercut the competition in terms of S&R anytime during the next decade.
The recent Linux craze is a flash in the pan and will not result in significant change.
Linux has already been around for 8 years, and is already resulting in significant changes.
Microsoft is a tenacious foe, and will continue improving its products at an amazing rate.
Continue? Microsoft has never improved its products "at an amazing rate". Unless you're willing to be amazed at how slowly they improve things. The only thing they're tenacious about is maintaining their market share, and they've always done that by means uncorrelated with the quality of their products.
One thing Microsoft understands that Linuxers do not is the importance of excellent developer tools and support. Microsoft has excellent IDE's for each of its languages
If Microsoft's tools are so wonderful, how come their products are so crappy? (I think you have just presented an argument that handcrafted code is the way products should be developed!)
support for these tools are horrible
No, it's wonderful. When's the last time you wrote directly to a Microsoft developer and got help with a problem? What's the average turnaround time between the discovery of a bug in a Microsoft product and delivery of a fix?
source code licensing is ambiguous
No, it's perfectly clear.
Linux lacks a component model beyond shared libraries (basically, DLL's)
DLLs are a caveman's attempt at shared libraries. For some reason Linux's shared libriaries never resulted in a phrase like "DLL hell."
Linux needs something like COM
Linux has "something like COM." And just as with DLLs, the Linux variants are better thought out.
Microsoft has a lot of weapons on its side
Well, yes, here's something we can agree on. A monopolist's grip on the market. All the lawyers, marketers, astroturfers, and congressmen that money can buy. A paranoid monomaniacal asshole coordinating all their efforts against any competition that peeps up out of its foxhole. Yessir, they have lots of weapons. But nothing that would seem to lead to a better product.
Linux advocates should seek to emulate them.
No, Linux advocates should view them as damage and route around them.
I have had to use Visual Basic at work since the version 3 days. Although it has come a long way, I do not feel that it is a truely reliable language. If it is so good, then how come MS does not use it for its own applications? VB is really built on C/C++ active-x controls anyway. Why not just use Visual C++?
Basic would probably be dead if it was not for MS pushing it. Since they control it they have no competition in making compilers for it. Other languages, like C, C++, COBOL, Algol, Pascal, LISP, Java, etc. have competition. Their are choices and this forces the manufactures to imporove their systems to compete better. MS doesn't have to do that with VB.
The basic language does not support many things that programmers need. The implementations of many of their types are hidden from the vb programmer. A good programmer needs to know these things in order to write efficient code. For example, variants and arrays are totally hidden from the vb write. I have worked with passing VB types to C DLLs and have seen how these types are implemented. It is pretty ugly. One line of basic code can turn into many lines of C code and even more lines of assembly.
Finally VB forces the developer to rely on vast array of active-x components that are also hidden from developers. Only by trial and error can a developer guess how a control is implemented. If there is a problem a developer has two choices: write their own control or work around it. Either choice takes a considerable amuont of time.
VB represents everything that is bad in Microsoft. All flashy and pretty on the outside but ugly on the inside.
-- soldack
Customization: How many window managers are there for Windows? How many are there for Linux?
The term window manager really doesn't apply to Windows, at least not 9x, you'd have to totally rewrite USER, GDI, and possibly parts of KERNEL.
People get mad when their apps crash. People notice that for some reason their apps don't crash as much when they run them under Linux. Netscape seems to be an exception, at least for me.
FWIW, the glibc version seemed to crash quite a bit on my Caldera 2.2 system. I replaced it with the libc5 version and problems seemed to have been resolved...
If two things were equal in every way except that one was $90 and one was free, which one would you buy? And what if it turned out that the free one was better?
At least for corporations, Linux really isn't free (beer.) You have to pay for support, and having a CD and dead tree manual is really, really a good idea.
My journal has hot
The implication would be that there are no Open Source projects that have "died out" for lack of continued development. This is just plain silly. Open Source project will continue as long as people are interested in the effects, but the network effect that he describes is important here. If another alternative is available and better, then the development will stop or never start. Would GIMP have started if Photoshop was free (or even dirt cheap) and available on many platforms?
Actually, just picking some nits here, but there are viable cheap alternatives to Photoshop in the closed source world...
Two are entirely commercial: Micrografx Picture Publisher used to be very very expensive, but is now a viable alternative to PS for most uses for about $50. Recent versions of Corel Photopaint ($100) are really really nice and have good features to rival Adobe's product...
Another is shareware: Paint Shop Pro is available fo a fraction of PS's $800 street price and it supports special effects, layers and masks like the other packages I mentioned...the usual features you look for in a PS replacement...
My journal has hot
No, I agree more with the original post, though I think he/she is a bit extreme. Linux API documentation and knowledgebase management is indeed disorganized compared to Microsoft's MSDN.
Yes, but the availability of source code is a powerful countervailing factor. Microsoft is famous for giving the developers incomplete API specifications and then releasing their own appplications which have performance made possible only through the use of undocumented system calls. The result in user space is applications from ISV's that use undocumented API calls merely to be competitive. This leads to upgrade incompatability issues.
Having open source is the ultimate insurance of complete knowledge of the OS. Microsoft can have the fanciest 'Knowledge Base' imaginable, but if they choose to make some os functions 'undocumented' this Knowedge Base is in fact corrupt and untrustable.
If Linux gets an "LDN" sort of system going, it would definitely be a good thing.
Design of a real LDN would be I think nearly impossible given the nature of Linux development.
There is a "law" or marketing that basically says, "Name recognition = $"
The way suits think, MS is right just because they make money. It's the same way IBM was before.
However, according to the rules for the new economy, the more money you make, the more money you make. This is because this is a war for standards. Why do you think VHS won out over BETA? There were more VHS products out there. Not that VHS was better. People didn't want to buy a product that you couldn't get movies for.
That's different for Linux. Linux adapts to "standards" because it has a user-driven development process. If there is a need for any kind of product, that product will appear. It just depends on attaining critical mass of people with know-how and people with use-how.
Everything is a matter of time. I think the user-driven development can succeed in the long run because it has a faster turn-around time than conventional marketing-driven development.
I do what the voices on my console tell me to do.
Interesting piece, though I think it skirts one of the more interesting aspects of network effects. Namely that companies often have the ability to make network-type products compatible with others if they choose.
If some way of ensuring interoperability is guaranteed, e.g. via open communication standards, then *everyone* benefits from the network effect. It ceases to be a case of "my network vs. your network" since users can gain the advantages of both without being tied to either.
Of course, companies with potential monopoly power don't want competition, so may deliberately place barriers in the way of interoperability. This reduces competition, allows them to reap monopoly profits but most seriously prevents consumers from being able to gain the advantages of a unified network. The internet wouldn't be as great as it is today if it was fragmented into incompatible proprietary segments.
As I see it, if antitrust legislation in the next century is going to be effective then it will have to adapt to take this into account. Specifically, it should be made illegal to artificially create barriers to isolate your network from others in order to gain monopoly power in any sector. Such incompatibilities are an enormous source of consumer harm.
The relevance of this to the Microsoft case is pretty obvious, and I am sure everyone can think of a hundred other examples where companies have abused their ability to build barriers to interoperability. This is likely to become more and more important as networked technology plays a larger part in our lives.
Therefore, open standards are definitely the way forward from an economic perspective. If that occurs, then true market competition will reign.
Microsoft probably won't want to try to conquer Linux. Sure they have the money and the influence, but Linux, to the DoJ and friends, embodies anti-Microsoft-monoply, and if Microsoft starts messing with Linux, DoJ will jump on them. Microsoft knows this.
GNOME and KDE have made significant strides on making Linux a contender in the desktop environment. In some ways Linux has already overtaken Windows:
I'm sure I missed a few. Sure, Win has its advantages. But they are probably not going to stay for long.
Ken
Have you considered that Linux is overtaking the server market and not the client market? When you have to pay thousands for NT server, getting something more reliable, less expensive, and easily administered remotely from anywhere is something any sane administrator would do in a heartbeat.
Granted, Linux doesn't fit everywhere yet, but it can be used for many server applications right now.
It also adds to your skillset, and your employer loves you if you're saving money, so you can demand more money for your skills.
The desktop market is a different story. Given enough time, it can certainly gain market share there as well.
Oh gee, and the same level of support and bugfixes we've come to expect from Microsoft.
I like free software. I would pay lots of money for it if I had to. It's about being able to get a problem fixed as soon as it happens.
A monopoly is only 'bad' and regulated if it stifles any possible competition yet gouges consumers.
If AT&T had put their earnings into research on how to lower long distance rates, and had consistently dropped price, nobody would have tried to break them up.
If MS had made something like BeOS, small, fast, stable, pretty, and charged $45 for it, and published the APIs, etc, nobody would hate them.
The problem is that AT&T gouged consumers and user unfair trade practices to stop competitors. Much the same as Microsoft.
Even if Linux became a virtual monopoly, with it being hard to buy a PC without Linux installed, it wouldn't matter.
A company can't demand money for their distribution, so they can't hold companies at ransom with license fees. It doesn't cost the consumer any more to have Linux preinstalled. It would be done during testing anyways, at no cost.
Linux is served by being open, accessible... APIs are published and documented because they benefit from the same network effect as everything else.
Nobody would be harmed by Linux even if it was on 99.9% of new computers. Thus, nobody cares if it becomes a monopoly. Not only is there no central company to exploit it, but GNU/Linux (the GPL is important in this) can't be leveraged this way.
The only ones who could lose are Microsoft and other companies seeking to limit information.
Continue? Microsoft has never improved its products "at an amazing rate". Unless you're willing to be amazed at how slowly they improve things. The only thing they're tenacious about is maintaining their market share, and they've always done that by means uncorrelated with the quality of their products.
To be fair, Microsoft has shown tremendous capabilities for improving some of their products. The two prime examples are Windows 3.1-->Windows 95,and Internet Explorer. They have the deep pockets to be able to add features to a product at a prodigious rate. I think it is accurate to say that no organization -- commercial or OSS group, has ever demonstrated the ability to integrate complex features onto a project at a rate like MS can.
I believe that there is a fundamental bug in the MS revenue model that requires them to lead customers into to new technologies, not to adhere closely to customer's real world needs. Once you've sold everybody in the market a word processor, where is your next buck coming from? If quality were a matter of having the largest feature set, and innovation a matter of making rapid changes, Microsoft would have the best software in the world.
Microsoft's revenues are dependent upon their making rapid changes in their products -- the more frequent the releases the more frequent the revenue. But this is not necessarily the sames as continuous improvement. The OSS model is extremely fast at making patches to specific defects, but otherwise relatively slow. It's ready when it's ready. How long have I been salivating over KOffice? On the other hand, if you look at what now appears to be an extraodinary rate of improvementin KDE and Gnome, I think it is very clear that it isn't the raw rate at which features accrete into the system, but the surety that each step is a clear improvement over the last. When strung together these amount an impressive rate of improvement over the long haul.
If MS and OSS were chess players, the MS would make moves very rapidly, defeating its opponents by throwing them off balance and leading them into mistakes. OSS would be play with more deliberation, at each step improving its strategic position. If you set these players down at a chessboard, non-chess players would first be impressed by the speed with which MS made its moves, in seconds versus minutes for OSS. They might at first think that MS was a better player, but inevitably, MS would start to lose strategic position (e.g. a year ago MS was poised to capture the server market, now nobody thinks that's going to happen), followed eventually by material losses. It gets worse when you figure that MS is actually playing a game of simultaneous chess against an array of opponents, each OSS project has its own game running against MS, whereas MS's games have to be linked by a common strategy (e.g. dominance in tools is linked to dominance in OS).
I don't think that Microsoft is ever going to be forced out of the game. However, OSS is going to change the nature of the game it plays, force it to produce better products at a more deliberate rate, to listen more to their customers. I think that perhaps in five years there will probably be some MS products that I actually use because I like them.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
"It's only a matter of time, and everybody (ie., people other than slashdotters, tech-junkies, and other members of their species ;-) ) will be aware of the defiencies of MS products and the availability of better, cheaper alternatives."
Interestingly this is exactly how the author of the article described FUD: just wait and everything will get better!
The way I see it is that OSS works for software commodities, that is software that has been around long enough to loose its initial attractiveness. Once this attractiveness is lost people are no longer willing to pay enormous amounts of money for it.
Categories of software that can be classified as such are operating systems, word-processors, spreadsheets and so on. Interestingly this also applies to serverside stuff: mail servers, ftp servers, webservers and even database servers.
Companies selling software in the above list all use the same tactic: they somehow add value to the default functionality of the product thus making it special: ms bundles all sorts of stuff with their OS (which is still dos), they also bundle all their office stuff in one integrated package which in its turn mixes very well with the before mentioned OS. IBM and Oracle add management tools to their databases. Webservers are made attractive by offering management tools and integrating it with other stuff. The same applies to mailservers and ftp servers.
OSS is moving into new territories. The desktop, ms greatest added value to their OS, is becoming a commodity. So OSS will take over that too. MS has long realized this and has started to integrate all sorts of stuff in their OS. So far this tactic is working very well since the most used argument not to migrate to linux (on the desktop) is that windows applications just integrate better and are more feature rich. In due time this tactic will no longer work since OSS can provide users with the same level of integration (KDE) and functionality (many companies are working on or pondering linux version of their products).
What I hope I showed with this is that OSS is not suitable for all software, only the commodity software. Luckily, increasingly more software falls in that category.
Jilles
I wish people would use scare quotes every time they used "desktop environment" with the restrictive and dedicated sense of that computing environment used by people who don't really know anything at all about computers and who don't want to, as opposed to the more intuitive and generic use in which it means "the set-up on people's personal computers", which obviously isn't what you meant even though it's what you said. It's as strange to many of us as using the term "personal computer" to mean "Intel-based IBM PC using whatever Microsoft wants you to use" rather than the more intuitive and generic "computer used by a particular person".
Sometimes it's really as though the Linux people were speaking Unix with a Microsoft accent. Be very afraid. That's how they want you to speak, and thus, to think. :-(
Not everyone derives their working vocabulary from the insidious Microsoft propaganda virus, you know--especially in this forum of all places. The whole idea of controlling the word-choice in order to control the implicit agenda and by entension, the entire world view is surely some Orwellian nightmare invented by somebody's marketing department trying to create tacit corporate branding on innocuous, general-purpose words. It's like the pro-life/pro-choice spin. Or the PC police (please take that in multiple senses) all over again trying to weed out choices that have a message they don't like.
Nothing you list there is inherent to the Linux kernel, nor even to any of the dizzying panoply of Linux-based operating systems that use some flavor of the same. I certainly don't see how any of these GUI toys really have anything to do with Linuces in particular rather than any of the myriad other Unix systems running X11.The only thing that seems to distinguish itself is the price-point criterion, but it's unclear how important that truly is. First off, the Linux-based family of operating systems are hardly the only free OSes around. But more important, really, is that the price for even costly OSes continues to go down. So I can't see that, even if it were one today, this would long remain that distinguishing a factor. For many situations, a two-digit dollar amount is hardly enough to notice compared to all the other factors. It just gets lost in the noise.
As long as Linux remains an open source project, its success will be limited to only a niche. Linux lacks organized and official testing, and it lacks paid engineers who make their living improving it. People are expected to fix bugs for free. But this is the real world, and people only do so much bona fide work. Sure Red Hat might begin to address some of these problems, but they will fail due to lack of integration bonuses that you get with Windows. By the time Linux gets the ease of use and ease of integration that Windows has, Windows will have the stability and reliability that Linux has, *AND* will still have 90% market share. The recent Linux craze is a flash in the pan and will not result in significant change. Microsoft is a tenacious foe, and will continue improving its products at an amazing rate. One thing Microsoft understands that Linuxers do not is the importance of excellent developer tools and support. Microsoft has excellent IDE's for each of its languages. Each of MS's language tools targets a specific type of developer. Each of MS's language solves differents sets of problems. And MS keeps a great on-CD Web-available knowledge base and reference called MSDN. Linux, on the other hand, has a rag-tag bunch of disparate tools, rely on the command line and makefiles, and has only Usenet has a poor excuse for a knowledge base. Furthermore, support for these tools are horrible, source code licensing is ambiguous, and documentation is a confusing array of man pages, which is nothing compared to the hyperlinked MSDN web interface. Finally, Linux lacks a component model beyond shared libraries (basically, DLL's). Linux needs something like COM to help organize its uncoordinated development efforts. Microsoft has a lot of weapons on its side, and Linux advocates should seek to emulate them. Linux advocates should be learning from Microsoft, not blindly bashing everything they do.