EU hits Microsoft with Unprecedented Fine

[L233]

Another good read:

Press conference on imposing penalty payments on Microsoft – introductory remarks
Neelie Kroes
European Commissioner for Competition Policy


I'll just quote the whole damn thing:

The European Commission, ladies and gentlemen, is entrusted by the EC Treaty to verify and ensure compliance with EU law.

In the case of Microsoft, I sincerely regret that the company has still not put an end to its illegal conduct.

It is now more than two years since the Commission’s March 2004 Decision that found Microsoft to be in violation of the EU’s anti-trust rules by abusing its dominant market position.

It is more than 18 months since an Order from the President of the Court of First Instance required Microsoft to comply with the Decision without delay.

The European Commission cannot allow such illegal conduct to continue indefinitely.

No company is above the law. Each and every company, large or small, operating in the European Union must obey EU law, including competition law, to the benefit of all companies and European consumers.

That is why the Commission has today fined Microsoft a total of 280.5 million euros for its failure to comply with the March 2004 Decision as regards the requirement to provide complete and accurate interface specifications, so as to allow other companies to make their products interoperable with Microsoft PCs and servers. The fine corresponds to 1.5 million euros per day, during the period from 16 December 2005 to 20 June 2006.

This is the first time ever, in the 49 year history of the European Union, that the Commission has had to fine a company for failure to comply with an anti-trust decision.

I hope that it is also the last.

In deciding to impose this fine, the Commission has taken a cautious approach to assessing Microsoft’s compliance.

The Commission has looked carefully at numerous reports by the Monitoring Trustee – a highly capable professional recommended by Microsoft itself – who has been assisted by experts of world renown. It has looked in depth at Microsoft’s documentation. It has considered very carefully the replies by Microsoft and its experts to the objections raised by the Commission.

Having carefully weighed all of the evidence, only one conclusion was possible. From 16 December last year to 20 June of this year:

Microsoft did not even come close to providing complete and accurate specifications
Microsoft therefore remained in breach of the Commission Decision and
The Commission had no option but to impose a penalty payment on Microsoft.

The fine is a substantial amount of money. The Commission has shown restraint in setting the level of the fine, seeking to do no more than is necessary to induce Microsoft to comply with the Commission’s decision.

Microsoft has told me that they are now devoting substantial resources to compliance.

It is a great pity that they did not do so two years ago, before there was the threat of imminent fines.

If the compliance effort had begun in earnest in March 2004, the burden on Microsoft’s staff would have been much lighter.

I don’t buy Microsoft’s line that they did not know what was being asked of them because the March 2004 Decision is crystal clear.

The Decision states that they have to provide complete and accurate specifications that will allow the development of interoperable products. In the period covered by today's decision, Microsoft's documentation fell significantly short of this requirement.

If the latest documentation is not sufficient to comply with the March 2004 Decision, the Commission will once again be forced to consider the imposition of fines.

And in order to increase the incentive for Microsoft to comply, the Commission has decided that the ceiling for potential fines will be raised. Should Microsoft not comply with its obligations to provide complete and accurate specifications at an appropriate level of remuneration, then – with effect from 31st July – the potential fines will be raised from 2 million euros per day to 3 million per day.

I sincerely hope that further penalty payments will not prove necessary for the quality of the documentation, the remuneration required for access to and use of the documentation, or for any other aspect of compliance with the 2004 Decision.

Full and complete compliance with the March 2004 Decision would be the best outcome for the company and for the company’s European customers.

 

[DemoCoder]

That's wrong. They can bundle it, they just have to offer a version without it - which they did months ago (and no one really cared btw). I kind of agree that the media player issue is a bit silly but you got it wrong.

Pointless. Why should they have to unbundle it? I could see an argument for saying that system builders should be able to bundle other media players and set them as the default system player for various mime types, and users can have the choice of choosing what default they like, but I see no point to removing the code from the operating system. With with Windows Media Center built into Vista, along with a HD-DVD playback, are they going to force MS to unbundle these as well? The whole raison d'etre is to work in the living room, and be slick and easy to use.

The bigger issue has always been interface documentation wrt non-MS work group server solutions and that's what the current fine is about. It's never been about opening the source code either, that's simply MS FUD.

More pointlessness, since you can drop a Samba server into a Windows network and have it become the PDC, WINS server, share files and export printers, and it all works fine. I am an ardent supporter of open standards, and I am a member of IETF, W3C, OMA, and OASIS standards organizations, but I really don't see the point in compelling Microsoft to release documentation here. Microsoft is in fact, being gradually eroded in the work group arena by open source alternatives, as well as Web oriented applications. File sharing works fine with SMB, Kerberos interoperability means you can drop in your own single-sign on server, etc

The reason why MS continues to hold marketshare when free alternatives exist is not the cost of the software but the human element. Frankly, MSCEs used to administrating all-Windows networks are not likely to want to manage a Linux/BSD or OS X alternative, when it is less of a headache to know only one operating system and have a single company to go to for support.

Secondly, groupware is a commodity. It is not going to be a compelling market to offer competiting implementations. When you install a commercial Unix, do you go out and buy a commercial NFS implementation? No, you use the one that came with your OS because it is "good enough" Only in rare cares to people buy specialized non-bundled implementations, usually because they have specialized needs.

The government is behind the curve. By the time they try and legislate some technology decision, it's already irrelevent.

I don't think it's all that unreasonable. There's no denying that MS has leveraged its de facto monopoly in PC operating systems to stifle the competition. If MS were to design their OS more modular, i.e. enable system builders to easily customize it with regard to certain components like media player, web browsers etc. they wouldn't have all these problems. The completely unneccessary tight integration of certain components is IMO clearly designed to force their solutions down everyones throats.

Pick-and-choose modules would lead to an end-user support nightmare, as well as a detriment to user interface consistency. Integration *matters*. On Mac OS X, integration between apps is very tight. So spotlight appears in every app, Bluetooth from iSync appears also in my system address book. RSS from my browser appears in iTunes, can affect my calendar, can integrate with iPhoto and iWeb. iSync can publish iWeb files automatically, and so on.

You could make public all of the interfaces, but that can also impede innovation in some circumstances where it is best to keep implementation details private *so that they may be changed*. If Apple, for example, had a private interface between its apps for sharing say, bluetooth received vCards, and for some reason, this interface was poorly designed, it could change it, and rebuild *ALL* of the system apps that use the interface, without worrying about breaking existing apps.

If instead, the interface were made public, they would be forced to continue to support the old interface in all applications, as well as introduce a new Interface2. We've seen this DLLHell before, Microsoft has it bad enough, mostly because a truly staggering number of their interfaces *are* public.

I don't have a problem with system builders being able to include alternatives as default, but I think it is stupid to force MS to unbundle their versions. For example, although Firefox can stand-in as a webbrowser default, IE is used for far more than just browsing, and Firefox is not feature-for-feature identical to IE, so if I, as a developer, with to use IE HTML component in my desktop application, and use IE specific extensions, I can't, without distributing IE itself with my application.

Also, MS has always been very tight lipped about certain interface specifications, document formats and such which is also a deliberate attempt to block competitors from offering interoperable products.

Why should they? Everyone is free to create their own document formats. If customers don't like vendor lockin, they can migrate to those formats. There are plenty of alternatives to Office, there is no need to compel MS to open their formats up. The open source community has demonstrated that given time and effort, they can clone most MS functionality needed with an open alternative.

I am no lover of MS. I hate what they tried to do with OpenGL, SMIL, etc and there is no doubting that they try block open alternatives when they perceive they can. I just don't think the government needs to be involved in software design decisions, about whether a particular protocol or interface is made public to developers. It is not *always* beneficial to do so.

I could arguably make the case that although MS tried to kill OGL for gaming, MS improved graphics architecture ALOT and competition with DirectX forced the truly glacial pace of OpenGL to speed up, resulting in a net-benefit for OGL users. Had there been no DirectX, I highly doubt OGL would have had DX9 or DX10 level features at this point.

 

[DemoCoder]

Read L233's post. It's all about interfaces.

In software design, programmers make careful decisions all the time as to whether an interface should be public or not, depending on expectations of whether it needs to change in the future or not, whether multiple implementations make sense, or whether dynamic linking/late binding, or static linking is going to be used.

Having the government compell software companies to open interfaces without regard to whether or not it is the best software design is the fallacy.

 

[L233]

Democoder, I cannot claim to have any insight into the the whole interface issue wrt group ware. I'll just take your word for it and I wouldn't be surprised to the EU was in fact "behind the curve".

Ultimately, this is irrelevant. MS was required to to comply and they didn't. They had plenty time to do so. They were either trying to stall, had some really shitty legal council or assumed the EU would pussyfoot when it comes to enforcing the decisions that were made. Someone at MS apparently made a bad call and it cost the company some spare change. Tough shit. I'm still too much of an "OMGz teh eVIL MeGaCoRpz" tinfoil hatter to feel sorry for them.

The fine wasn't about MS's lack of compliance with the WMP ruling - they did that and it obviously wasn't too hard to do, so I still feel your first post missed the mark.

Anyway, I don't see much of a problem here. Why would modules lead to an end user support nightmare? The support is the problem of the OEM that makes the changes, it's not Microsoft's problem because MS doesn't offer support for system builder licenses and refers you to the OEM. MS only supports fully packaged products (which btw aren't available anymore in Germany) and even then it's only two support requests for free.

Also, I think your user face consistency argument is funny because WMP is likely the worst offender. Not only does the interface suck, it's blows. Not even it's visual design is consistent with the rest of WinXP. The reason why I kind of agree that the WMP issue is a bit silly is because the damn thing is such a piece of shit that I haven't used it in years. Still, I don't mind the EU assraping MS over the WMP issue because it maybe ever so slightly interfere with MS's plans to shove their Digital Restriction Management crap down everyone's throat and make us all the content industry's bitches and I don't like being a bitch.

Different standards apply to companies like MS. That's where your Apple analogy falls short. MS domintes the market for desktop and workstation OSs and this requires regulation. I think you're looking at the issue from a perfectly reasonable techie perspective and you ignore the bigger issue here. Remember what happened after MS leveraged their OS monopoly to push Netscape out of the browser market. Even if you think that MS including IE in windows made perfect sense (and IMO it did), the final outcome was a disaster. MS had won and they abandoned the product. How many billions of dollars of damage could have been avoided if MS hadn't stopped developing IE until they were forced to continue developement by Firefox? A good part of all this spyware, maleware and worm shit was only possible because MS didn't lift a finger to improve IE after they had achieved their goal of market dominance. That's a pretty solid argument for regulating the living shit out of MS on principle.

The goal of competition laws isn't neccessarily to make to most reasonable decisions from a tech perspective or even to be "fair", it's to maintain a functional market because in the end that's what's best - even if you break some eggs in the process.

 

[DemoCoder]

Netscape was pushed out of the browser market because of inferior software design. Netscape 1.0 was a hacked Mosaic codebase. They continued hacking the codebase all the way through Netscape 4. MS IE 1.0-3.0 were hacked Spyglass codebases. MS made the gutsy move to rewrite the HTML renderer for IE4, offering sophisticated incremental layout, better CSS, DHTML, XML, et al support. Netscape did not bother to rewrite their rendering engine, which made it literally impossible to support anything but DOM Level 0, and no DHTML except the <layer> tag. Netscape also did not offer a COM control version of Netscape because they did not want the Netscape "brand" hidden, therefore developers could not easily integrate Netscape into applications except through OLE/DDE.

When Netscape finally released the source code to the public, everyone saw how TERRIBLE it was. The Open Source community couldn't do much with it, and it was practically rewritten. It took a team of former netscape engineers and the open source community 2 YEARS to get a solid version of the rewritten Gecko HTML renderer done. It was in this two year gap, between Netscape 4 and Mozilla, that Microsoft destroyed them.

The Netscape browser was truly a piece of shit. Anyone who is a programmer and took a look at it knew it. There was no way they were going to compete with IE4/5/... without a complete rewrite. The management of Netscape Communications are at fault for not plowing all company resources into a rewrite of the browser engine. Instead, they fscked around with servers (killed off by Apache and IIS), and "push" technology, oh, and they wasted time making Java development tools.

People outside the netscape development community don't realize this, and think that the only reason Netscape died was MS bundling. The reality was, IE4/5 were lightyears better than Netscape 4, and Netscape Communications was thoroughly unable to deliver anything competitive due to bad management.

Even the current design of firefox has problems, ergo, "stripped down" browsers like Camino, et al. Why? Firefox has memory leaks out the wazoo, and the user interface is built entirely in XUL and JavaScript instead of C++, which leads to more memory usage, and poor performance on some platforms. I like Firefox alot, and it is my only web browser, but they really need to fix the memory problems. A web browser should not chew up 200-500 megs of VM.

 

[L233]

LOL, yes, Netscape was a piece of shit, alright. I remember it quite well. Terrible piece of software, even from a mere user's perspective. I still think that doesn't really invalidate my argument with regard to MS abandoning IE. It's a showcase for what can happen if one product is allowed to dominate the market and that justifies regulation, even if it may make little sense in the short term.

I don't use Firefox, mainly because I hate the fact that I have to spend days to find and install crappy plug-ins just so I can emulate some of the features I sorely miss from Opera. If the "Orca" browser (the Gecko version of the Avant Browser frontend for IE) ever gets released, I may switch to that.

 

[_xxx_]

Netscape was pushed out of the browser market because of inferior software design.

Errm, Netscape 4 kicked some serious shite out of IE back then features and usability-wise. Total domination.

 

[Frank]

More pointlessness, since you can drop a Samba server into a Windows network and have it become the PDC, WINS server, share files and export printers, and it all works fine.

The again, it took them six years to reverse-engineer, and they're only allowed to use the NT 3.0 RPC interface, as Microsoft won a court case that disallows the Samba team to directly use the 4.0/4.2 or AD interface. And yes, I know you can get quite some way with only LDAP, Keberos and such, but while you're able to access the files, things like managing security still require you to use multiple different tools in different OSes. So it's still easier just to use Windows.

And that is the point.

 

[Sxotty]

Err, those "features" are mostly crap anyway.

Also, I don't think MS will pay this, they'll drag this through courts for years and it'll eventually be settled.

Demo: The problem with the Media Player was that you couldn't choose not to install it or were not able to uninstall it after the windows installation. Same goes for IE.

You can unistall media player, and btw they already have a version in EU countries w/o that stuff guess what no one wants it. Why would anyone want to pay tte same and get less? I understand the desire to open the playing field up for 3rd party developers, but this is utter crap. The EU is just playing politics with it and it is pretty shameful. Yes monopolies are bad, but the way to get around it is not to sue for including IE or media player, it is to allow other people to make things like lindows etc... Anyway there does need to be open source stuff, if the basic kernel could be open sourced so others could make GUIs etc that would be nice, but from what I understand that will not be as easy in Vista to separate them.

Still in the ultimate nice world there would be a common very generic base that would allow programs to run, and then they could use different OSes over the top, but the problem is then programers would have to do more work and I doubt they would want to.

Anyway let's hope OpenGl doesn't die so at least DX doesn't become the only game in town as well.

edit: I see you already realized most of what I said. Not complying with a verdict that is wrong is stupid I admiyt that. MS should have paid then appealed afterward.

 

[_xxx_]

You can unistall media player, and btw they already have a version in EU countries w/o that stuff guess what no one wants it.

That happened much later though, after the pressure from the EU. I think it was XP SP2 which brought the possibility to uninstall Messenger, IE, Outlook Express and the rest of the malware which some people actually like to use (or be used by as I see it ).

 

[DemoCoder]

The again, it took them six years to reverse-engineer, and they're only allowed to use the NT 3.0 RPC interface, as Microsoft won a court case that disallows the Samba team to directly use the 4.0/4.2 or AD interface. And yes, I know you can get quite some way with only LDAP, Keberos and such, but while you're able to access the files, things like managing security still require you to use multiple different tools in different OSes. So it's still easier just to use Windows.

It will always be easier to just use windows. NT RPC has a history of introducing security flaws you can drive a truck through, and many companies choose to deactivate it. The benefits to security of using AD/Windows instead of LDAP/Kerberos are dubious.

Any company which has a heterogeous network of machines (Windows, Linux, OS X) isn't going to want to use *Windows* specific group-ware features, and any company that is absolutely homogeous is going to want an all-MS solution anyway and isn't likely to want to buy a third party solution for something that costs them nothing nowadays anyway.

The government is trying to preserve competition in an area of the market that should be commodified.

 

[KimB]

There are other issues than networking at stake here, DemoCoder, such as development of 3rd-party software like media players, web browsers, word processors, etc.

Edit:
And why the hell should networking interfaces be commodified under Microsoft technology?

 

[DemoCoder]

Errm, Netscape 4 kicked some serious shite out of IE back then features and usability-wise. Total domination.

You're on drugs. Netscape 4 was inferior to IE4 in every way conceivable. Netscape 4 didn't support XML, CSS support way *TERRIBLE*, had really buggy table support, didn't support incremental reflow, had a horrible "resize window" problem that would fubar up the page or *reload it*, was much slower than IE at rendering, did not support dynamic modification of DOM or CSS, except limited layer positioning, had a terrible editor (Composer), had demonstrably faster Java and Javascript performance, the list goes on. It was a piece of shitty spaghetti code that was continually hacked on from the Mosaic codebase days. The first Netscape version was slapped together and hacked quickly to get it to market, and the company managed the result of that legacy for the next few years.

I was a devout NN user from the very first alpha version until MS shipped IE4, I never looked back after that. Netscape was a badly managed company full of hubris that they were going to kill Microsoft, and true to form, they focused on branching out into too many areas trying to fight MS rather than focus like a laser on the browser.

"It's the browser, stupid!" should have been their mantra. Atleast Opera seems to have understood this.

 

[Frank]

It will always be easier to just use windows.

Well, that depends. I really love simple terminals to manage stuff, you can go anywhere from everywhere that way if you know your stuff, and most people who have the experience required do so as well. Which is very hard and takes much more work with Windows. RDC or VNC is what you need for that, and slow as hell if you haven't got a fast link. And automating simple things in a Windows network is prehistoric, compared to *nix.

NT RPC has a history of introducing security flaws you can drive a truck through, and many companies choose to deactivate it. The benefits to security of using AD/Windows instead of LDAP/Kerberos are dubious.

Definitely. And, because of some serious flaws in the OS/API itself, I had to circumvent most of the security to get some simple things to work. So I know of and had to use a few of the most glaring holes.

Simply put: half the RPC calls that require authentication are circumvented by Windows itself by elevating the rights to admin level when the caller is a simple user...

Any company which has a heterogeous network of machines (Windows, Linux, OS X) isn't going to want to use *Windows* specific group-ware features, and any company that is absolutely homogeous is going to want an all-MS solution anyway and isn't likely to want to buy a third party solution for something that costs them nothing nowadays anyway.

Exactly. So, give us the opportunity to develop the tools we need, legally! Because cracking Windows to do it won't score you points, if admins or Microsoft learn about that. And it's illegal by the US Millenium Copyright Act, which makes circumventing security measures a criminal offence.

The government is trying to preserve competition in an area of the market that should be commodified.

Yes/no. Can you explain a bit more how you see that? I agree readily that it should be possible for anyone to write whatever they want to. But that the OS should be bullet-proof enough to make that possible, while keeping security intact. Which would require quite some effort from Microsoft, as half their security exists of "non-disclosure". Which is very bad, and half the reason they don't want to open up.

 

[DemoCoder]

There are other issues than networking at stake here, DemoCoder, such as development of 3rd-party software like media players, web browsers, word processors, etc.

You can develop third party media players today, Microsoft isn't stopping anyone. You just can't *sell* them effectively, because MS is giving them away for free. I mean, do we really want to bring back the days of having to PAY for web browsers? The people whining over this are commercial companies trying to sell commercial software.

And why the hell should networking interfaces be commodified under Microsoft technology?

Because it's um, Microsoft's operating system? If you run Linux, then you network interfaces are commodified under a batch of Linux protocols. If you run OS X, they were commodified under Apple (Boujour for example) There's nothing stopping anyone from selling alternate file sharing. Infact, at the last two companies I worked, they didn't use MS file sharing, they used NFS on Windows and WebDAV. NT already permits user installable file systems, user installable pluggable authentication, user installable print services, ad infinitum.

The debate here is over whether third parties should be allowed to sell products which clone MS's proprietary protocols and can replace MS versions. For me, this is a false debate, since I don't give a shit about MS proprietary protocols, as long as I can install a suite of Open Protocols, it's fine. The all-windows-shop has made their choice, if they don't want to use MS products, then migrate your clients to a non-MS protocol.

MS's interfaces are only relevant for out of the box all-windows-shop functionality. If you don't want an all windows shop, the relevant software is there and nothing stands in the way of third parties.

In the case of file sharing and authentication, the solution was there already for a long time, long before Microsoft dreamt up NetBIOS, WINS, AD, SMB/Lanman, et al. I've been in this industry for almost 15 years, and we were doing the same things way before MS introduced their solutions to the problem, so I could care less about idiot MSCEs who ran their shops into the ground with an all-MS network and then figured out later they want to get a third party implementation. They made the wrong decision and the correct solution is not to buy a third party MS protocol implementation, but to buy /install an Open Solution.

MS has no ability whatsoever to kill off competing workgroup protocols, none. They also have no ability to stop people from using third party workground software. I see zero cause for forcing them to open up these protocols. I see zero benefit to the end user. What's more, treating MS protocols as some kind of open interface with multiple implementations merely puts more power into MS's hands because MS *still controls the definition of the interface*.

An interface isn't open unless other people are permitted to evolve it. So a completely 100% documented SMB would still be 100% owned by MS just like Java is owned by Sun and .NET is owned by MS.

Companies that want otherwise need to stop legitimizing MS protocols and get involved with open movements like W3C/IETF/OASIS/ODF/etc, otherwise, a world of 5 implementations of SMB just locks people into to SMB, with MS owning SMB's definition.

Would you like it if MS owned all the rights to control the API of OpenGL, but they simply documented 100% of any changed they made to it? It still gives them a huge competitive advantage.

 

[DemoCoder]

Yes/no. Can you explain a bit more how you see that? I agree readily that it should be possible for anyone to write whatever they want to. But that the OS should be bullet-proof enough to make that possible, while keeping security intact. Which would require quite some effort from Microsoft, as half their security exists of "non-disclosure". Which is very bad, and half the reason they don't want to open up.

It's simple. Commodification means that file sharing should be built in, out of the box, and that it should be mature enough that it doesn't change often (in features) and therefore vendors really only compete on price. Since the marginal cost of producing software is zero, the price of a commodified piece of software should be zero.

That means file sharing should a) be built into the OS and b) be free. That means there is no market for competition.

Typical examples of commodity: standard consumer disc file system, GUI, TCP/IP stack.

There was a time when multiple companies sold competiting implementations. Rememeber Trumpet Winsock? Remember KA9Q?

Do you think that Microsoft should "unbundle" the Windows TCP/IP stack so that third parties could sell competiting implementations to consumers? It's a commodity now.

 

[KimB]

In the case of file sharing and authentication, the solution was there already for a long time, long before Microsoft dreamt up NetBIOS, WINS, AD, SMB/Lanman, et al. I've been in this industry for almost 15 years, and we were doing the same things way before MS introduced their solutions to the problem, so I could care less about idiot MSCEs who ran their shops into the ground with an all-MS network and then figured out later they want to get a third party implementation. They made the wrong decision and the correct solution is not to buy a third party MS protocol implementation, but to buy /install an Open Solution.

But this is a huge obstacle to businesses moving to other operating systems. When buying a new PC, you're not going to want to change the software of the entire network: you want it to work just fine with what you've already got set up.

And when have we ever seen Microsoft, or any other company, change their protocols to rub out the competition? It just doesn't sound feasible, because it would cause a whole lot of their own software to stop working: that is, there would be a very large cost associated for Microsoft to attempt any such change.

 

[Frank]

Simple statement: I would want all my software to be Open Source. Including the stuff I write myself.

But we still aren't there, as we still haven't got the business model for it. Although for what I use and do, providing services is close enough. The only thing I cannot get around is Windows. Mostly because that's the basic requirement, whatsoever. I hate .NET for that, but not with a passion.

 

[KimB]

It's simple. Commodification means that file sharing should be built in, out of the box, and that it should be mature enough that it doesn't change often (in features) and therefore vendors really only compete on price. Since the marginal cost of producing software is zero, the price of a commodified piece of software should be zero.

That means file sharing should a) be built into the OS and b) be free. That means there is no market for competition.

But without c) Operable with other operating systems, you still are locking users into a single platform. OS developers won't be competing on price, but instead those companies that have used Windows in the past will find it very hard to migrate to a different OS.

 

[Frank]

It's simple. Commodification means that file sharing should be built in, out of the box, and that it should be mature enough that it doesn't change often (in features) and therefore vendors really only compete on price. Since the marginal cost of producing software is zero, the price of a commodified piece of software should be zero.

That means file sharing should a) be built into the OS and b) be free. That means there is no market for competition.

Typical examples of commodity: standard consumer disc file system, GUI, TCP/IP stack.

There was a time when multiple companies sold competiting implementations. Rememeber Trumpet Winsock? Remember KA9Q?

Do you think that Microsoft should "unbundle" the Windows TCP/IP stack so that third parties could sell competiting implementations to consumers? It's a commodity now.

Agreed.