Microsoft patents: Linux to become VFAT-free?

In light of the recent Microsoft-TomTom VFAT lawsuit and its subsequent settlement, Linux users have been wondering what Microsoft's intentions with regard to the presence of VFAT functionality in the Linux kernel might be.

The Linux Foundation continues to maintain that Microsoft's VFAT patents are not valid. And of course the United States is one of only a small handful of countries that even recognise software patents.

However, according to CNet, the Linux community has recently been taking steps to work around the entire VFAT patent issue. But what is the nature of the VFAT patents, and how can Linux bypass them?

The historical point of VFAT

MS-DOS and older versions of Windows (versions prior to 95) could only read filenames in the short 8.3 format. VFAT was designed to accommodate the use of older software programs written for MS-DOS and pre-95 versions of Windows by allowing any given file to have a long filename (technically up to 256 characters) and a short 8.3 one. Newer programs would see the long version and display that to users. Older programs would see the short version.

This kludge was not perfect. If a user created a DOC file named, say, "Letter to the Prime Minister.DOC" in Word 95 or higher, the VFAT functionality would automatically and invisibly create and store a short version of the name (here, it might be something like "LETTE~01.DOC") so that older software that could only recognise the 8.3 format would be able to see and open the file. However, if that file were then opened in, say, Word 6.0 (as "LETTE~01.DOC"), edited and then saved again, the long version of the filename would be lost.

Imperfect as this functionality may have been, it did allow for a more or less smooth transition to long filenames. Nowadays, however, almost no one has a need for this functionality. Yet it is built into the standard FAT/FAT32 file format used on almost every USB and flash device out there.

Because Microsoft claims patents regarding this VFAT functionality, it was able to sue TomTom and extract from that company concessions in an out-of-court settlement. Linux users more generally--and particularly corporate Linux users--have since worried that Microsoft may be wanting to sue them for the simple reason that, while Linux does not contain Microsoft code (and so does not violate any copyrights), it does contain the VFAT functionality (and so could be taken as stepping on Microsoft's patent).

Fixing the problem

The TomTom case has had Linux coders working to alter the way Linux works to avoid any potential claims by Microsoft. The current favourite, now being vetted by Linux patent lawyers and considered for inclusion in the kernel, involves filling the short version of a given file name with dummy data, rendering useless the VFAT functionality claimed in the patents.

The key to the patents involves allowing a long filename to be seen and edited and saved by software that can only see its short version. The Linux fix keeps the framework of long and short filenames intact, but breaks that functionality--effectively, you can use the long filename without any problem, but the short filename would become unreadable by older software programs that demanded it.

This loss of functionality, however, would not matter to the vast majority of people out there, as almost no one these days uses pre-Windows 95 operating systems.

Whither VFAT?

At the end of the day, Microsoft may never seek to attack Linux directly over its VFAT patents. Those patents, as noted, only apply in the United States and a small number of other countries. They do not apply, for instance, in the EU and most of Asia.

Besides, its VFAT patents were already once ruled invalid in the United States. Although Microsoft was later able to get them reinstated, it is unlikely they would want to risk a second invalidation.

Finally, the use of the FAT/FAT32 format with VFAT functionality in billions of devices across the world actually limits Microsoft's ability to act on their patents. If they seek to enforce those patents more broadly, they may find themselves in more trouble for abusing their monopoly.

Report a problem with article
Previous Story

Confirmed: Windows 7 set to RTM on July 13

Next Story

IE's market share falls below 60%, Firefox surpasses 30%

34 Comments

Commenting is disabled on this article.

Airlink said,
Or Linux-NTFS. Why is anyone still using FAT anyways?

NTFS is not ideal for external flash drives due to the way it writes data. FAT is still much more ideal but really we need exFAT support on Mac systems and then we can all use that instead.

shinji257 said,
NTFS is not ideal for external flash drives due to the way it writes data. FAT is still much more ideal but really we need exFAT support on Mac systems and then we can all use that instead.


NTFS is 'heavier' but doesn't enforce a policy that that is not ideal. See note at the bottom.

As for performance or space usage, NTFS is not longer easily edged out by basic FAT or lower end FS technologies. With Vista and Win7 a lot of work went into the optimizational writes of NTFS as well as 'knowing' the type of device media - something that benefits from the Object I/O model in NT.

Initial MFT use is larger with NTFS, but overall savings once the drive is filled offsets the MFT size almost entirely.

Performance is also on par with FAT and on regular Hard Drives will exceed FAT performance when the system is dealing with fragmentation reads and a few other tricks that steam from the NTFS MFT structure.

This has been a big debate in the exFAT and NTFS debates in the Microsoft circles, as NTFS can be just as efficient as FAT, but with NTFS you can 'add' on features on a file by file basis, so you can use compression, rock tight encyrption, and even take advantage of copy on write (Shadow Copies).

NTFS also offers better data integrity with journalling and full volume journalling features introduced in Vista.

It is hard to corrupt an NTFS volume, even on removeable media when you rip it out while doing multiple writes, as the journal should catch any inconsistencies and NTFS will rolback, ignore, or repair any failed writes even to 'system' areas of the volume.


With that said, there is a problem with NTFS on portable devices. And it goes back to Microsoft and Licensing. FAT was easy as most people thought they could get away with using it because of its history, with the FAT32 naming being the only stumbling block as the Tom Tom utility found out.

Microsoft owns NTFS, FAT32(vFAT), and exFAT, and requires basic licensing for all of them - even though Apple and Linux pretend their code is legal to read these volumes, technically it isn't unless they have signed an agreement with Microsoft.

Consumer devices are not going to want to pay a Microsoft fee, so you are back at the begining of the argument.

Microsoft needs to step up and pick either NTFS or exFAT as the consumer device standard. Since exFAT was a WinCE technology, and with recent NTFS changes, there is really no reason to support exFAT when it is not backward compatible.

Hence NTFS would be the best choice for consumer device media, and would also allow for secure distribution of software on media, as the vendor could use NTFS encryption features.

Nothing is going to work in the consumer world unless it has inherent Windows support, so it has to be Microsoft to standardize this and let go of the fees.


(*Note that Vista and Win7 that NTFS notices if the drive is Flash based, and uses the Microsoft random writing system so that it doesn't degrade the device. This technology was born from the ReadyBoost technologies to have a high use Flash device and not wear its 'writes'.

NTFS and FAT under Vista and Win7 enjoy this technology, as it occurs at the NT Ojbect Model level, and is a benefit of a having a true object based OS I/O model as opposed to a generic I/O model like UNIX uses.)

Dead_Monkey said,
FAT is quickly becoming obsolete just based on the minimum size of our storage capacity these days. Seems like a pointless battle.


Exactly...

Question why does the gpl 3 specifically go against licensing patents? What if say microsoft invents say a warp drive. Something that is needed. Linux would rather have people steal what makes it tick and put it into linux instead of doing the right thing?

Ps I cant find no real neutral articles on this case.

majortom1981 said,
Question why does the gpl 3 specifically go against licensing patents?

The goal is to create software that can be freely reproduced and modified by everyone, everywhere, forever. If you license MPEG-4 and create a player/encoder you have the rights to distribute the code but can't necessarily grant other's the right to use your licensed tech. If you can't distribute that license then the code is still partially "owned" by a third party and by revoking your license (or not granting new ones) they can effectively kill the project.


What if say microsoft invents say a warp drive. Linux would rather have people steal what makes it tick and put it into linux instead of doing the right thing?

What are you trying to say?

The FSF would argue "We want software to be free: free to edit, free to give away, free to do anything but lock up from others. If you can't have that then it's better to do without." If there is some sort of feature that can only be implemented by using patented technology (like say H.264 decoding) then we'll either create our own version that isn't patent encumbered (ogg/vorbis) or do without (vfat without backward compatible short names).

evn. said,
The goal is to create software that can be freely reproduced and modified by everyone, everywhere, forever. If you license MPEG-4 and create a player/encoder you have the rights to distribute the code but can't necessarily grant other's the right to use your licensed tech. If you can't distribute that license then the code is still partially "owned" by a third party and by revoking your license (or not granting new ones) they can effectively kill the project.



What are you trying to say?

The FSF would argue "We want software to be free: free to edit, free to give away, free to do anything but lock up from others. If you can't have that then it's better to do without." If there is some sort of feature that can only be implemented by using patented technology (like say H.264 decoding) then we'll either create our own version that isn't patent encumbered (ogg/vorbis) or do without (vfat without backward compatible short names).


So then if osx and windows have something thats a must have and advertised as such but needs a license linux wont implement it because its not free?

With the way the gpl is makes it seem like there could be a lot of stuff in linux that actually does violate some patents. Considering you have all these people trying to put stuff that has patents into a free os . A lot fo that would eventually go against patents.

So technically no software should be encoding things to mp3 in linux right since you technically need a license on mp3 to encode and that violates the gpl

majortom1981 said,
So then if osx and windows have something thats a must have and advertised as such but needs a license linux wont implement it because its not free?

If they can't possibly work around patent issues (which would be unsual) then yes: they would rather do without. The free software foundation's goals are "Free software" and they consider patents that can't be licensed royalty free contrary to that goal.

With the way the gpl is makes it seem like there could be a lot of stuff in linux that actually does violate some patents. Considering you have all these people trying to put stuff that has patents into a free os . A lot fo that would eventually go against patents.

The linux kernel is a GPL2 project so the GPL3 issues aren't important here.

There's nothing to stop the end-user from mixing and matching licenses on their own system. You could, for example, take Microsoft Office source code and work it into the Linux source to produce your own piece of software. The GPL would forbid you from distributing those changes but you're free to use it if that's what makes you happy.


So technically no software should be encoding things to mp3 in linux right since you technically need a license on mp3 to encode and that violates the gpl

Most (all?) of the modern linux distributions don't ship with an MP3 encoder or decorder. They force the user to download one after the fact and install it after—I'm making an assumption here—making the user read the license and realize that it's "not free" software.

You can't make a GPL3 MP3 decoder. GPL2 doesn't have the patent requirements that GPL3 does.

Wasnt the problem one of the 3rd party utilities that tom tom used that was the problem not linux itself? Is the lniux community spreading false info to badmouth microsoft?

I think people are getting way too much excited with this patent thing. I personally don't know the settlement details between Microsoft and TomTom, but having been involved with patents and the business behind it, I have my doubts about "Microsoft going after Linux". It's a very daunting task that would probably irritate anti-monopolistic agencies (heck, even the government uses Linux so imagine the aftermath if Microsoft was to go after Linux...)

Corporations usually use patents to protect themselves more than to attack others. For instance, whose to say that TomTom did not itself have a patent that Microsoft was interested in or was already infringing? Maybe there were early negotiations to resolve the matter, maybe there weren't. Maybe Microsoft went too aggressively by suing instead of negotiating first. We don't know that (I personally don't).

And yes, I must post the usual disclaimer that I am no Microsoft fan boy, just like I am not Linux' nor Apple's. This old tirade is getting very old. I work with all OSes in my day to day life and I can see first hand that each OS has its strengths and weaknesses. And despite what the Apple fan boys might think, Microsoft is no more evil than Apple (you just have to look at Apple's history to see their own mishaps).

Some of you Linux guys(I am one) amaze me with your comments. Especially daPheonix and his off the wall "Apple-esque fanboy comment. Patents are meant to stifle competition its meant to protect that you have created. Yes MS can create things and "not steal" the ideas from someone else. Get off the Linux high horse and come down to reality. If they did indeed use a code or a method which another company already claimed the rights too then they need to change it, ask for permission or move on to something new. Not circumvent the method with another method leaving the original method in place but rendering it useless to bypass a lawsuite or court ordered injunction.

So its okay to sue MS but its not okay for MS to sue anyone else?

The EU is a joke and I wish that MS would just stop selling over there and let them figure out where it went wrong. Poor cry babies asking for EU help to sue the big bad wolf. All the while Apple just seems to bundle everything in their re-packaged updates and call it Snow Leopard. DUH why do you think its so cheap to upgrade? NOT A NEW OS APPLE FAN BOYS.

Linux just needs to find another solution or work with this one and ask MS for permission. Or do as they are doing and "work around" it legally.

daPhoenix perhaps you should read up on "Lemelson, Jerome".. It's not always about the patent...

Fear is the path to the dark side: fear leads to anger, anger leads to hate, hate leads to suffering.

Yes, I sense much anger in you.

VFAT patent has been considered invalid in EU not because MS is evil or is american but because is obvious and is under prior art.

Also VFAT patent has ben considered not once but twice invalid in Us of A but later a jury considered the opposite.

Magallanes said,
VFAT patent has been considered invalid in EU not because MS is evil or is american but because is obvious and is under prior art.

Also VFAT patent has ben considered not once but twice invalid in Us of A but later a jury considered the opposite.


But its been found valid in the US where this lawsuit took place.

Minimoose said,
I really love it when people say that Microsoft should stop selling in the EU, it really makes me laugh a lot :)

I'd lol if they did, but its never gonna happen because they'll make less munnies from sales!

No, the title is "Microsoft patents: Linux to become VFAT-free?".

It simply shows us a dangerous precedence as to what Microsoft has in mind for the future - if they cannot compete in price or quality, they will use their patent portfolio to stifle competition - not that it's surprising in the least.

Linux guys are simply implementing a feature that makes interoperability "legal" in the sense that it doesn't offend any patents, whether they have any validity or not.

FoxieFoxie said,
So the title is: Linux finds a loop hoel to scam microsoft

I don't think Microsoft is being scammed by anyone. While Microsoft may own VFAT petents, they do not own the users' data, even if the user uses Microsoft products to create that data, and they do not have a right to prevent users accessing their own data with software from another vendor.

I fail to see how MS is being "scammed", MS have a patent on a certain technique (stupid yes, but still) so the way around the patent is to not infringe it, i.e. write code that doesn't use that technique.

So if we're to say MS is being scammed because people aren't infringing their patents, then where does it stop?

daPhoenix said,
No, the title is "Microsoft patents: Linux to become VFAT-free?".

It simply shows us a dangerous precedence as to what Microsoft has in mind for the future - if they cannot compete in price or quality, they will use their patent portfolio to stifle competition - not that it's surprising in the least.

Linux guys are simply implementing a feature that makes interoperability "legal" in the sense that it doesn't offend any patents, whether they have any validity or not.


What about the gpl3 then that basically states if you do not steal a technology you cannot use linux software. The part where it states if you license a patent from a company you cannot use linux software.

Why cant tom tom license a patent that has been proven valid by court cases? Gpl 3 states they cant.

majortom1981 said,
What about the gpl3 then that basically states if you do not steal a technology you cannot use linux software. The part where it states if you license a patent from a company you cannot use linux software.

Why cant tom tom license a patent that has been proven valid by court cases? Gpl 3 states they cant.


Holy crap!

Do you even know what the GPL states? Because your statement is so out in left field that I have no idea where to start any reasoned discussion with you.

Let me just start by pointing you to the GPL v3
http://www.gnu.org/copyleft/gpl.html

and, if reading and interpreting a legal document isn't your bag, here is a easier to digest guide
http://www.gnu.org/licenses/quick-guide-gplv3.html

FoxieFoxie said,
So the title is: Linux finds a loop hoel to scam microsoft

more like "found about 10 years ago and may get in trouble now". The operative word being MAY. And seeing as this feature only matters for old versions of Windows or Dosbox this is hardly a major scam, just an infringement of a patent on an old technology.