Long file names in FAT

Was this the group where recently there was a thread about long file name implementation?

If so then take a look at

formatting link
and search for "Common Name Space for Long and Short Filenames" in
formatting link

--
Regards,
Richard.
 Click to see the full signature
Reply to
FreeRTOS.org
Loading thread data ...

Yes, it was.

My main concern about this is that Microsoft may want to establish that the patent is valid (at least over there in the US). If they do, then what is currently a secondary issue, the long filenames, could become a primary issue in the next lawsuit which could be against against Linux itself.

It will be interesting to see if Microsoft intend to target purely European companies.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
Reply to
Simon Clubley

I don't think you have any problem. IIRC the patent has to do with the generation of 8.3 filenames from arbitrary long filenames, not with the use of filenames. After all, files have been given names for some time.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
 Click to see the full signature
Reply to
CBFalconer

Yes, I know (but I didn't express that very well above). The problem here is that Linux includes the various FAT filesystem drivers, (which include this long filename to 8.3 mapping when required), as standard.

I just can't believe Microsoft got a patent over something as simple as that. It's the kind of thing which should belong in a computer science book in the chapter on examples of implementing backwards compatibility and not in a patent.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
Reply to
Simon Clubley

Does umsdos count as prior art?

--
Mit freundlichen Grüßen

Dipl.-Ing. Frank-Christian Krügel
Reply to
Frank-Christian Krügel

I'm not a patent lawyer, but as far as I read the patents, they indeed only apply when you actually have a common namespace of 8.3 names and long names. Linux does that, that is, when you have a file "Microsoft", you can also access it as "micros~1". My implementation doesn't. It only fills in the "short file name" fields to not leave them empty (because it seems Windows refuses to access files with empty or non-unique short names).

5,758,352 requires "creating a first directory entry for a file wherein the first directory holds a short filename for the file". It's never used as a file name, it's a random number.

Likewise, 5,579,517 requires "accessing the first directory entry with the operating system". Again, the "first directory entry" is never accessed with the operating system, all references go to the long names.

That aside, the Rock Ridge Interchange Protocol *might* serve as prior art. RRIP does to ISO9660 what VFAT does to FAT. There is a document describing RRIP, dated three months after the MS patents. Because its version is "1.10", there could very well be versions dated before.

IANAL. YMMV.

Stefan

Reply to
Stefan Reuther

I had totally forgotten about umsdos.

The honest answer is that I don't know. I've just read up about how umsdos saves Linux type file attributes, including the long filename. See:

formatting link

It appears that umsdos stores the additional attributes, including the long filename, in a file called --linux-.--- in the same directory as the files. I don't know if that's close enough to the long filename to 8.3 mapping patent to be able to invalidate it.

OTOH, it would be absolutely _beautiful_ if this Microsoft patent was invalidated by Linux prior art. :-)

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
Reply to
Simon Clubley

The additional data is clearly *not* stored in the same namespace as the long file names - only the --linux-.--- file is in the same namespace. Close, but no cigar.

Reply to
Clifford Heath

I don't think that Linux predates MS LFN besides CPM had sub directories and other things before MS-DOS..... :-)

I think that a lot of things in MS-DOS that are still in MS OS's were in CPM or other OS first... so it could get quite silly. Especially in the current climate going to court over SW patents that AFAIK are only enforceable in the US seems suicidal.

Your other major problem is most people who worked on Linux were programmers who are employed (or were at the time) by some one and most contracts of employment give the employer rights on anything you produce. This is not often enforced but was recently between the makers of Barbie and Bratz.

So it could be argued (at great expense in court) that even if some things in Linux predate MS the things in Linux or other FOSS actually belong to the programmers employers....

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

Are you freebasing? MS's first release of an LFN-supporting OS was in

1993 IIRC. The Linux kernel dates to 1991, but in fact the filesystems supported by the oldest versions are older.

If they have been disclosed to the public - which they have - they are unpatentable, so the worst that could possibly happen is a /copyright/ suit, not a /patent/ suit. The copyright issue could be worked around in weeks (days?) by a cleanroom type reimplementation of the code in question.

Reply to
zwsdotcom

Don't see how as SW patents are only valid in the US? Also targeting only European companies would be difficult to defend in a court of law and bring other reprocussions

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

Sorry, forgot that Linux is that old and LFN is that new,.

True but the IP can be said to belong to the employer. It is rarely invoked though as an example I think a designer of the Bratz Dolls was previously working at the place that does Barbie.

The Barbie people (Mattel?) I think successfully sued that he used information from Barbie or did the Bratz designs (in his own time) whilst still at Barbie

SO programmers working on FOSS in their own time could find their employers claim the IP..... Note morals and being right or wrong have no place in a legal argument :-(

But the situation whilst the lawyers and courts wrangle over it and people stop using it until it is decided could be years.....

This sort of tactic has been used before. Companies will not use sw or technology that is (or is about to be) the target of a legal dispute. So if a large company threatens patent/IP injunctions and court cases many customers of the threatened company get nervous and in some cases it is enough to if not kill off certainly incapacitate a competitor.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

Just like people stopped making, selling, buying or using Linux while the SCO court case was going on... That was almost certainly Microsoft's aim when they sponsored the case.

You could probably justify claiming that a few potential Linux users were discouraged from taking a "risk" with Linux during the legal battles, but the effect was pretty minor. I don't expect you'll find many companies that drop the use of non-MS-licensed FAT just because of this current case.

I think there are several potential outcomes, which may all be mixed. TomTom might give up and pay the license - this will hopefully increase pressure to get some changes to the IP laws (in the USA in particular), and perhaps also the pressure to change the way such cases are handled. If the case goes all the way to court, I don't expect the FAT-related patents to hold out (and probably not the "computer in a car" patents either), which would be a good thing for the rest of the world. It might also encourage companies to fight rather than pay protection money in the future. I also expect that patches will be made to make MS-style LFN's on FAT optional in Linux - if MS tries to push for more general licensing of these "patents", Linux users can simply disable the use of MS-compatible LFN's.

Reply to
David Brown

It affected some but I am not sure it did much harm.

It will have an effect. More so in the US than elsewhere.

Hopefully

Why?

Yes at this point in time as it would remove uncertainty thought most of the world ignores it anyway. Also SW patents are not enforceable outside the US AFAIK.

It depends....

It's not just linux. Actually linux users in most of the world can simply ignore MS as, AFAIK SW patents only apply in the US?

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

To be patentable, an idea must be new, useful, and not obvious to experts in that field. MS's LFN FAT patents boil down to using additional directory entries to hold the long file name alongside the original 8.3 filename. Using long file names in FAT is not new - UMSDOS and OS/2 both did it before MS Win95 (AFAIK). OS/2 certainly let MSDOS and Win3.x programs run natively and access files with LFNs on FAT drives, though the implementation details differed from those of MS Win9x. Secondly, it's hardly innovative - it's a fairly obvious possible method for getting LFN support in FAT. So the FAT patents will fail on two of the three required criteria. MS will, of course, claim otherwise if the patents are contested, but I don't expect them to stand up to serious and fair judgement.

There have been a few patents awarding in Europe that are basically software patents, and plenty of discussions about how valid software patents may or may not be in the EU. The general point is that pure software patents are not valid at the moment, but there are plenty of grey areas. It is still very far from the situation in the USA, of course.

Yes - in particular, it depends on how much it costs TomTom, and whether they can counter-sue or in some way recoup their costs if they win against MS. Even if they don't get their money back, standing against a corporate bully and winning is great PR.

Unfortunately for that argument, very few large companies are unaffected by the rules in the US. Either their development is at least partially in the USA, or their sales are partially in the USA. Even for the average Linux end-user outside the USA, the American laws are a pain. Big American-based Linux distributions already have limits on putting things like libdecss (for reading DVDs) in their main distributions or making them easily installable - often they have to have FAQs pointing users to sites hosted in free countries to download such software. If things go really badly with the FAT LFN patents, we might see the same thing for LFN FAT access in Linux.

Reply to
David Brown

... snip ...

While CP/M preceded MSDOS in many things, a nested set of sub-directories is NOT one of them. And this has nothing to do with MSDOS' alleged patents.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
 Click to see the full signature
Reply to
CBFalconer

DOS 1.0 didn't have subdirectories, either. I remember. I developed applications for national sale at that time, under DOS 1.0 in IBM's assembler. Directories were added in v2.0.

I've no comment on this.

Jon

Reply to
Jon Kirwan

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.