Micro$oft to license FAT

royalties

The patents were granted in 1996 or so, and their first implementation was

1976, how can they retro-actively seek compensation for this when they allowed businesses to be built on this for 20 years? Also, the patents all seem related to long filename handling. If the device doesn't use/support long filenames (and most simple devices don't), why would you need a license?

Rob

I'm not a lawer and hence probably should shut up, but the patent numbers listed on that M$ page seem all to deal with long & short filenames hence FAT32. So, while it's probably not save to asume that, eventually this patent thingy is limitted to use FAT32. Does anyone knows better?

Really highly motivating this page I must say...

Are there *SIMPLE* filesystems around which are public domain?

Markus

In other words, it appears M$ wants to LZW-ize FAT.

Long filename support has nothing particular to do with FAT32. You can have LFNs on FAT16 and even FAT12 filesystems, just as easily (e.g. on a plain 1.44 MB floppy disk).

--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Yes, I did not expressed myself clearly enough. Based on the patent numbers I would think that not FAT iteself is patented, but the long/short filename thingy. Nothing more than sheer hope though...

Markus

You probably know this, but for the sake of others, patents can be looked up at

.

To discuss patents with "the experts", try the newsgroup misc.int-property .

Mike

LFNs are supported on FAT12 and FAT16 as well as FAT32.

Yes, but the four patents mentioned specifically relate to LFNs, which are an extension to FAT. By implication, the "pending" stuff they talk about is newer than LFNs, so probably of no interest to embedded developers unless you're making a very computer-like device. In the enormous majority of cases, the only reason people use FAT in embedded systems is for data interchangeability with PCs. So a lowest-common-denominator filesystem is perfectly acceptable for us.

Anyhow, the tone of the document is not one of MS actively seeking to close down extant third-party FAT implementations, but more like MS trying to keep future enhancements to the FAT format closed.

(Side note: The XBox uses FAT32, not NTFS - and it's FAT32 with some extensions for slightly more robustness than regular FAT).

I take the document that they (MS) are trying to follow the success of licensing of MP3 -- make the specification available, and years later ask for fees because there are so many implementations.

[I am not a lawer, opinions are strictly mine :]

Vadim

A common misconception is that FAT32 and long filenames are in some way interrelated - not so.The long filename trick works just fine on FAT12 and FAT16 too,FAT12 being used on floppies for example. It's also quite possible to have a short-filenames-only FAT32 volume.

A skim read would suggest that it's mainly flash based consumer gadgets that are being clamped down,with no mention at all of other platform desktop systems like FreeBSD,nor embedded uses like oscilloscopes etc... Sprow.

On Thu, 04 Dec 2003 17:00:01 +0100, Markus Zingg wrote in comp.arch.embedded:

LFN's for FAT were introduced with Windows 95 first edition, which did not have FAT32 but did support VFAT16. FAT32 support was introduced with OSR2 of Windows 95.

And as others have pointed out, LFNs can be used on FAT16 and even FAT12 floppy disks.

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html
comp.lang.c++ http://www.parashift.com/c++-faq-lite/alt.comp.lang.learn.c-c++ ftp://snurse-l.org/pub/acllc-c++/faq

First implementation of MS-DOS was ca. 1980. The IBM PC was announced in summer of 1981.

There were damned few microcomputer disk operating systems around in 1976, mainly because there were damned few microcomputers in 1976. About all you had at that point were Apple Disk BASIC, ISIS-II on the Intel MDS, and CP/M. I don't remember when UCSD PASCAL came out, or when the TRS-80 Expansion Interface and Disk Drives came out, but it WASN'T in 1976.

I don't remember when software patents actually became legal, but I think it was a few years after this all went down.

Microsoft had not even HEARD of a disk operating system until around 1980, when they acquired QDOS from Seattle Computer Products. (It is not completely clear how Seattle Computer Products got QDOS. There is a story floating around that involves a room full of Microsoft and IBM lawyers, Gary Kildall, a PC running MS-DOS, and a Digital Research (Kildall's company) copyright notice in an Easter Egg.

If I had an unlimited legal budget, I'd be sorely tempted to ask Microsoft to back up that claim of FAT first being implemented in 1976.

To you and the others who replied, - I know. I just did not expressed myself precisely enough - sorry for that. The page - and that's my main concern - while only listing LFN related patent numbers claims to license FAT in it's whole and apart from other posters opinions the "tone" on the page to me clearly indicates that M$ want's 0.25¢ per device and from the media manufacturers 0.25¢ per preformatted media.

While the page clearly lists certain mainly protable devices, it does not at all exclude any other useage as being free or such but states that a special arangement must be made. Bearing in mind how M$ usually behaves I have a feeling that this does not mean that they don't care. It just means that the fees would be subject of the agreement and with lower quanties I'm afraid that the fees would be higher - not lower.

I also wonder what the impact of this will be for the linux comunity where FAT implementations are used also. Do we see a FAT free linux in the future? :-)

Well, let's see how things develop.

What I really wonder is if there is a *simple* hirarchical filesystem available that is in the public domain. I'm just curious.

Markus

... snip ...

A fundamental principle of US patent law (IANAL) (AIUI) is that all applications must be made within one year of first publication or marketing. That's application, not issuance.

At the same time the FAT system was built into Billies early Basic, which functioned without an OS, again AIUI.

--
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
   Available for consulting/temporary embedded and systems.
     USE worldnet address!

Markus,

If I read their text it appears their license is a license to use *their implementation* of FAT, which includes LFN and possibly other stuff. It doesn't seem to cover other implementations of FAT per se. It's a bit like this: You can probably come up with your own way of implementing malloc() and then try to license it to others, but that doesn't mean others are

*required* to license your implementation.

I have requested additional information from Microsoft to explain what is and isn't covered by the license.

Rob

Hi Rob

Oh, let's hope that this really is the case. Your statement along wiht what Chuck F said makes me feel a lot better!

Thank you for the effort. Would you mind posting the result of this here?

Markus

This seems to be twofold. MP3 case shows that you can have your very own implementation of it (hardware, software, anyware), but are still required to pay Thomson Multimedia for the privilege of using the algorithms. Patented algorithms, btw. Which is not the case with FAT -- the references to "methods and apparatus" to deal with long file names have nothing to do with plain vanilla FAT. Dunno how can they get it patented *now* .

All thoughts that can get into your mind while you are reading this personal opinion are subjects of intellectual property. Unlicensed cogitation of aforementioned thoughts is prohibited :-)

Vadim

Maybe. Anyway, it's not a very serious issue - $0.25 won't break me, but I will fight it if necessary simply out of a desire not to pay Microsoft anything. FAT12 at least is closely based on CP/M on-disk format anyway, isn't it?

(I guess since nobody wants Windows CE, MS has to find some other way to get a revenue stream out of the embedded market).

I believe the first "pure software" patent was granted to Tektronix in

1987 for a basic graphics anti-aliasing algorithm, which oddly enough I'd learned in a college course in 1972. Prior to then the only software patents I'm aware of were so-called "system patents", in which hardware was complemented by specialized software.

Jim McGinnis

Not really. I'm reaching back 20 years in my memory, so I may be wrong. Seems to me that CP/M did the allocation of clusters/sectors right at the directory entry, not in a FAT. This severely limited the file size you could have and still have a reasonably length directory. A later kluge was "extents", where once a directory entry was full of allocations, it chained to another one and that was used.