MS patent on FAT a problem for ros?

[tomleem]

http://www.pcpro.co.uk/news/82088/micro ... atent.html
"On its website Microsoft offers to licence the FAT IP. Pricing for this license is $0.25 per unit with a cap on total royalties of $250,000 per manufacturer. Among those who might find themselves on the receiving end of a royalty demand from Microsoft are likely to be Linux vendors and memory stick manufacturers." and ROS users?

[Baldomero]

No FAT? No posibility to read floppy from anyone on the world using *any* MS Win version or dos? neither compact memory...

[oiaohm]

PLEASE READ

Linux is untouchable by this fat problem. All vendors will refer Microsoft back to the kernel.org.

Its not like MP3 where the creator of the interface code had licence changed.

Number One Microsoft and IBM have a shared agreement between them regarding filesystems. Fat file systems is included in the aggreement.

Number Two IBM developers work on the linux kernel with IBM's blessing.

SCO case is over this very problem because SCO is saying that IBM inserted their code into Linux. Linux is protected by IBM IP agreements.

Problem is we are not. Microsoft demarding payment for Linux will not work.

Thank You Microsoft we will see more Linux based devices for sure most of the other light end os's don't have IBM's Blessing.

[temarez]

No FAT? No posibility to read floppy from anyone on the world using *any* MS Win version or dos? neither compact memory...

This patent is about "Common name space for long and short filenames". So it doesn't makes usage of FAT illegal as I understood. But usage of both long and short name for each file in one volume without Microsoft permission is illegal

[Mrkaras]

Can there be a special type of fat that doesn't store the short file names but creates them as needed. Short file names ore only usefull for dos anyway and fat 16 would have only short names therefor also not a problem.

[temarez]

Can there be a special type of fat that doesn't store the short file names but creates them as needed. Short file names ore only usefull for dos anyway and fat 16 would have only short names therefor also not a problem.

This is one of suggestions. Let's hope besides that it is not end because FAT patent fight not over yet

[Matthias]

There is absolutely no reason to use HPFS. Unlike FAT and NTFS it can't be read or written by any Windows NT Version > 3.51, it doesn't support large Volumes and every Open Source Driver for HPFS i know of lacks write support. NTFS would be a much better choice, because it can be read by Windows, and there is a free driver (in the Linux Kernel) with limited writing support (e. g. you can edit files, but you can't create new files or delete old ones).
ext2 lacks journaling, but you could later upgrade to ext3, which is backward compatible and would allow users a soft transition, once the ext3 implementation is finished. Also you wouldn't have to worry about any patent stuff. ReiserFS might be an even better choice, because it's faster. A GPLed IFS driver is available here (beta stage).
@Patchworks: i get a 404 error when trying to access the OpenBFS Web Site. Still it might be an interesting candidate, after all i've read about it.

[menn]

i remember reading the LONG discussion on the mailing list about this. of note:

1. as they said in the mailing list any normal patent on FAT would have expired now or not be valid otherwise. ie FAT patent misleading by omission: 8.3 filename fat NOT An Issue. ianal, but as i see it, this issue is chiefly with long filenames the way fat32 does long filenames, and fat32 etc.

2. ntfs is not the way to go. it's propriatary, not entirely documented, and way more of a patent issue than fat is, oy. i'm not against ntfs support or any other support it just isn't a solution to this problem.

3. i look forward to seeing compatibility with ext2 and ext3, it's about time. but part of windows compatibility really is fat. i mean, cmon...

4. we should be able to get away with anything the dosemu project gets away with.

5. lastly, consider that the entire industry uses fat without paying royalties. it's too damned late for MS to do anything about it- oh, i could be wrong. but with *everyone* having to pay extra to make the devices they already sell have the same functionality as before, they're going to fight this one for us, just like ibm is on foss's side against sco. and they're going to win, i believe. so let's stop jumping the gun, can we? there is going to be a version of reactos with fat support, thank you europe, i love you dearly- and the only hypothetical issue is going to be whether or not american businesses have to use a crippled-8.3-only-fat version of ros. beyond that a fat patent is unenforceable, you can't stop most users from downloading the euro version, and besides- 8.3 + 4dos-type driver would let ros simulate long filename handling in a way that Wouldn't violate the fat32 patent on the offchance it became an issue. ok, done with my rant. now relax.

[Matthias]

2. ntfs is not the way to go. it's propriatary, not entirely documented, and way more of a patent issue than fat is, oy. i'm not against ntfs support or any other support it just isn't a solution to this problem.

I believe that NTFS support is necessary sooner or to allow the users a soft transition from Windows to ReactOS (once ReactOS is actually productively usable)

[temarez]

HPFS ... doesn't support large Volumes

Wikipedia article says that max volume size is 2 TiB. Max file size is 2 GB but this is not limit for HPFS. Max size in this FS depends on sizes of field storing file size and file descriptor (4 bytes in OS/2). So it is possible to overcome this "limit".

every Open Source Driver for HPFS i know of lacks write support.

I have seen many open-source HPFS drivers with full write support. Linux kernel driver claims himself to support HPFS in write mode. One of old ones even can store owner and acess permissions information in extended attributes.

Besides HPFS is closer to Win then ext2/ext3, Reiser and others. The question is can Microsoft request something for usage of HPFS (in view of it was developed by IBM also). Who knows current status of this FS?

So, maybe JFS is answer then?

NTFS is the way to go

There is no good write support for NTFS (AFAIK Captive is based on MS driver so this is not good decision). Again, NTFS was develolped by Microsoft. Don't you think that they'll take out a patent for it when normal full-featured open-source NTFS driver will be available?

So IMHO NTFS is the good way though it must not be the only way.

[menn]

i hope it isnt necessary, because i'm not 100% sure how possible it will ever be.

but like i said, i'm all for support of any and all filesystems. i just think fat is important, not worth worrying about (quote the mailing list: "...in [about] ten years..." heh.) and that you can't fix an MS patent issue by switching from fat to ntfs, because ntfs is an even greater patent issue if fat is one at all. i do agree... nfts support is a desirable feature that would help- immensely with transition. i think everyone sees how useful ntfs support will be if it is possible to implement.

[oiaohm]

Can there be a special type of fat that doesn't store the short file names but creates them as needed. Short file names ore only usefull for dos anyway and fat 16 would have only short names therefor also not a problem.

Sorry No. The short form is the base of the filesystem. The long ones are fake.

4. we should be able to get away with anything the dosemu project gets away with.

Sory to say dosemu does not really do fat at all. It freedos that does fat.

Freedos is older than us and most likely on it own. Only hope is to find prior art that is close enough to get the patent thown out or it does not hold because of entrapment since they should of acted sonner.

dosemu can run MS-Dos it really does not give a stuff.

I am thinking about a new version of fat(someone a name.)

Almost everything the same as fat bar the directory. This will allow us to create some security.

[fat table entry][32bit attribs][first 8 bytes of file][size][chksum][varnam len]

First 8 bytes so we are not tail dependant for file id. Get us out of trouble Minor tweeks change the directory layout and leave everything else the same.

Get it worked out and get pic and avr code interfaces build.

Break away from microsoft if they want to be idiots can solve alot of our problems.

[menn]

Sory to say dosemu does not really do fat at all. It freedos that does fat.

details! dosemu relies on freedos in enough ways to be for all intents of purposes. you say it could rely on msdos - i think not, absolutely not, even if on a *purely technical* level it's true.

Freedos is older than us and most likely on it own. Only hope is to find prior art that is close enough to get the patent thown out or it does not hold because of entrapment since they should of acted sonner.

amen- and i tried to find freedos people worried about this like ros community worries about this and i can't find it. i think this whole thing is an overreaction. MS doesn't want to kill support of fat they want to leech a little more $ off of hardware manufacturers... i'm betting.

dosemu can run MS-Dos it really does not give a stuff.

yeah think about it we're talking about a linux project. you really think they'll say "oh screw it ust use ms dos" i think they'd just bolster support of freedos. foss is the ties that bind. ros and linux and dosemu are all gpl at that.

I am thinking about a new version of fat(someone a name.)

sure, many fat compatible but patent noninfringing ideas will come out of this just as 4dos came out of fat- look at the NEW iso 96?? cd format: iso9660 did not have long names... MS joliet added that. new iso has long filenames. there is a free ticket in there somewhere.

Break away from microsoft if they want to be idiots can solve alot of our problems.

i'm sure they love to be but there is a limit to everything even a limit to stupid. i know some days it doesn't seem that way, but there is a limit to any day.

fat is important to many people using ros and many devices compatible with win and until it becomes a real issue it would be foolish to abandon it.

if MS wants to fight ros someday i don't think they will make issue of fat anyway, they will complain about something else. like ntfs. you can't think for a second they haven't got a zillion patents on that, of which they will enforce about 4. MS is where the $ is. the threat ros poses MS is not a threat ros support of fat poses. MS only attacks something legally when convenient and profitable. no $ here. can't squeeze blood from a turnip. they will hit the hardware people.
SCO will have a fit about this, and MS will support them from a curtain, and userfriendly will point and laugh, and reason will actually win this one. sco sets the precident: fight foss and someday you'll be laughed out of court. it may take hard work and eff but no worries. must stop ranting. thanks for patience.

[tomleem]

"Sources indicate that recent FAT patent win news for MS will not be the end of the story. "
http://www.eweek.com/article2/0,1895,1910318,00.asp

8)

[oiaohm]

Some Serious points my Dosemu does run with MS Dos and PC Dos.

So its not a tech option its fact.

Dosemu is not required by linux so loss of it will not be important. There is no reason for linux personal to protect it. Even linux supporting companys.

DosBox when its speed up can and will replace all its options. Note no Fat problem at all ie no Fat. Freedos can be built with out long filename support so Dosemu will work but limited. Its the long filename support that is protected.

With out hardware makers we don't have a market.

Foss links Open Source together true.

MS joliet no problem prior art fixs that problem. Unixs had a extention like joliet. Rock ridges so this will never come if it does it will die very quickly. Also iso9660 is a standard patent on top of a standard has never stood in a court if the company placing the patent did not create the standard. Yep iso9660 existed before Microsoft.

Ie joliet is just a case of Microsoft being Microsoft and breaking standards does not give them a right to place a patent on it.

NTFS toxic if patents stands. Basicly we have to except that we have to stand on our own feet. Depending on fat a base filesystem will not work. Basic fat compad without longfilenames is safe. No user what that.

A patent lock can stop developers in the USA and any other country where patents work from working in a infringing product. NTFS driver and VFAT driver will have to be split away from reactos just so developers don't get locked out.

SMB ie Microsoft filesharing. This is safe they stole this from Sun Microsystems. This could get really interesting if they try to place patents on it.

Foss does not prevent attack from patent holder they will help.

IBM will not allow this to happen to linux they will lose to much money. What cash do we make companys protection unlikely to come.

They will try to over turn a patent this does not happen that often.

People have to learn that floss runs on cash not good will. Other patent fights have been let slip due to no requirements. MP3 for one.