MS patent on FAT a problem for ros?

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

inclusivedisjunction
Posts: 23
Joined: Wed Mar 07, 2007 8:09 am

Post by inclusivedisjunction » Sat Jan 19, 2008 11:14 am

I thought Windows did that with most system files anyway. I can see how that might suck for testing some applications that are designed for 2000 / XP/ Vista, and don't have to comply to such insanity, but it would only have to last until you ported a "non-infringing" file system to ReactOS. Other than the possible performance overhead, what would it hurt to drop FAT32 / vFAT support?

linuxgx
Posts: 170
Joined: Wed Mar 29, 2006 4:18 pm

Post by linuxgx » Sat Jan 19, 2008 11:36 am

Your usb pen drive wont mount when we get usb support.
Thats personaly a big one to loss in my eyes

inclusivedisjunction
Posts: 23
Joined: Wed Mar 07, 2007 8:09 am

Post by inclusivedisjunction » Sat Jan 19, 2008 11:45 am

linuxgx wrote:Your usb pen drive wont mount when we get usb support.
Thats personaly a big one to loss in my eyes
You can format a pen drive of up to 2 GB with FAT16, on which the patents would have long since expired. Even more if you partition it.

Besides, by the time USB is working, ReactOS should have a new native file system, I would think.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Sat Jan 19, 2008 12:49 pm

Ok showing how little you know. FAT16 does not have lfn out box.

Long Filenames is basically a after fact hack for fat. So the patent applies to that. So no matter what fat you use it cannot have long filenames if the patent is upheld somewhere.

inclusivedisjunction
Posts: 23
Joined: Wed Mar 07, 2007 8:09 am

Post by inclusivedisjunction » Sat Jan 19, 2008 1:02 pm

oiaohm wrote:Ok showing how little you know. FAT16 does not have lfn out box.

Long Filenames is basically a after fact hack for fat. So the patent applies to that. So no matter what fat you use it cannot have long filenames if the patent is upheld somewhere.
So don't use long file names. It's a little awkward, but it's not the end of the world.

ntoskrnl.exe - 8.3
comctl32.dll - 8.3
uxtheme - 7.3
dinput.dll - 6.3
explorer.exe - 8.3
browseui.dll - 8.3

Pretty much every system file I can think of has an 8.3 file name.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Sat Jan 19, 2008 1:14 pm

And pritty much every user I know uses long filenames.

Its kinda a rude shock to them. Its the handling. If we cannot use it basically have to avoid FAT to keep users happy.

inclusivedisjunction
Posts: 23
Joined: Wed Mar 07, 2007 8:09 am

Post by inclusivedisjunction » Sat Jan 19, 2008 1:23 pm

Restricting ReactOS's support to FAT16 would only need to be a temporary measure, until a more suitable alternative, like JFS or ext3 was implemented. I'm basing this all on the premise "What if Microsoft pursues patent royalties on FAT?" Reader may have thought this was a problem because he read that ReactOS ONLY supports FAT. My claim is that:

1.) ReactOS can run on FAT16.
2.) FAT16 is no longer under patent protection
3.) Support for other file systems is being developed, so ReactOS can continue to be used and developed, just without FAT32.

linuxgx
Posts: 170
Joined: Wed Mar 29, 2006 4:18 pm

Post by linuxgx » Sat Jan 19, 2008 4:11 pm

Everyother OS has support for fat32

You all need to do your homework, but the fact is this who leagel fight was droped a long time ago, as MS went with NTFS, they have taken no legil move agianst anyone since. This is all a waste of time. And should only be discussed if problems come from it. Now NFTS support should be the concern

inclusivedisjunction
Posts: 23
Joined: Wed Mar 07, 2007 8:09 am

Post by inclusivedisjunction » Sat Jan 19, 2008 4:25 pm

linuxgx wrote:Everyother OS has support for fat32
None of which makes it legal to use it if Microsoft tries to collect royalties.
You all need to do your homework, but the fact is this who leagel fight was droped a long time ago, as MS went with NTFS, they have taken no legil move agianst anyone since. This is all a waste of time. And should only be discussed if problems come from it. Now NFTS support should be the concern
Microsoft created NTFS long before they created FAT32. I doubt Microsoft will actually try to sue over FAT32. I think their real reason is to prevent anyone from making an enhanced version of FAT, like they have done with exFAT.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Sun Jan 20, 2008 12:37 am

inclusivedisjunction Microsoft legally cannot sue of FAT32 alone since its purely based on FAT12 design.

Sorry they cannot prevent people making a enhanced version of FAT. Novell has the prior art for exFAT. So only one that could have patents over exFAT is Novell not Microsoft. Kinda explains why Microsoft is paying Novell for patent usage.

Please put FAT12 FAT16 and FAT32 disk structs side by side. They are the same only thing different is bit width. So patents on the base of the file format cannot be taken out since FAT12 is the prior art to FAT16 and FAT32.

There are no active patents on the base of Fat since they have to be expired by now.

Long Filenames feature is what they can sue over its a latter feature. That is not part of base fat design. ACL support in exfat MS might be able to try to patent but most likely covered by Novell patents.

Feature of TFAT cannot be patented any more its prior art is too old. Its running atomic writes. Form of fat that would be good in ros reduces screwups. Atomic system can be applied to most filesystems. Of course Atomic comes at a price speed and fragmentation. It avoids the need for a journaling while keeping the data on filesystem solid.

http://linkinghub.elsevier.com/retrieve ... 9X03001729
Just like Microsoft to give something a new name so it appears to be a new thing. Note some clones of dos also could be set to run atomic writes. Not there is no reason why ROS could not alter the start tag of its FAT32 parition with TFAT features to stop other OS's picking it up and messing with it unless they use TFAT like features.

Free Space Bitmaps feature of lots of Unix and Linux file systems.

Basically after pulling apart exFAT there is nothing much they can protect prior art kills most of it.

ps OPPS I though I pasted the short link. Both point to exactly the same place.
Last edited by oiaohm on Sun Jan 20, 2008 1:57 am, edited 1 time in total.

linuxgx
Posts: 170
Joined: Wed Mar 29, 2006 4:18 pm

Post by linuxgx » Sun Jan 20, 2008 1:08 am

Thank you oiaohm

counting_pine
Posts: 237
Joined: Fri Nov 26, 2004 10:44 pm
Location: Fallowfield

Post by counting_pine » Sun Jan 20, 2008 1:43 am

Would you mind editing that long link? It's making the forum much too wide for my screen.
Thanks :)

marsilies
Posts: 3
Joined: Wed Feb 27, 2008 11:59 pm

Post by marsilies » Thu Feb 28, 2008 12:43 am

oiaohm wrote:Sorry they cannot prevent people making a enhanced version of FAT. Novell has the prior art for exFAT. So only one that could have patents over exFAT is Novell not Microsoft. Kinda explains why Microsoft is paying Novell for patent usage.
Microsoft doesn't pay Novell for their patents. The two companies have a patent share agreement, which means each can legally utilize the other's patents for their commercial products.
http://en.wikipedia.org/wiki/Novell#Agr ... _Microsoft

Even if exFAT's patent is owned by Novell, that doesn't mean Microsoft doesn't own patents on the underlining FAT structure, nor that exFAT is open for Open Source use.
So patents on the base of the file format cannot be taken out since FAT12 is the prior art to FAT16 and FAT32.
Patents certainly can be based on prior art, and have been all the time. This is why patents expire, so that new works that build on previously patented work can occur. This is why Open Source software is released under GPL or some other variation thereof: because if they were just released into the public domain without any license restrictions, companies like Microsoft could build on that code and privately patent/copyright their modified versions. With the GPL, companies are required to release modified code as open-source.
Long Filenames feature is what they can sue over its a latter feature.
That's interesting, as LFN support is a feature of VFAT, which is a modified FAT16.
Feature of TFAT cannot be patented any more its prior art is too old. Its running atomic writes.
TFAT's a specific implementation of atomic writes on exFAT. If exFAT is under patent protection, then by extention TFAT is as well
Not there is no reason why ROS could not alter the start tag of its FAT32 parition with TFAT features to stop other OS's picking it up and messing with it unless they use TFAT like features.
Unfortunately, this would kill interoperability, which would be the whole point for using FAT32 in the first place.

oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm » Thu Feb 28, 2008 9:08 am

Please look at the money changing hands around that deal.

The deal involves upfront payment of $348 million from Microsoft to Novell for patent cooperation and SLES subscription. Novell will pay around $40 million to Microsoft over 5 years.

Novel is pay cents for all of MS Patents really. MS had to pay a heck load to get the deal. Novel could pay Microsoft its amount out the interest on the money MS payed them. Note deal only lasts 5 years too. So when 5 years is up if Novell still hold the key patents MS will be paying another big lot of cash.

Fat12 and base Fat16 was a joint project shared with IBM.

Patents issued over prior art are invalid if challenged. There is a difference between a Issue patent and a valid one. Stupidly patent offices don't have to make sure the patent is valid before issuing patent.

Valid feature must be something new. Not the prior art itself. Patent Fat32 disk structs would be straight out prior art infringement. Only thing that has changed is the bit width same as between fat12 and fat16.

Many companies have learn that the hard way when trying to apply patents to Redhat. Since as normal operation Redhat does a prior art search to attempt to destroy the patent before paying.

Vfat was not a joint project.

Putting patents in way kills interoperability anyhow. So ros having its own operation partitions not past options.

marsilies
Posts: 3
Joined: Wed Feb 27, 2008 11:59 pm

Post by marsilies » Thu Feb 28, 2008 4:09 pm

oiaohm wrote:Please look at the money changing hands around that deal.

The deal involves upfront payment of $348 million from Microsoft to Novell for patent cooperation and SLES subscription. Novell will pay around $40 million to Microsoft over 5 years.
In that deal, only $108 million is for Novell's patents. And in turn, Novell is to pay out at least $40 million over 5 years.

http://www.linux-watch.com/news/NS7235986827.html
Note deal only lasts 5 years too. So when 5 years is up if Novell still hold the key patents MS will be paying another big lot of cash.
I haven't read anything that states the agreement is only good for 5 years, just what Novell's payments for the first 5 years would be.

Also, I can't find anything that supports your claim that Novell owns the patent on exFAT, or even on the "prior art" to exFAT. Everything I can find states that it was developed by Microsoft off of FAT.
http://en.wikipedia.org/wiki/Comparison_of_file_systems
Fat12 and base Fat16 was a joint project shared with IBM.
I don't get the point of this statement, since you claimed that both of those FAT weren't patented anymore, meaning that Microsoft can patent their new versions of it without any licensing issues.
Patents issued over prior art are invalid if challenged. There is a difference between a Issue patent and a valid one. Stupidly patent offices don't have to make sure the patent is valid before issuing patent.
According to the articles about their patents, Microsoft's patents on FAT were challenged, and were ultimately upheld in court.
Valid feature must be something new. Not the prior art itself. Patent Fat32 disk structs would be straight out prior art infringement. Only thing that has changed is the bit width same as between fat12 and fat16.
There are more differences between FAT32 and FAT16 than just the bit width:
http://www.computerhope.com/fat32.htm
"The Major differences between FAT32 and earlier implementations of FAT are as follows:

Two new partition types are defined: OxB and OxC. Both indicate FAT32 volumes; type OxC indicates a FAT32 partition that requires extended INTI3 support (LBA).
The boot record on FAT32 drives requires 2 sectors (due to expansion and addition of fields within the BPB). As a result, the number of reserved sectors on FAT32 drives is higher than on FATI6, typically 32. This expanded reserved area allows two complete copies of the boot record to be stored there, as well as a sector in which free space count and other file system information is stored.

The FAT is now larger, because each entry now takes up 4 bytes and there are typically many more clusters than on FATl6 drives.

The root directory is no longer stored in a fixed location.

A pointer to the starting cluster of the root directory is stored in the extended BPB.

The on-disk format directory entries is unchanged, except that the two bytes previously reserved for Extended Attributes now contain the high order word of the starting cluster number."


As for prior-art, that was presented during the court sessions over Microsoft's patents, and they weren't convincing enough evidence.

Arguing over whether Microsoft should have a patent on certain forms of FAT is fruitless anyway, since legally they do. What we should be spending time on is deciding what this means for ROS and it's support of FAT's various incarnations.
Putting patents in way kills interoperability anyhow. So ros having its own operation partitions not past options.
If ROS isn't going to play nice with standard FAT32 anyway, why not just go with ext2 or some other exiting file system that will afford some interoperability, instead of creating a new FAT that may or may not have patent restrictions, but will certainly not be usable on any other OS?

Locked

Who is online

Users browsing this forum: DotBot [Crawler] and 3 guests