Long discussions about long names

Posted May 5, 2009 1:02 UTC (Tue) by k3ninho (subscriber, #50375)
In reply to: Long discussions about long names by JoeBuck
Parent article: Long discussions about long names

The USPTO allows people to start using an invention publically, even selling it, before filing for a patent as much as a year later. Then also, they allow people to apply and say that they were the first to invent the thing in a patent owned by someone else.

It may well be that, following the original nullification of the LFN patents, Microsoft showed that they had invented and used privately (maybe for testing purposes) their long file name technology. Perhaps they showed proof that their invention was before the publication of Rock Ridge. Maybe there was some way that they could argue that the parallel development of Rock Ridge didn't render their method obvious to a person having ordinary skill in the art. But we'll never know because the hearing was private and so my (perhaps unreasonable) assumption is that the convicted monopolist cheated again.

to post comments

Long discussions about long names

Posted May 5, 2009 2:00 UTC (Tue) by drag (guest, #31333) [Link] (3 responses)

> The USPTO allows people to start using an invention publically, even selling it, before filing for a patent as much as a year later. Then also, they allow people to apply and say that they were the first to invent the thing in a patent owned by someone else.

Ya.. but in this case RockRidge is a extension to allow POSIX compatibility with the ISO 9660 file system. Its draft is RRIP 1.1.2 has copyright references from 1993.

So.. we are dealing with multiple patents here:
U.S. Patent 5,745,902 -- filed 1992
U.S. Patent 5,579,517 -- filed 1995
U.S. Patent 5,758,352 -- filed 1996

Then IBM holds a patent on FAT, too..
U.S. Patent 5,367,671 -- filed 1990, for extended attributes on FAT. Used by NT, Linux, OS/2.

So only IBM and one of the patents exist prior to 1993... (I would of expected that one to be cross-licensed years ago with Microsoft from the OS/2 days)

But even possibly more important then RockRidge is UMSDOS. Which I remember quite fondly because that was used by the first Linux system I ever used.. which would be Zipslack.
http://en.wikipedia.org/wiki/UMSDOS

UMSDOS is a method for making FAT POSIX-compliant filesystem for the sake of Linux compatibility. Quite nice for booting Linux from Zip drives.

It was started in 1992 and was accepted into the Linux kernel in 1994 with 1.1.36 until it was dropped in 2.6.11. If I remember correctly it stored extended information in extra hidden files.

-----------------------

So at least with UMSDOS it pre-dates all of Microsoft's patents. With Rockridge it only certainly pre-dates 2 of them.

-----------------------

Now everybody knows there are dozens and dozens of different manners to do stuff in software. This is one of the problems with software patents. To any problem there are thousands of solutions. (and each one is patentable)

So even though Rockridge and UMSDOS cover one way to have long file names it's likely that they do not do it in the same manner that is described in Microsoft's patents. Just a guess, I am not going to try to pretend to understand any of the patent language. So I suppose that would be why they don't immediately invalidate the patents.

----------------------

Probably the read-long-only patch would work great for embedded developers. Since they do not write to the FAT flash, generally, then it's only the users that really care about long file names. When you plug in a device to a PC then it's that PC's software that writes to the flash media and not any software on the device.

Just guessing, but it seems reasonable.

Long discussions about long names

Posted May 5, 2009 3:49 UTC (Tue) by JoeBuck (guest, #2330) [Link] (1 responses)

But Microsoft didn't patent their own way of storing long names in a length-restricted system. None of the claims refers to anything in particular about their system. Rather, they patented pretty much the very idea of storing long names in a length-restricted system by encoding the filenames.

Long discussions about long names

Posted May 5, 2009 22:48 UTC (Tue) by man_ls (guest, #15091) [Link]

This same concept appears in Kernighan and Plauger's "Software Tools" as early as 1980. And in this text it is proposed as an exercise! Surely the obviousness test is not passed either.

Long discussions about long names

Posted May 5, 2009 4:42 UTC (Tue) by Oddscurity (guest, #46851) [Link]

So even though Rockridge and UMSDOS cover one way to have long file names it's likely that they do not do it in the same manner that is described in Microsoft's patents. Just a guess, I am not going to try to pretend to understand any of the patent language. So I suppose that would be why they don't immediately invalidate the patents.

Still, reexamination in that case ought to have the effect of significantly limiting said patent's scope.