Micro$oft to license FAT

I have a Motorola phone which has the earpiece in the flip part and the mic in the main body [and no push buttons, just a touchscreen].

This is unique, novel and non-obvious to those skilled in the art? Loading a register before calling a function? A bios call?? Please...

I disagreed with the phrase "Clearly it was not all that trivial", because it was.

--Gene

Well, just load the EAX register instead of AX and the patent won't apply :-)

... snip ...

They did, and it was available much earlier. It was the 'description' field via 4dos, which could (and can) be searched on, include blanks and funny characters, etc.

If you look at the download section of my page, URL below, you will see those "long filenames" carried to a ridiculous extreme.

Their target audience for license are music and camera manufacturers. How many of those would you think are using an Intel x86 type processor in their camera or MP3 player? Haven't seen too many with fans and a 20Ah battery attached yet ;-)

Rob

There was also a utility in Norton Utility that could add descriptions to file names. I think it was in 4.0.

That utility *was* that very same 4DOS, licenced by Norton and renamed 'ndos'.

Markus,

I've been in touch with one of their licensing specialists. He requested confidentiality so I won't tell you in detail what he told me, just that it isn't as bad as it sounds. Suffice to say they encourage but don't demand using the license under certain circumstances.

I can however tell you what I told him... I explained to him that the license offer had created quite a lot of confusion in the embedded developer community (this group, slashdot), as it's unclear under which circumstances it would be needed. He promised to relay that to the proper people for consideration in a reviewed text. If you want to know exactly how/when you'd want this license, go ahead and e-mail the address on the page. If you're concerned that your current products or projects will be impacted by it, same thing, use the e-mail link on the page. They provided me with a decent answer, I'm sure they'll do the same for anyone else asking.

Rob

Hi Rob

Thanks for shareing this with us. Seems like I have to contact them then also.

Markus

I think you guys are being a bit silly. Obviously the "description" in DOS was not the same as a "long filename". If it was, it would have been patented, not the long filenames that we are discussing. If the description was prior art, even though it may have been from the same inventor, it would preclude the patent on long filenames.

A lot of people think things are not valid patents because "they are obvious". That may well be true now that they are in common use. But someone had to be the first to invent it and that makes it patentable. The "obviousness" is intended to cover things like using a screwdriver to open a paint can, not the first time someone applies a tool to a new area, like changing the shape of the screwdriver to self center in a Philips type screw. You could say that the Philips screw and driver were obvious once the need for a machine operated driver required self centering. But someone had to be the first to do that.

But then some people just like MS bashing (myself included), but there are much better topics to bash MS on.

(Unfortunately.)

So what is your point? Anyone could have added long filenames starting with DOS 3.1 then. I seem to recall that there was about 5 or 6 years between DOS 3.1 and Windows 95.

So what is your point? As you indicate, what MS did was different from what was done in these other systems and was both new and novel (not obvious).

(which was actually 4dos, licensed and relabeled)

On the contrary, the description provided all the functions of the later MS long filenames, albeit in not quite as integrated a manner. It could be searched, converted to a short file name, the results opened, deleted, renamed, etc. although some coding might be needed to provide all those services. I believe it would be perfectly possible to provide all the functionality of the lfn DOS calls through a DOS call interception TSR module providing the same API as the MS system and using the 4DOS system. The major cost would be the loss of the filename "DESCRIPT.ION", which is less likely to bite than the loss of NUL, CON, etc.

Not that I am advocating doing it now :-) (Unless someone wants to pay me to do so)

Yes, that is my point. Motorola has made many variations and I expect they are all covered by their patent.

The functionality of long filenames is not at issue. A patent covers the "how" of doing it. If you can do it a different way, then you are free to use that method. But the method that MS chose was not obvious and not trivial.

But other people had already invented the concept of mapping long and short file names to each other in a backwards compatible way. All Microsoft did was choose a different method of doing that. If you asked a class of students to implement a method whereby the FAT filesystem can support long filenames without breaking legacy operating systems then you'd get several solutions, some of which could resemble what Microsoft did.

If once you state the problem the solutions present themselves immediately, then the solutions wouldn't qualify as nonobvious to a practitioner of the art. In this instance the problem statement was known and at least two solutions existed (OS/2 and Rockridge).

Further, I think that Microsoft's solution could only have been created by Microsoft anyway, even if someone else thought of it first. If anyone other than Microsoft used that exact method those changes could easily have broken with the next release of Windows or DOS (ie, suddenly the reserved bits are used for something).

I don't think your test of obviousness is anything like what the patent office uses. This is not a term from the dictionary. This is patent jargon and has a specific meaning just like code means something specific to us which is different from what it means to the average person.

Your test of a classroom reinventing the MS lfn approach is not useful. The bottom line is that the method was not in use for serveral years showing that it was not obvious. MS invented it and patented it. Why does so many have a problem with this? This is what patents are all about. They allow companies and individuals to profit from their inventions without someone stealing them. I know of several examples of small companies or even individuals who were able to defend their patents and reap their just rewards. If there are lots of alternatives to the MS lfn approach, then use them. I don't see why this is even being discussed.

Perhaps a lot of people feel that such a thing shouldn't be patentable. Many also feel that software patents shouldn't exist. Quite a lot of people actually feel that the entire system is broken and open to abuse. Microsoft is merely taking advantage of this.