Microsoft licensing

Microsoft's new licensing announcements are stirring some controversy.

The ClearType license issue seems pretty much cut and dried. If you like what ClearType can do on LCDs and want to short circuit the process of getting there, license it from Microsoft and be done with it. There doesn't seem to be a way to accidently use it (although I can certainly see ways to possibly violate the patents).

On the other hand, the approach with FAT licensing is going to be interesting. There will be a lot of arguing.

Flash cards, memory sticks, MP3 players, et. al. have adopted FAT because there are drivers for mounting virtual FAT devices everywhere. The whole reason Microsoft has patents on long filename handling in FAT is because the file system is so old that it didn't handle reasonably long names. They added long names with what amounts to a patented meta data extension system. Without those extensions, FAT would be useless to third parties seeking a virtual file system for their devices. It was adopted because it was ubiquitous and Microsoft has basically waited until everyone was using it to decide that they didn't really mean to make it as open as it appeared to be.

U.S. Patent 5,579,517
U.S. Patent 5,745,902 ^[1]
U.S. Patent 5,758,352
U.S. Patent 6,286,013

It's an interesting signal for Microsoft to send as it appears to say to developers 'be wary of adopting our standards because we can and do bite'. Not exactly a surprise, but they aren't usually up front about it. This has been my biggest problem with later versions of SOAP and WS-Security; Microsoft and IBM both claim IP rights.

Finally, I don't buy the premise in this /. post:

I had a conversation with one of their licensing officers as I was afraid my 10-a-year GPS logger project was in danger. He explained this was an encouragement to have everyone implement FAT32 and LFN the same way, by using their reference design. All this to prevent incompatible implementations down the road.

¹ For what it's worth, I was surprised to see Bill Bruffey's name referenced and had to check out the related patent (4,945,475) for HFS data organization (which some of us know as B*-Tree's).