Re: Article: IBM wants to "clean up the license" of Linux

David Feuer (feuer@his.com)
Sun, 20 Dec 1998 20:51:33 -0500


"Albert D. Cahalan" wrote:
>
> > I am kind of monitoring this list until I get up to speed but could you
> > go a little further in depth explaining some of the issues involved with
> > the patents on these technologies? Also are the expiration dates on
> > these patents known? I would appreciate it alot as I am sure others
> > would.
>
> Well...
>
> >> IPv6 routing Several good algorithms are prohibited.
>
> Alan Cox mentioned this to me when I pointed him to a nice trie-based
> system for keeping track of routes.
>
> >> compression Arithmetic encoding (bzip), MPEG, audio, GIF...
>
> Oringinal bzip uses arithmetic encoding for the second stage, which works
> better than the huffman encoding currently used. The same exact problem
> hits jpeg, with everyone forced to use the inferior compression.

Are all arithmetic coders patented? This is evil.

>
> Many kinds of perceptual compression algorithms for audio and video
> are patented. They remove the data that humans don't notice, and get
> far better compression.
>
> GIF compression is an old and well known problem. Lots of people
> violate it though. PNG can do a better job with plain images, but
> the MNG animation format is not yet well supported. I doubt this
> one has many years left on it.
>
> >> encryption Idea, RSA, some kinds of key exchange...
>
> PGP hit some problems here. RSA (the company) holds many.
>
> >> color management The gimp can't do it. Damn.
>
> At least 3 patents hit us here. The gimp site has a web page about it.

Could you explain?

>
> >> filesystem updates We can't do delayed ordered writes.
>
> Old AT&T patent.
>
> >> OSPF The IETF is not happy with this one!
> >> quality-of-service Multi-protocol Label Switching (MPLS) QoS
>
> Both IBM patents. Cisco has a few nasty ones too AFAIK.
> These make it hard to do some networking stuff.
>
> >> register allocation Ever wonder why VC++ code can beat g++ code?
>
> Various people AFAIK. Something about graph coloring...

?????????

>
> >> PPP Can't return a useful error code to the other side.
>
> Use grep on the kernel source. Motorola holds this one.

What???? Error return is obviously not a new idea! It's been around
since...... hmmmm.... telegraph times?

>
> >> drivers/net/zlib.c Must use code with nonlinear time requirements.
>
> 2 guys hold it. Not serious. You can't do gzip with a nice time bound.
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/

-- 
David Feuer
feuer@his.com
dfeuer@binx.mbhs.edu
Open Source: Think locally; act globally.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/