Return to patent.doc CVS log

Up to [GPL Stacker compression] / dmsdos / doc

DMSDOS 0.9.1.5

Concerning the relationship between dmsdos and patents I keep receiving mail 
like this one... 

> In article <[email protected]> you wrote:
> : dmsdos 0.9.0 has just been released. It supports read and write access to
> : compressed dos filesystems [...]
> 
> I think you should make sure no patents are violated. I am afraid that you
> MUST get permission to distribute your code. I wish it was impossible
> to get a software patent, but unfortunately companies do receive these
> patents. There are many patents in encryption and compression technology!
> [...]
> Again, I wish it was possible/legal to write a free variant for every
> existing commercial variant. Unfortunately it is not always possible.

Let me say some gereral things about this. The patent problem is indeed as 
old as the first dmsdos version, which was released approximately three 
years ago.

First, please allow me to correct some statements that are not litterally
written above, but that sound like a misunderstanding. Dmsdos is not a 
"free variant" of an "existing commercial variant". Even both "free" and 
"variant" are wrong. Dmsdos is not free, but copyrighted and distributed 
under the GPL. 

Dmsdos also is not a "variant" of some commercially available software
package. It is just a tool that makes it possible to read from and write
to some variants of compressed dos filesystems. Dmsdos is surely not a
filesystem compression software package like stacker and doublespace, as 
such a package needs a lot of more software e.g. all kinds of creation and 
maintainance tools. I know it would be nice to have them under Linux, but 
I won't write them for well known reasons.

If you follow the dmsdos history, i.e. the dmsdos documentation of older
releases, there have always been some comments on the possibility of patent 
problems mentioned in the documentation, and thus some features users liked
to have in dmsdos were not implemented. Not because I knew they are 
forbidden by patent but because I thought they *might* be covered by a 
patent.

Yes, I did contact the respective companies (no need to say which ones :) )
for legal issues and, of course, I asked whether they were willing to help 
developing the code by providing documentation about their filesystem. 
So what? When I really had luck and got an answer, it was of no value. A 
stripped down version of one answer was published once in the dmsdos 
documentation. I removed the name of the company and the name of the person 
since I didn't want to blame one person. But I wanted to show the level 
of interest of these companies - on the one hand in support and cooperation,
and on the other hand in a Linux version of their code. It was absolutely 
zero.

So I did my own research. It was surely not exhaustive. Patents are things
for lawyers, but I'm not a lawyer myself and I don't have the money to
get a bunch of lawyers study all the software patents concerning data 
compression. Furthermore, I don't earn any penny with dmsdos.

It lead to the result that there are a lot of patented compression
algorithms. Also the compression algorithms that the original dos software
uses for filesystem compression are covered by patents. So what. Dmsdos
doesn't use them. Dmsdos was developed without official documentation, and 
it turned out that its compression algorithm even reached a higher 
compression ratio (but was much slower).

I still tried to contact the companies, rarely, but it became more and more 
boring. I must admit that I gave it up some time ago. I also must admit
that I didn't add all the patent problem related stuff to the dmsdos
documentation when I rewrote it some day. I considered it simply dead.
This implies that some features are still missing in dmsdos and will
probably never be added because I don't just want to be exposed to the
risk of violating a software patent and provoking a company owning it.
The community of dmsdos users on the net seem to have accepted this.

I also must say that dmsdos is not at all fully my own work. I just happen
to maintain the code currently, and I'm not doing this on my own. The 
documentation used to implement dmsdos came from a lot of people on the
net and even from a previous sample implementation(*) that was released under
the GPL. If you know the GPL you also know that it has a very restrictive
patent section, so I considered this quite safe.

In fact, the compression and decompression routines in dmsdos are something
like a collection of parts of free or GPL'd software. Most of them have
meanwhile been rewritten from scratch for better performance. They use very
common compression techniques. And you don't receive a patent for something 
that is well known. You can receive a patent, e.g. for a special, highly 
optimized algorithm, but, let me repeat this, dmsdos does not use any 
patented compression algorithms.

So what can I do? Just throw away the dmsdos code and remove it from
the servers? This is like a snail going back into its house and staying
there though nothing is happening outside. Just continue trying to contact 
the companies? I'm bored by their answers if I happen to get one. Sorry.
Just giving dmsdos maintainance into the hand of someone else? Heh. That 
would solve the problem probably for me, but not for others.

If you just happen to know more about the patent situation than me, please
let me know.


(*) ftp://sunsite.unc.edu/pub/Linux/system/Filesystems/dosfs/thsfs.tgz
    written 1994 by Thomas Scheuermann (current email address unknown)

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.