So what do you think about software patents? It's a subject surrounded with controversy and contradictions.

I became concerned that the rules might be open to abuse when I read the patents for a well-known disk partitioning tool. I chose this tool because I had once written some low-level disk utilities, and I thought my experience might be relevant.

I was surprised to see the patent read more like the description of a business process - move some data from one place to another, and so on - than of a software technology.

I guess this is the nature of the beast. After all, a patent becomes invalid if prior art - the existence of the same thing before the patent was issued - is proven.

But there is a danger that people could obtain patents for technologies that they have not yet created. This scenario is not as unlikely as one might think.

Well, I became interested in the patent-free Gzip compression algorithm once I noticed that Microsoft had begun using it. It was at a press briefing about Exchange 2003 recently that I found out that the new version of Microsoft's groupware uses Gzip compression to speed up network performance from the email server to certain types of client device.

Now, I had come across Gzip on various Linux and Unix systems, and in popular Windows tools, such as WinZip, and I guess Microsoft must have first used it for the Compressed Folder feature in Windows XP. So I was interested that Microsoft's use of open compression is growing.

I was particularly interested as Gzip was developed specifically as an open - IETF RFC 1952 - compression algorithm that did not contain any patented technologies. Given that Microsoft did not simply go off and grab a ready-to-run version of Gzip from the web, it surely chose Gzip because it's an open standard that does not include patented technologies.

Well good for Microsoft.

I guess it chose Gzip so that both Microsoft-designed devices and third-party devices could connect to Exchange using the same compression. Now if Microsoft's choice of a patent-free compression algorithm is an irony, it is not the only one.

Among the rather dubious patents issued in the past by the US are several for perpetual motion machines. Eventually the folks at the US patent office came to their senses and stopped issuing patents for what is commonly accepted to be something that is impossible to invent.

But this has not stopped them from allowing several patents of compression algorithms that claim to be able to compress random data. For the record, storage experts agree that it's not possible to compress random data.

This type of oversight might be innocent enough, but it does not help to build confidence in the various national patent systems that each affect us all.

Perhaps the final irony is that, probably, the only way Microsoft could be sure that Gzip did not actually infringe or include any patents was because it could inspect the specification and various source code implementations. Unfortunately companies buying commercial products cannot usually be so sure.