There has been some discussion on the Linux AU mailing list in the past few days about Software Patents, and in particular asking for specific examples of a situation where development of a peice of Open Source Software has been hindered by software patents. The best example given was an application called rproxy created by Andrew Tridgell (of Samba and Rsync fame) and indeed rproxy uses the same underlying principles as rsync, it uses delta compression to speed up HTTP browsing. The problem is some company in America has a patent on this. Tridge managed to get a “gentlemans agreement” that the patent wouldn’t be enforced against rsync, but couldn’t secure the same agreement for rproxy and hence the development stalled.

I found this example fascinating, rsync is a hugely useful tool that I literally use every day, so it’s a great example to bring home to me why software patents are a threat to the open source community. More details of the rproxy situation at:

• The Tragedy of rproxy
• RFC 3229
• Old rproxy homepage

In light of all this I found the following IEEE article quite interesting because it proposes a solution to the software patent debacle rather than just further bemoaning the current situation. I also like it because it lines up with my personal ideas of what should and should not be patentable.

• IEEE Spectrum: Software Patents Don’t Compute