From as early as January, but since I'm currently cooking potatoes, it reminded me again: The End of a Potato Variety: Soon No More Linda. About the attempt by the previous patent holder to restrict the availability of the potato variety Linda to avoid competition - a good example of what patents and free market really mean: ultimately only monopolistic tendencies and monopolistic consolidations.
Does anyone really believe that potatoes and algorithms are so different that such activities will not occur in the IT industry when implementing software patents? Who is the winner in such a situation - the customer for whom a potato variety disappears, the small farmer or greengrocer who can no longer supply a well-known product to their customers, or the large corporation that holds the patent? Actually, the patent would have expired, actually, real competition could now begin. Actually ...
By the way, the loss of the potato variety Linda is more than annoying - it was one of the few varieties I like and that also survived Jutta ...
Free Pascal is a Turbo Pascal and Delphi compatible Pascal compiler. The new version supports significantly more CPUs and platforms - including Mac OS X. And for Panther there is also an XCode integration. Finally, a Pascal that works for the Mac. Not that I would do much with it today - but somehow Pascal was a part of my programming history (after all, I took care of the Gateway software Erwinsgate for the MausNet, which was written in Turbo Pascal, for a long time).
May seems to be the month of vintage languages for me.
Who wants to see the full horror of software patents in real application: search for the patents DE69901832 (German patent) or EP1081612 (European patent) in the patent database (unfortunately, there are no permanent links to patents in this patent database - why not? Stupid software).
A truly great invention: the identification of a session in a web application is not transported in the path part of a URI, in a cookie, or a URI parameter, but in the hostname (wildcard A-record makes it possible). A patent has been granted for something like this - although there are supposedly no pure software patents in Germany. Where exactly is the definition of technicity in this patent? The connection to a tangible device? This is a pure algorithm patent.
If you read through the description and claims, you get the chills: in the description, it is written that extracting the encoding from the path part would be complex and require a lot of computing power - probably this is supposed to be the reason for patentability, so to speak, more efficient access to information. And in the claim, nothing else is described than the use of HTTP/1.1 virtual hosts in combination with a wildcard A-record (or many A-record entries that point to the same IP), so that the session can be extracted from the hostname (which is sent as a header). A truly great intellectual achievement - presumably the entire creative height of it lies in the formulation of the patent application, but certainly not in the actual algorithm or the encoding ...
The patent office simply patents any nonsense that is submitted. There seems to be no control at all - anyone who wants to do something against a patent must first laboriously apply for and enforce the deletion of the patent. If the patentability of algorithms is further facilitated, we will find even more such absurdities. Because if the patent office is already overwhelmed today to carry out these basic checks on general patentability, it will certainly get worse with the new regulation. And this is supposed to bring economic growth? Probably only for patent lawyers ...
Found in the dead-tree version of the current Linux Magazine on page 102.