[Fsfe-uk] Symbian reanimates software patent debate

[Alex Hudson]

Nicked off some other lists I read; I haven't seen this echoed here yet. 
Symbian recently got a Court decision overturning a UKIPO decision to 
refuse them a software patent:

   http://www.bailii.org/ew/cases/EWHC/Patents/2008/518.html

Actually, Symbian has *two* patents they are appealing. This decision 
concerns the first of these two challenges:

http://www.ipo.gov.uk/patent/p-decisionmaking/p-challenge/p-challenge-decision-results/p-challenge-decision-results-bl?BL_Number=O/209/07
http://www.ipo.gov.uk/patent/p-decisionmaking/p-challenge/p-challenge-decision-results/p-challenge-decision-results-bl?BL_Number=O/238/07


One imagines if they succeed with the first, they may well succeed with 
the second. These patents in my best lay person's explanation:

  1. All about using DLL libraries. The idea with a DLL is that your
     program can call functions from a library, and you have to
     identify that function either by name or by the index into the
     library. This patent claims a slightly different way of doing the
     look-ups so that when libraries are changed in the future, the
     linkage still works. (Actually, I can't claim to really understand
     this - it seems to be a way of keeping an ABI consistent, but not
     much of it makes sense in the context of how I understand
     libraries to work ;)
  2. A method of accessing files in two directory hierarchies. This is
     worse than it sounds - it basically sounds like a method of
     "mounting" a second directory hierarchy onto the first, but not at
     a file system level - e.g., they propose to use string
     stripping/replacing as one way of doing it. As an example, if you
     replaced "C:\USB\" with "F:\", then all the files on F:\ might
     appear under C:\USB\.

These are both pretty seriously awful patent applications. They should 
fail for a vast number of reasons, but also because they are pretty much 
"pure" software patents. I don't see that they teach us anything new, or 
contribute technically.

I think the Judge in the Court case was basically bamboozled. The key 
text to my mind is this:

     62. I doubt whether very much is to be gained by trying to make
         some kind of direct comparison between the invention in
         /Autonomy /and that in the present case. In paragraph 21 of
         his judgment Lewison J said of the claim he had to consider that:

             /"What is of significance here is that the claimed
             invention required no new hardware or arrangement of
             hardware, did not fix any perceived technical shortcoming
             in the computer itself, and was purely concerned with the
             processing of data. This was done and done only by a
             computer program. "/

     63. In the present case there is a perceived technical shortcoming
         caused by modification to the DLL as a result of updates to
         the computer's functionality. This is not a case where the
         invention is limited to the processing of data. If an increase
         in the speed at which the computer works can take the program
         out of Art.52 (3) (see /Aerotel /at paragraph 92) it is
         difficult to see why the improved reliability of the machine
         brought about by the re-organisation of the DLL in its
         operating system does not.

Actually, this DLL thing is *totally* about processing data - it just 
happens that the data is code. A basic bit of computer architecture 101 
ought to have sorted that out, but clearly this judgement wasn't 
properly opposed: and, a patent which is entirely non-technical looks 
like it could get the nod through because it looks fundamental to how 
computers work.

Hopefully, if we in the UK can get our act together, we can stop this 
from getting approved: I don't think this patent is very strong, 
particularly for the above reason.

Cheers,

Alex.