[Debian-uk] Symbian get UK patent on linking to a DLL -
analysis wanted
Will Newton
will at misconception.org.uk
Mon, 19 May 2008 23:01:17 +0100
On Mon, 2008-05-19 at 22:13 +0100, Magnus Therning wrote:
> Ben Hutchings wrote:
> > On Tue, 2008-03-25 at 17:16 +0000, Wookey wrote:
> > <snip>
> >> We (FFII and killsoftwarepatents.com) would be interested to know
> >> which pieces of software will be infringing this - can anyone here
> >> give examples, or other analysis of this? Feel free to pass this mail
> >> on to people who might.
> > <snip>
> >
> > Windows seems to have prior art:
> > <http://blogs.msdn.com/oldnewthing/archive/2006/07/19/671238.aspx>.
> >
> > However, since this patent apparently refers specifically to linking by
> > ordinal, and that's rare on Win32 and unknown in ELF dynamic linking, I
> > can't think of anything infringing this.
>
> Doesn't wikipedia say that linking by ordinals on Windows is more common
> than that blog post?
>
> http://en.wikipedia.org/wiki/Dynamic-link_library#Symbol_resolution_and_binding
>
> I also vaguely remember in my Amiga days that it wasn't too uncommon
> with link errors based on numbers, so that system might have used
> ordinals more extensively than more recent ones.
>
> Then I can't help but be amazed that identify-by-ordinal provides an
> inventive step over identify-by-name. I guess that shows just how naive
> I still am when it comes to swpat ;-)
For those interested the text of the appeal is here:
http://www.bailii.org/ew/cases/EWHC/Patents/2008/518.html
The core of the patent is the extension of a DLL by another (using an
ordinal linking scheme). I'd guess that plugin-based software using an
ordinal linking scheme could fall foul of this. The conclusion of the
appeal was that the patent office had followed incorrect procedure in
denying the claim, but the court did seem pretty keen on allowing a
wide-ranging application of software patents IMO.