Gauss patents, where?

[Carsten Svaneborg]

On Thursday 07 April 2005 14:26, Roland Orre wrote:
> OK, now I've found 117808 patents, even though I've not yet checked
> if the list is unique, I found the following:

What you find below ~zqex/PatData/pat is unique.
Deleted patents are in ~zqex/Patdata/backup


> but the blacklist contains 121731 entries so some are obviously missing.

The blacklist just identifies known EP numbers, but i could have downloaded
and deleted these previously. So it can't be used for anything except which
patents are "known".


> I don't see the point having the patents in so many directories.
> but I'll keep the structure if you don't have another opinion.

It's more a organisation of the search results, such that I can
remember what I searched for, for instance recently I redid all
the word searches, by a simple perl script that extracted the
search words from the previous search output file.

> Are you continue downloading EPO patents, as I understand there
> is nothing after March 15th?
No.  This is done.

>> * Is there a complete list of all EPO patents?

Look at ep.espacenet.com under coverage you can find the range of
EP publication numbers. As today the latest patent is EP1521515.

>> * Is there a list of clear software patents versus a
>>   list of real patents? The larger the better for the
>>   classifier.

No. I would guess that most of the patents in the DB is software
related [1], and most patents not in the DB is not software.

[1] except for the ones that aren't.

I think the most sensible thing would be to clean the DB for non-swpat,
that gives spam for the classifier, and then iteratively it should be
possible to identify and remove more and more spam. But that
requires some manual work to identify spam.

You can start with the patents in ~zqex/PatData/backup which were
identified by OleTange.


>> After that you can run the script on gauss.ffii.org.

I also have to implement XML IO for pleech.pl and the import script.

>> /scratch/pat/people/wagner/EP520400.html
Aha. The blacklist is generated on my desktop computer where the
paths differ. /scratch/ -> ~/zqex/PatData

-- 
  mvh. Carsten