[Gauss-parl] [Fwd: Re: EP 1 040 409 B1 (Microsoft) - JVM able to use Java classes precompiled into a DLL]

[PILCH Hartmut]

> If there are (as in this case) systematic differences between what is 
> available online in HTML (?) and the B1 PDF-docs (granted), we should 
> look into Hartmut's OCR project.

Start by linking to the OCR'ed pages, e.g.

	http://swpat.ffii.org/patents/txt/ep/1040/409/

The database on genba tells you which pages are ocr'ed.

Btw we need someone to talk to the author of GOCR and get him to patch
gocr so that it reliably recognises the two-column format of EPO
patent descriptions.