Archive for January, 2013

What can we ask of the USPTO?

Wednesday, January 23rd, 2013

The USPTO is organising roundtables to get suggestions from software developers. Below are my initial thoughts on what changes we can ask for. If you have other suggestions, please contribute to this wiki page:

A big problem with this sort of hearing is that the audience isn’t specialised in knowing how the USPTO can influence patent policy, and which court rulings bind the USPTO in which ways. So there will be a lot of simply invalid suggestions (which should go to Congress or the courts). Let’s try to reduce that waste.

The other problem is that because the patent office’s role is so limited, we can’t ask for much.

Here’s what I think we can usefully ask for:

  1. I’d start by saying that they don’t need to ask the software community. Judges consistantly invalidate software patents, much more than non-software patents, as shown in this study:

    Surely the USPTO can read those rulings to see what they’re doing wrong. For a suggestion to be valid, a proposed change has to be compatible with the courts’ case law and be within the USPTO’s scope of competence. Software developers don’t know these things, so it’s very hard to come up with valid suggestions, but judges are trained and experienced in exactly this, and they write a legal reasoning each time they invalidate a patent, so the answer is there in their rulings.

  2. Second, I’d highlight that the Supreme Court is over the district courts and the CAFC, and that the USPTO should apply the Supreme Court’s Mayo ruling to remove the computer or storage medium from applications when they’re examining them. Putting software on a computer is a trivial and expected step that shouldn’t contribute to patentability.
  3. Third (but I don’t know how much leeway the USPTO has on this), instruct examiners to send unclear applications back to applicants asking for a clearer description using more words that have definitions in the industry. This should “improve clarity of claim boundaries” so that hardware patents (a “point of sale”) stop spilling over into software (online shop).
  4. Finally, I’d point out their ongoing lack of resources and that raising or lowering standards doesn’t reduce the amount of examination necessary, but narrowing the field of patentable subject would. So, while the USPTO mightn’t have the power to determine what is legally patentable, I’d recommend not being afraid of proposals to remove a sector of any field from eligibility (and we know that software is the only sector for which exclusion is seriously discussed). If this happens, the USPTO will have more resources for improving evaluation quality in the remaining patentable sectors. By removing the least-loved sector, this will also reduce the amount of negative press about patents and curb the increasing discontent with the patent system in general. Everyone will be happier.

Probably not worthwhile:

  • Ask them to change the fees structure so that they’re no longer paid for what they accept. Problem is, this means reforming the USPTO across the board, which means no deal until pharma a few other other patent-intensive sectors are sure that they win something and lose nothing. (Just look at how much time and watering-down was required to get the America Invents act passed.)

Definitely not wanted:

  • Patent applications have to include a source code example. This would just add copyright problems on top of existing problems.

Those are my initial thoughts. Please add yours here: