Ask N.YC: What is the best way to research patentable ideas?

[kingnothing]

I've been thinking about a new idea for a couple of hours today. As is often the case, I got side tracked by another idea, and I think it may be novel and patentable.

How have you researched ideas you thought were patentable to determine whether or not you should step it up and talk to an IP lawyer?

[kirubakaran]

I found this book to be very useful:

Patent Pending In 24 Hrs

http://www.amazon.com/Patent-Pending-Hours-Richard-Stim/dp/1...

I also bought (but haven't yet read):

http://www.amazon.com/Patent-Yourself-David-Pressman/dp/1413...

http://www.amazon.com/How-Make-Patent-Drawings-Companion/dp/...

[sarosh]

Most IP attorney's will do quite a bit of research themselves. If you, or your university, have access to it, IBM's Delphion is an excellent resource (not free system).

I think your question is too broad and needs to be refined.

If you are looking to patent a software algorithm, the approach to research ideas would be very different from that for particular plants. As a general tip, search by patent class http://www.uspto.gov/go/classification/

Alternatively, search the pending applications of a large company in the field you are researching. For example: http://appft1.uspto.gov/netahtml/PTO/search-adv.html

And in the query field type:

AN/google

It is from the pending applications that you can look at 'prior art.' Prior art is the list of the patents and other materials upon which a patent relies. Here is a software analogy: If a patent is source code, prior art is an <include> directive or a <lib> call. Only by mastering the existing libraries can you produce new code. Look at some source code. Then examine the libraries/include files. And then start tinkering around on producing your own source code.

[kingnothing → sarosh]

To get a bit more specific, I found an answer to a problem dealing with the implementation of encryption in a particular environment that, as far as I know, hasn't been done yet. I suppose that would fit in to the software algorithm category.

I've been reading up on encryption in general today in order to figure out if the solution is novel or not, and so far, it seems as though it is. Looks like I have a lot more reading to go, though.

Do you have any more tips for researching software patents?

[sarosh → kingnothing]

If it's encryption software, browse around the Cryptography class: http://www.uspto.gov/go/classification/uspc380/sched380.htm

After you find the subclass, click on the red P to get a list of the granted patents. But note that these are only the patents that have been granted, you still have to check out the latest patent applications.

This system can seem a bit confusing, which is why there are so many service providers who specialize in this area. But with a little elbow grease, you can get through it yourself.

The defect with resources such as freepatentsonline is that they do not have the whole database. A good alternative is http://www.patentlens.net/patentlens/simple.cgi

If you are handy with regexpr/sexpr/xml search tools, you may want to take a look at the raw patent data from the FTP archives as well as the Patent Grant Red Book. More info at http://www.uspto.gov/web/menu/patdata.html From the raw data you could roll your own patent search engine.

Please also note that as a general habit you should research the academic periodicals in the Crypto arena to see if anyone else has published any information about the topic in question. If so, you will need to include it in 'prior art.'

If you are very serious about the process, check with a local attorney. If the prices quoted seem out of your ability, contact the nearest (and largest) law school in your area. Speak with the IP faculty who can point you in the right direction for local counsel.

[kingnothing → sarosh]

Thank you, that's some great info.

[kf]

http://www.google.com/patents

What you need to do is read ALL of the patents for anything similar to what you are doing. You need to hope that you find patents for things that are similar enough to imply that your idea is also patentable but that no one has actually claimed what you are trying to claim.

This takes an extraordinary amount of time to do which is why lawyers charge so much money. But you can definitely do it yourself, if you want.

[kingnothing → kf]

Thanks. That's kind of along the lines of what I was thinking I would have to do. I figure if I can do a bunch of research and find out that what I'm thinking of is not patentable, it would save me a boatload of money. On the other hand, if I don't come up with anything, it would give me more of a reason to think that it is patentable and to go ahead and meet with a lawyer.

[pierrefar]

One thing you can do is to ask a company to do a patent landscape for you. There a few around (can't remember their name) and they charge a few grand to do it for you. The results usually come in within a week or two.

The instructions for them would be to look at encryption in general with particular focus on the environment you're talking about.

Also, try freepatentsonline.com instead of the USPTO's website. I find it's much easier on the eyes and the search a bit more human-friendly.

Pierre

[wmorein]

One thing to be careful of here. If you intend to implement the idea (rather than just patent it) doing the search puts you at risk to some degree. If you do the search yourself (possibly not finding the relevant patent) and then do the implementation anyway, you are at risk for "willful infringement". This is all really murky and you probably want to talk with an attorney if you are serious.

Link with some relevant info: http://wistechnology.com/article.php?id=1303

[boucher → wmorein]

Well, you can't patent ideas in the first place. And doing a google search isn't going to get you sued for willful infringement.

[wmorein → boucher]

Most large tech companies have an explicit prohibition on non-lawyers doing patent searches. At all.

Doing the patent search won't get you sued, but if you are sued and they win, the fact that you did the search means that they can hit you for triple damages.

I've leave the "you can't patent ideas" piece alone here because I think that is semantics.

[DaniFong]

There are two steps, looking for similar patents, and reading them to decide if you'd be infringing.

I can't offer any tips for the former, but for the latter you may want to skip ahead to the 'claims' section: this is the part with the legal meat in it.

[goodmoney]

I used Global Prior Art, a boston based IP consulting firm for some quick and dirty patentability searches. They were cheap and very good. Saved us months of work by avoiding some patented approaches.

[pixcavator]

Yes. :) Not a very useful answer, sorry.

Seriously, my advice is to search the PTO site. Unfortunately that's not very useful either. Just too much stuff.

[eru]

No.

[kingnothing → eru]

What?

[eru → kingnothing]

No, I have not researched those things. 'No' was not a very useful answer, sorry.