3

I'm releasing some source code as open source which includes a new, highly effective and efficient method of automatically classifying graphical data (image, video frames etc) as either pornography or not pornography. I do not want a patent, I want it to be free for everyone to know and study.

I'm concerned only about someone coming along and trying to claim the invention and patent it while it's still relatively unknown, as I lack both the saddle bags of cash and will to fight someone to prove it's mine. Does some mechanism exists where I can file information about this with the patent office, but only to inform them of this "prior art" for consideration when processing new patent applications?

  • Are you concerned about this in a particular country, or world-wide? – vallismortis Nov 2 '15 at 12:03
  • 1
    Well probably the U.S. because from what a patent lawyer told me once upon a time, people generally start there and expand out to nations that have some sort of treaty, I don't know. I believe my question may be a duplicate of this question actually patents.stackexchange.com/questions/83/… – user15342 Nov 2 '15 at 16:11
4

I'm with the many others who suggest to place your sourcecode on a public open-source management system of some kind. This will date the files appropriately.

However, look at any software patent, the very thing you are trying to prevent, and you will see flow charts of the process or method the patent protects. Source code doesn't really figure into much of anything except to implement the process or method.

Thus, there is one more thing you should consider publishing... the method or process you are implementing in your software. Somewhere you probably have a flow chart of the steps your software takes to classify the graphical data. Maybe its on paper... maybe it's in your head. Wherever it is, you should get it into a flow chart or something similar that describes the method/process so you can publish it along with your code. A good home for this method/process documentation is in the /doc directory next to your source code directory.

I suggest to document the process at several levels. The first should be a very very high level and very broad description of the basic steps you use to classify the graphical data. The next should provide more detail on the intermediate steps involved. Keep adding more flow charts with more detail until you are satisfied you have covered your process or method. By doing the above you have, sort of, defined broad claims on your process/method along with more detailed claims. Of course you aren't going to claim them on your own patent application, but you have effectively made your idea prior art.

Not an attorney am I... just my two cents.

  • Nice answer, I've been investigating doing a paper and trying to get it published in something like arxiv, only thing is I've never done such a thing and wouldn't know where to start lol, I'll figure it out. +1 – user15342 Nov 3 '15 at 21:19
0
  1. Making a work publicly available before someone else files a patent application on it automatically makes it 'prior art' for them - even if they don't see it, and even if the patent issues (so long as it is actually prior art, and you could demonstrate it was public).
  2. After it is 'publicly available' for 12 months, nobody can patent it.
  3. For a little more detail about prior art read patent/IP attorney Gene Quinn's articles pre-AIA and post-AIA.

One way you could easily publish it is by releasing the source code to the public on a public portal such as GitHub, SourceForge, etc. This should serve to make it 'publicly available' and prevent anyone from 'stealing' the IP from you. Note that if you ever want to patent it, assuming it's patentable - you have only have 12 months to file an application. Further, any 'offer to sale' is arguable a 'statutory bar' that immediately bars the invention offered for sale (which can be construed rather broadly) from being patented.

Hope this helps!

NOTE: I am not an attorney and this is not legal advice.

  • I wonder if me posting precompiled binaries and offering the application under an open source license could be considered "being offered for sale". Thanks for the information. – user15342 Nov 3 '15 at 4:17
  • 1
    Hmmm...Interesting concept. That would be a good question for an attorney! Perhaps if you received something in exchange - or, for example, were offering related services that hinged on you offering the binaries - than it could be argued that you were, in a sense, offering it for sale because of the commercial potential of receiving consulting/etc fees due to releasing it. – SRDC Nov 3 '15 at 5:20
  • Yeah I'll have to consult I guess, not that it really matters. I'm not actually interested in getting a patent myself, but rather preventing anyone from doing so and opening up all the code for educational purposes and to seek contributors to make it better. – user15342 Nov 3 '15 at 6:53
  • Document it and publish it on the web. A blog post is good enough. If you can get it published in a journal, that's even better. The main concern is making sure that your documentation is still easily available via a Google/Google Scholar search when the examiner is examining a hypothetical application covering the same invention in the future. – user4545 Nov 3 '15 at 15:41
  • @user4545 - exactly. The primary thing is to make sure that it is independently verifiable that it was made public by you. As one of the articles I referenced pointed out, a self-published blog/website could be suspect. Releasing the source code (which would typically be where the 'invention' is) on a 3rd party website is typically going to be pretty hard for someone to try and argue. – SRDC Nov 3 '15 at 18:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy