Getting a Software Patent

cardine

...
Jan 9, 2008
3,582
77
0
wordai.com
As I've been looking at all the patent litigation going on, and even reading some of Google's patents in search for myself, I've realized that I have several algorithms of my own that might be worth patenting. Things that I think will "solve" some longstanding problems in NLP, and things that I have already begun implementing in my own products. And things that I think nobody else has come up with - I have watched video after video from big name college professors on the payrolls of big companies like Google and I'm getting far better results compared to the algorithms that they are bragging about. And, most importantly, these algorithms have a big commercial application right now, and a potentially huge commercial application 5-10 years from now.

So although I think the patent system is broken and stifles innovation, it exists, and since it exists I want to use it to my advantage.

So my question is, has anybody had any experience getting a software patent? What is the time and cost associated with it? And if I wanted to get one (or five), how should I proceed?
 


I've done a lot research on this since I'm getting a few myself, I think it's broken too but it exists in the real world and since we have to operate in the real world we have to use it. I spoke with a local angel investor who is a bit of a "patent fiend" as he describes himself and he told me normal patents can be completed for about $4,000 - $6,000 but tech/software related patents are closer to $6,000 - $12,000+ depending how complex your patent is since you either need to find a lawyer with solid tech experience or a lawyer who's smart enough to know he needs some outside tech consultants. The wait time from the date you file until the date you actually have a patent issued can be as little as 2 years if years if your lawyer does really detailed research and drafting + a little luck to as long as 5 years if you have to wrangle with the patent office over the claims/prior art/amendments/etc.

Provisional patents can be filed online and on your own for about $200 but all it really does is give you an official filing date and 1 year of "patent pending" protection but it doesn't include any claims which is kind of the heart of a patent and it's never actually examined by the patent office so it's probably not something you would want to swing at a competitor in court but it is a good starting point. If you file a provisional patent you have 1 year to file your utility patent otherwise it expires.
 
Back in the day when I actually had lawyers, after my corporate lawyer finished drafted a bunch of agreements, during one of our meetings he brought up patents. The discussion never furthered after the original chat, but it was clear it was quite expensive in terms of both money and time.

Main issue was, technology is a difficult thing to patent, and even if you do, then you need to enforce it. Even with a patent, you still need to source out companies who are infringing on it, then take them to court and prove they're using your technology. Not easy, and only worth it if we're talking multi-millions.

That, and for example, you mentioned algorithms. It's a simple one, minor tweak of an algorithm, and it's no longer your algorithm. How are you going to prove in court they used yours (without using corporate espionage)? And just lots of small things like that.
 
As I've been looking at all the patent litigation going on, and even reading some of Google's patents in search for myself, I've realized that I have several algorithms of my own that might be worth patenting. Things that I think will "solve" some longstanding problems in NLP, and things that I have already begun implementing in my own products. And things that I think nobody else has come up with - I have watched video after video from big name college professors on the payrolls of big companies like Google and I'm getting far better results compared to the algorithms that they are bragging about. And, most importantly, these algorithms have a big commercial application right now, and a potentially huge commercial application 5-10 years from now.

So although I think the patent system is broken and stifles innovation, it exists, and since it exists I want to use it to my advantage.

So my question is, has anybody had any experience getting a software patent? What is the time and cost associated with it? And if I wanted to get one (or five), how should I proceed?

Generally rule is over here that a patent will cost you at least £5k to get through, depending on the amount of resistance you encounter. In the US I've heard that the average is about $10k including preparation, filing & prosecution fees etc..

What's the average cost of a software patent in the US? - Quora

Fuck software patents though, and fuck patents in general. The whole system is a mess. Software patents don't stimulate innovation IMO, which was the entire purpose of patents in the first place. Patents are a gold mine though [look at samsung] if you can get a few good one's, so might as well exploit the broken ass system, assuming you have the funds in place to defend your patent if need be.
 
Main issue was, technology is a difficult thing to patent, and even if you do, then you need to enforce it. Even with a patent, you still need to source out companies who are infringing on it, then take them to court and prove they're using your technology. Not easy, and only worth it if we're talking multi-millions.

That, and for example, you mentioned algorithms. It's a simple one, minor tweak of an algorithm, and it's no longer your algorithm. How are you going to prove in court they used yours (without using corporate espionage)? And just lots of small things like that.

I agree with all of this, and is making my decision even harder. I think the algorithms/data structures I am working on have the ability to be ubiquitous in NLP, to the point where billion dollar products (Siri, Google Search, etc.) would be using them. But of course there is no guarantee of that ever being the case, and even if it were, it's not like those companies open source to the world how they are powering their NLP algorithms.

Fuck software patents though, and fuck patents in general. The whole system is a mess. Software patents don't stimulate innovation IMO, which was the entire purpose of patents in the first place. Patents are a gold mine though [look at samsung] if you can get a few good one's, so might as well exploit the broken ass system, assuming you have the funds in place to defend your patent if need be.
Agreed, software patents are stupid, broken, and stifle innovation. But as you said, as long as they exist you can't ignore them.


Appreciate the responses from all three of you (first time in a while a thread in STS has gotten three straight good replies). I'm still on the fence about whether or not I should do it. I might do a provisional patent to give me a bit more time to keep up on the academic literature in the field and talk to a lawyer who knows a bit about NLP so I can get his take on whether what I'm working on is worth patenting or not.
 
Skype me his week. Ill hook you up with a friend who's going through this (his retired lawyer dad is helping). One false typo and you gotta revile and all sorts of red tape. Need lawyers who are on point.

He may have good insight. I'm on vacay now tho.
 
Copyright and patent laws have flat out ruined our country. It stifles innovation to the point where still driving on 4 wheels ... this should be the Jetsonian era.

Open source your shit, scream it at the highest mountain and make your mark on the world. You'll live in infamy and still have lots of money thrown at you. The biggest difference is that you won't be spending your life in court trying to collect.

I know I'm not being helpful to you or even productive to the conversation, I'm just squarely on the other side of the fence.
 
If any of you are interested in patents, trademarks, copyrights, let me know. I know a very good intellectual property lawyer that is a one-man shop so his prices are incredibly reasonable.

Check your Skype cardine.
 
Copyright and patent laws have flat out ruined our country. It stifles innovation to the point where still driving on 4 wheels ... this should be the Jetsonian era.

Open source your shit, scream it at the highest mountain and make your mark on the world. You'll live in infamy and still have lots of money thrown at you. The biggest difference is that you won't be spending your life in court trying to collect.

I know I'm not being helpful to you or even productive to the conversation, I'm just squarely on the other side of the fence.

I wish everyone but me did this. It would make my life a lot easier!

But I'd rather be Bill Gates than Linus Torvalds.
 
I wish everyone but me did this. It would make my life a lot easier!

But I'd rather be Bill Gates than Linus Torvalds.

Here's my reasoning on why open sourcing your algorithms that aren't essential to your product is to your advantage.

  1. You build credibility
  2. You save in lawyer and court fees that may never come anyway
  3. You get a higher chance at major contract offers
As others have pointed out, proving software patent infringement is a bitch. If this was a mechanical patent, then I could see your point. Those are much easier and less costly to prove. We all know there's many different ways to program the same appearance of functionality. If there weren't, Apple and Oracle would have ousted Google from the mobile phone market years ago. So you'd need to successfully obtain unfiltered access to their source code or some public acknowledgment of them using your algorithm to even have a legal claim.


Now, if you're an avid github comitter, then you're probably already aware of this, but publishing your open source algorithms tends to land you several requests for similar, outside work a year as does blogging about it. So, comitting the basic functions of whatever NLP algorithm you've revised or made yourself would give you tons of brand recognition and credibility in the industry, possible speaking gigs, etc. You couple that with a link to your site for an actual service using said algorithm and you have more users using your product and requests for contracts than you did without open sourcing the basics.
 
You can file a provisional patent which gives you 1 year to get the official patent filed.

But after 1 year it becomes public info and if you haven't patented it by then someone else can.

The day you file a provisional patent you can start saying "patent pending" on your product.

To get the real patent you have to have a professional do a "prior art search" to determine whether it's already taken.

From there, you can either file it yourself if you're just trying to protect your implementation, or can hire a lawyer to write the broadest wording legally possible if you're trying to get a patent for litigation. If it's the latter check this book out: [ame="http://www.amazon.com/Drafting-Patents-For-Litigation-Licensing/dp/1570187398"]Amazon.com: Drafting Patents For Litigation And Licensing (9781570187391): BNA Books: Books[/ame]


Prior art search is the first step once you have the prototype. I'm saving up for a prior art search right now. But you need a software attorney, not just a patent attorney. I plan to file myself after finding out officially whether it's taken or not. There are many patents that aren't developed, so even if you don't see your idea being used it could already be patented. Attorney is the expert at determining this. Also in my case there are a handful of similar patents, but they're not quite the same, so an attorney would be able to tell me whether it's different enough.
 
there is also 1 more thing. there is national [per country patents] and global ones. they differ alot between each other [also in terms of possibility getting them].
there are companies that check local patents, then try to patent them in other countires/continents for themselves.
 
Copyright and patent laws have flat out ruined our country. It stifles innovation to the point where still driving on 4 wheels ... this should be the Jetsonian era.

Open source your shit, scream it at the highest mountain and make your mark on the world. You'll live in infamy and still have lots of money thrown at you. The biggest difference is that you won't be spending your life in court trying to collect.

I know I'm not being helpful to you or even productive to the conversation, I'm just squarely on the other side of the fence.

IMO this isn't the most effective way of being a positive impact on society. To have the biggest possible positive effect I believe you need to get extremely wealthy first and then dive in altruism. If instead you do it the other way around there's a higher chance your impact will be severely limited.

By patenting your stuff and growing your company fast you can then use your money to fund research you otherwise would never be able to.
 
^^ Beyond that, software patents are not like copywrited works. The first to make a mechanic is not protected by the fact they were first.

Say you/your team develop a new ad network, if it's not patented nothing stops another company from copying it and submitting the patent, then suing you and putting you out of business for patent infringement of your idea!

And open source code is stolen all the time for financial gain by businesses/individuals. You're delusional if you think an individual making open source code would have any (realistic) recourse against someone who steals it. How are you going to pay for the lawyers and court fee's? How is your law team going to be better than the company who is marketing/making money off your idea's law team? Since you're not making money off it?

You make some open source code. Someone steals it and implements it in a product they're selling. You see product and think it's your code. You have to crack their software, prove it's a violation of your license, and then take them to court. They meanwhile, have a product they're selling which is making them money to fight you.
 
Here's my reasoning on why open sourcing your algorithms that aren't essential to your product is to your advantage.

  1. You build credibility
  2. You save in lawyer and court fees that may never come anyway
  3. You get a higher chance at major contract offers
As others have pointed out, proving software patent infringement is a bitch. If this was a mechanical patent, then I could see your point. Those are much easier and less costly to prove. We all know there's many different ways to program the same appearance of functionality. If there weren't, Apple and Oracle would have ousted Google from the mobile phone market years ago. So you'd need to successfully obtain unfiltered access to their source code or some public acknowledgment of them using your algorithm to even have a legal claim.


Now, if you're an avid github comitter, then you're probably already aware of this, but publishing your open source algorithms tends to land you several requests for similar, outside work a year as does blogging about it. So, comitting the basic functions of whatever NLP algorithm you've revised or made yourself would give you tons of brand recognition and credibility in the industry, possible speaking gigs, etc. You couple that with a link to your site for an actual service using said algorithm and you have more users using your product and requests for contracts than you did without open sourcing the basics.

I just don't see it being a good idea at this stage of the business cycle. A lot of the awesome work I have now that I think might be worth patenting came from standing on the shoulders of giants. I do not want others to stand on my shoulders.

So eventually when I have a team of 50 full time programmers and am making 8-9 figures/year in revenue I might be willing to open source some code, because at that point the people who would be building off of me don't have the resources, manpower, and focus to easily catch up. But until then, when it is just me and one other programmer, I risk lowering the barrier to entry and enabling other people to become potential competitors.
 
I just don't see it being a good idea at this stage of the business cycle. A lot of the awesome work I have now that I think might be worth patenting came from standing on the shoulders of giants. I do not want others to stand on my shoulders....risk lowering the barrier to entry and enabling other people to become potential competitors.

I agree with that. Which is why no one should release their service/product's essential offering to the public. Especially when your customer base could easily produce it given the proper examples as this community could.

What I meant was, as in your OP example, you were improving on already existing NLP and other libraries and trying to patent them (which you probably couldn't since the core would come from already free to use licenses), then commit those to GitHub or something. Something like that, since most of those libraries just return you sentence parts, key words, etc. isn't essential to your service.

Now, if you were committing the algorithm for your Turing section, for instance, then yeah - you're just giving people the green light for tossing up a new site in a couple weeks to directly compete with you since you'd have included how and what synonyms to substitute, how to define and work with grammatics, etc. That particular section I would class as an essential business offering, which only makes sense to release to the public after you have no key advantage of highlighting that feature of your service anymore and are looking to destabilize your competition, get free insights on how to improve it, etc.
 
Copyright and patent laws have flat out ruined our country. It stifles innovation to the point where still driving on 4 wheels ... this should be the Jetsonian era.

Open source your shit, scream it at the highest mountain and make your mark on the world. You'll live in infamy and still have lots of money thrown at you. The biggest difference is that you won't be spending your life in court trying to collect.

I know I'm not being helpful to you or even productive to the conversation, I'm just squarely on the other side of the fence.

This.

In my opinion, the only reason to worry about something like patents is the fear that someone else will do it better than you can.

Where intellectual property and patent laws mess everything up most is you can be vehemently against these laws, and go on creating, then someone else copies you to the letter, patents it, and sues you out of your own idea.

I haven't looked into if open source licensing protects you from this. The biggest turnoff to open source is Dennis Ritchie and his militant attitude against programmers turning a profit off anything they create.

In my opinion, its fine to add your own security measures to protect your product from being copied and whatever, but it is immoral and lazy to rely on government to do it for you.

There are so many means of preventing software from being pirated, and so many ways of benefiting even if your software is pirated, that from the perspective of a software company, there is no reason to need a copyright.
 
In my opinion, the only reason to worry about something like patents is the fear that someone else will do it better than you can.

There are so many means of preventing software from being pirated, and so many ways of benefiting even if your software is pirated, that from the perspective of a software company, there is no reason to need a copyright.

Having a patent can protect you from patent trolls who have no interest in developing a competing product.

Short scenario I ran into:

I wanted something different for advertising in games. I made a prototype where I can rent out virtual ad space in game. Any client that opens the game, connects to the server and dynamically updates the virtual placeholders' with the current "sponsor" icon and URL. There's a lot more to it I won't go into, but it grabs a 2d image from the server, converts it to a compressed format for mobiles, then applies it to 3d geometry. There's also incentives in game for clicking ads and tracking data for each user. Advertisers get exclusive use of the placement for a period of time.

Then some asshole pointed this out to me Patent US6954728 - System and method for consumer-selected advertising and branding in ... - Google Patents which is similar and different from my prototype, but the fact that ideas like that are patentable is ridiculous.

Which led me to look and I found that dynamically loading textures(images) from a server into a game, is patented as well (but is built into the API for the game engine I use, so they're not worried about it). Tracking user data in DB is patented, just about everything is patented. Things you do every day.

If you can get a patent, you're far more protected than not having one. If you don't have one using my example, the guy with the dynamic texture patent can easily sink you, people with advertising framework patents can nail you, because you have nothing. If you have a patent, you're a lot harder target to win from.

Many companies have patents just to prevent lawsuits, you sue me over patent A, which I may be infringing on, I'll sue you for patent C which you probably are infringing on.

If you program, you're likely at some point to be infringing on somebody.
 
Especially when your customer base could easily produce it given the proper examples as this community could.

Welcome to my life ... there's not shit anyone can do to keep things from getting cloned once you've got a product in the public's eye. It will get ripped with or without your source code or patents ... likely overseas. The ONLY remedy is being awesome all day every day ... and making sure that being first to market was insanely profitable.

It's simply impossible to fight reverse engineering.

You did a great thing with this hack, don't make your next few decades miserable by surrounding yourself with lawyers.