Wednesday, November 17, 2010
Obviousness and Cumulative Innovation
I've been spending time recently (way too much time, I think) participating in prior art searching for ArticleOnePartners, a company that pays its bounty hunters advisors prizes if they find prior art that invalidates certain patents. The company, in turn, uses the gathered research to license to patent defendants. The prizes used to be huge ($50,000 or so), but are now moderately big ($5K to $15K for the most part).
It's no mystery why I do it: I want to win some money. I haven't won a big prize yet, but I've got a couple "studies" that I am optimistic about. Even without winning, the endeavor has been extremely helpful in developing my understanding about innovation, obviousness, and prior art. That's what I want to write about below the fold.
The story begins with conventional wisdom. Whenever a high profile patent is asserted, the media reports the claims in only the most general terms, like "the patentee claims to have invented hyperlinking." People reading the article immediately protest that surely this inventor could not have invented something so a) important or b) broad.
The reality is that few patent claims actually read like they are depicted in the media. First, they are usually much narrower than the media reports. For example, the "hyperlinking" patent was really a much narrower claim about hidden data that points to blocks of memory retrieved from a single central computer using a keypad. It's possible, of course, that clicking a link using a mouse (where the URL is not really hidden memory) to obtain a web page from a variety of servers would infringe by the doctrine of equivalents, but it's not clear. The technology in the patent describes no way to communicate with more than one server. (Of course, the owner sued Prodigy, who may have connected users to a single server).
Second, such claims, though usually narrower than reported, are also usually filed much earlier than people think. The hyperlinking patent, for example, was originally filed in 1976. If someone really did invent hyperlinking in 1976, my view is give them the patent - they were years ahead of their time.
This is where Article One Partners comes in. The company knows that it is difficult to find evidence that even narrow claims were not new when the patent application was filed early. It only takes one document or product to invalidate a patent claim, but only if that one piece of prior art includes every element of the patent claim. Further, because we often don't know when the claim was invented, we can only be sure to invalidate the claim if the prior art is dated one year or more before the patent filing date. This makes information very difficult to find. Thus, Article One Partners enlists others to find that prior art with money as a reward for a successful search.
My searches have opened a frustrating but fascinating window into the nature of innovation. As I look at each patent under review, I almost always think, "No way they invented that." (Sometimes, it really does seem like the patent was a pioneer.) However, when I start digging, I find many instances that are close, but are not the same. Every time I think I've found the silver bullet, I review the patent claim and it includes some small detail that makes the invention different than what came before.
While "close, but no cigar" is not putting money in my pockets, it is helping me better understand innovation and obviousness. In just about every case (there are a few where I think I have found the silver bullet), you can see technology developing, with many companies racing to add the next feature, each building on the other. The most fascinating process to me is a claim for data storage in web pages. This is technology that every HTML form programmer takes for granted today - you put a variable name and value into your dynamic web page so that it remembers what you've typed in when you load the next page. This method is so basic that it is part of every manual on HTML programming: "Authors may create controls that are not rendered but whose values are submitted with a form. Authors generally use this control type to store information between client/server exchanges that would otherwise be lost due to the stateless nature of HTTP..."
While hidden fields are basic now, they weren't basic in December of 1993, the cutoff date to invalidate this patent. While I think I have found invalidating prior art, the method was certainly not a standard. Instead, it was in the process of being invented by others at exactly the time of the cutoff. I found fascinating discussions among those who invented the HTTP protocol about how to store information between pages, and I found draft standard documents that were published two days before the cutoff.
This, I think is the most fascinating part of the experience. We talk now about "crowdsourcing" as if it were new, but a lot of innovation has come about through crowded fields with many fingers in the pie. The innovation process that was artificially cut off in my search reached the identical solution in the patent soon thereafter (if not before).
My final point is that this innovation process affects how we should think about obviousness. Where the crowd is close and quickly moving, it seems that many of these inventions should be considered obvious. Obviousness is how one determines whether a close, but not identical, piece of prior art invalidates a patent. Patent law has an obviousness factor called "simultaneous invention," and I've come to believe that the factor is woefully underused. In some cases the patentee may make a leap away from the crowd, but I am finding that for many of these patents, the inventor won the patent only because he or she was the first to the patent office and it was difficult to find out how others were doing the same thing.
I am convinced that as search engines digitize and index more of the world's information, more patents will be invalidated as obvious. I am surely not the first person to be convinced of this, but I have now seen first hand how difficult it is to find do in practice.
TrackBack URL for this entry:
Listed below are links to weblogs that reference Obviousness and Cumulative Innovation:
Is the takeaway that a patent is nonobvious only if the inventor wins the race to invent by a furlong, but not if she wins by a nose?
That leads me to wonder why a particular field might be crowded. If the field is crowded because there is an unmet societal need for innovations in that field, then could we lose something valuable if the simultaneous invention rule incentivizes inventors to move to less crowded fields to innovate, where they can be more sure their labors will result in a patent?
Posted by: Jake Linford | Nov 17, 2010 11:37:27 AM
Jake - that's a great question. Obviousness is exeedingly complex and uncertain, so I would say that the takeaway can never be a single rule. It may be that there is a crowded art, but no one can figure out how to take the next leap - telegraphy is an example of that. Everyone was trying, but no one could make it work. I do think, however, that you have to look at the trend in a crowded art. If the crowd is trending toward the same solution, then you need the furlong to win the race, I think.
I don't worry so much about movement to less crowded fields for a few reasons. First, there are trade secret, first mover, and other advantages in crowded fields. Second, if the social need is that great, it will be met one way or the other, as a piece of the pie might still be big. Third, we want people to be pioneering in non-crowded areas; as money for basic science dries up, it wouldn't hurt to have companies move back to new technologies. The demise of industrial scientific research like Bell Labs is a great loss to society, and I think it is driven in part by moves to get short term gains in known markets. While basic science is unpatentable, IBM and Bell Labs did not have problems finding patentable inventions in pioneering research - research that has given us some of our greatest inventions.
Posted by: Michael Risch | Nov 17, 2010 11:44:13 AM
I had a lot of the same intuition when I was in practice. But you need to refine the simultaneous invention factor a little. The reason that courts have basically discarded the simultaneous invention factor is that, quite often, the classic "pioneering inventions" were all subject to simultaneous invention. Take Edison and Swann with the light bulb, or Bell and Gray with the telephone. The trick you need to do is to treat all of the patent-seeking people and treat them as a group, since if you deny the patent as obvious then all of those people would disappear. What you have left (and are trying to identify) are people who would create the same invention even without patent incentives. And that might be a far smaller group than the "crowd" that appears at first glance.
Posted by: TJ | Nov 17, 2010 12:59:42 PM
TJ - I agree with your concern. I think two or even three people working on something simultaneously (and perhaps secretly) do not render each others' work obvious. On the other hand, where multiple people/companies are working on standards and products that are publicly available, then "simultaneity" becomes more important.
In the former situation, the solution hasn't been found because the problem is difficult to solve. In the latter situation, the solution hasn't been found merely because no one has looked at the problem yet.
Posted by: Michael Risch | Nov 17, 2010 1:13:56 PM
This is a fascinating and insightful entry about the difficulty of determining the validity of a patent. I too am involved with Article One - gotta love the thrill of the hunt and the potential of winning up to $50,000 every time you submit a reference!
I came across a great white-paper they released titled, "Patent Quality: Leveraging Patent Quality Research to Reduce Litigation and Improve Patent Portfolios" - http://info.articleonepartners.com/patent-quality-whitepaper
In this whitepaper they discuss how high patent quality can ensure the success of litigation and low patent quality increases the costs to both the public and private sectors. Patent quality affects many players, and the recent increase in litigation and patent applications has thrust quality to the forefront of discussion.
Quite a good read!
Posted by: Greg | Nov 17, 2010 4:58:54 PM
The comments to this entry are closed.