By Jason Rantanen
For the past few years, I’ve been working on a project with the working title “the Federal Circuit Database Project.” The goal of this project is to develop an accurate, reliable and transparent database containing information about the Federal Circuit’s patent law-related decisions that would be of interest to scholars, commentators and policymakers. In keeping with the public nature of the University of Iowa College of Law, I wanted the database to be accessible and usable by anyone.
The first stage of that project, now called The Compendium of Federal Circuit Decisions, is complete. The Compendium contains records for all documents released on the Federal Circuit’s website in an appeal arising from the USPTO or District Courts. In practice, this means all written opinions back to late-2004 and all Rule 36 affirmances since mid-2007, along with a smorgasbord of orders and errata. Each record contains multiple pieces of information about those documents, including whether the decision was precedential, who the authors are, whether there is a dissent or concurrence, the year of the decision, and more. The data and its interface was structured from the ground-up in a format that would be useful for people seeking quantitative data about the court’s decisions, rather than for the purposes of traditional legal research. The data export feature allows for all the information to be used in your favorite analytical toolset.
In order to kick off the Compendium, I wrote a short article describing the database and reporting some core descriptive statistics about the court’s decisions in appeals arising from the UPSTO and District Courts. The draft of that article, called The Landscape of Modern Patent Appeals (forthcoming Am. U. L. Rev.) can be accessed here. It’s still in production–the graphs, in particular, are low-resolution placeholders.
Below are graphs that might be of particular interest to readers of this blog: the types of nonprecedential decisions in appeals arising from the USPTO and District Courts. The paper goes into more detail, but the short of it is that the frequency at which the court used Rule 36 affirmances fell substantially in 2017 (for district courts) and for both 2016 and 2017 (for the USPTO). 
There are many more descriptive statistics and graphs in the paper, including the proportions of Federal Circuit decisions that arise from the PTAB & BPAI versus the TTAB, degree of agreement among panel members, the rate of the court’s production of precedential opinions, and even per-judge authorship data.
Although the project is now at its first milestone, there is much more to be done. Future stages include adding new types of information to the records (such as subject-matter issues and disposition), adding documents in appeals arising from other sources, particularly the ITC, and improving the on-the-fly graphical functionality. Comments and feedback on the project are welcome if you email me directly.
Note: In order to continue with the soft opening, and to avoid the possibility of stress-testing the Compendium, I intentionally did not include a link directly to the database in this post. The Landscape of Modern Patent Appeals does contain the link. I’ll add it to PatentlyO in the near future.
As this data develops, especially because the outcomes at the CAFC are so panel-dependent, it could become a very useful tool for litigants to asses their chances. Advanced stats work well for the NBA 😉
Oracle v. Google
link to cafc.uscourts.gov
O’Malley, Plager, Taranto
CAFC finds Google’s use of the Java API packages “not fair” as a matter of law. District court’s decision to the contrary is reversed; case is remanded for a trial on damages.
…so much for “just logic,” eh Malcolm?
In your weak mind, you just totally decimated me! Very impressive.
“just logic”
LOL Are you quoting me?
What is the point of your reply?
I brush away the mindless ad hominem, and, there is nothing left.
“not fair use”
Fixed that for you.
What I would like to see, but did not see in this post, are the two graphs in this blog post rendered as percentages of all opinions rather than as absolute numbers. If nonprecedential opinions and Rule 36 affirmances are becoming an ever greater percentage of the court’s output, then the exception is swallowing the rule. Figures 6 and 8 in the paper appear to be relevant, and from them, it seems as if Rule 36 affirmances are holding steady at around 30% but nonprecedential opinions are slowly creeping ever upward.
I think that you may have confused the exception and the rule as it may apply to nonprecedential opinions.
One should expect such to “creep ever upward.”
When you believe that the way to build certainty in the common law system is to have every decision carefully considered such that it has precedential force, the whole idea concept of nonprecedential decisions is anathema. It’s especially unfair for any reversal to be marked nonprecedential. If the law was so clear, how was the error made below? If the argument is that what we really have is a systemic problem of the USPTO and/or the district courts just plum not following the law, then that issue needs to be drawn out and addressed, not concealed in the shadow law of a large volume of nonprecedential opinions.
I have to disagree – not every decision has to somehow “be different” enough in its own right (and thus earn precedential status) in order to build certainty.
In fact, quite the opposite: MOST decisions should merely be the application of the developed – and consistent – law.
Your view of “each” necessarily being precedential is the opposite of consistency, as there would be NO consistency in such a system. There would be NO “fixing” of law, as each decision would bring a change to the law to be applied.
As to “especially unfair for any reversal” – not buying it.
A reversal may – or may not – warrant being precedential, and decisions below sometimes (and sometimes far more than they should) are the result of the lower court simply misapplying what should be consistently applied.
You beg the question be presupposing that shadows exist in “a large volume of nonprecedential opinions” – as that should be the rule rather than the exception.
It is in fact axiomatic that the rule is defined by the large volume and the exception is defined as, well, the exception to that large volume.
RKS What I would like to see, but did not see in this post, are the two graphs in this blog post rendered as percentages of all opinions rather than as absolute numbers.
Definitely. Also a comparison with the trends in non-IP cases would be useful.
Bravo.
Comments are closed.