Study Suggests Patent Office Lowered Standards To Cope With Backlog 96
An anonymous reader points out a story at Ars about how the "significant reduction" in the backlog of pending patent applications may not be all that it seems. "...a new study suggests another explanation for the declining backlog: the patent office may have lowered its standards, approving many patents that would have been (and in some cases, had been) rejected under the administration of George W. Bush. The authors—Chris Cotropia and Cecil Quillen of the University of Richmond and independent researcher Ogden Webster—used Freedom of Information Act requests to obtain detailed data about the fate of patent applications considered by the USPTO since 1996.
They found that the "allowance rate," the fraction of applications approved by the patent office, declined steadily from 2001 and 2009. But in the last four years there's been a sharp reversal, with a 2012 allowance rate about 20 percent higher than it was in 2009."
Correct buried among incorrect bits in article (Score:5, Informative)
First, the incorrect or misleading:
But in the last four years there's been a sharp reversal, with a 2012 allowance rate about 20 percent higher than it was in 2009.
It can also be looked at as 12% higher [patentlyo.com], since it went from about 58% to about 70% in that time. "20 percent higher" is misleading. "120 percent of the 2009 rate" would be correct.
Calculating the real allowance rate is tricky because inventors can submit the same application multiple times. "From the perspective of the patent office, a 'final rejection' doesn't get rid of an application," Quillen told Ars in a December phone interview. If an application is rejected, the inventor can make minor changes to the application and file it again. "The only way you can reduce your numbers and get rid of somebody is to allow the case," Quillen said.
There are a number of different ways to re-file applications, with names like File Wrapper Continuations, Continued Prosecution Applications, Requests for Continued Examination and Continuation-In-Part Applications. But in all cases, the upshot is the same: the applicant gets another shot at convincing examiners to grant him a patent.
These are talking about two different things, mainly because the journalist doesn't understand the distinction:
1) An application can be "finally rejected" by the USPTO (meaning that it was rejected on specific grounds, the applicant replied, and the Examiner wasn't persuaded and "finally" rejected the application on those grounds), and the Applicant can amend to narrow the claims and file a request for continued examination or a continued prosecution application (the same thing, but for design patents).
For example, say you were Toyota and were patenting the Prius, and you originally had a claim of "1. A car, comprising: four wheels, an engine, and a battery" and the Examiner comes back and says "duh, that's every car." If you tried to argue that you meant a battery running the engine, but the Examiner wasn't persuaded since that distinction isn't in your claims (and there's other prior art with electric engines), they'd finally reject it. If you then amended your claims to recite your novel planetary engine dual-powered transmission, you'd have to file a request for continued examination so that it could be considered.
It's not really "submitting the same application multiple times" but several iterations of narrowing the application and arguing that as narrowed, it's patentable, until it finally is narrow enough to be allowed.
2) New applications can be filed as continuation applications or continuation-in-part applications, but they're not the "same application" by definition. Continuations and continued prosecution applications have the same specification, but different claims. Like, say, to save money, you wrote a patent application that described two separate, but related inventions. Like say, a new machine for more efficiently turning horses into glue, and a new offset gearing system for use in that machine or others that has increased torque with reduced tension. You could file a single application describing both, but only claiming the horse part. Later, you could file a continuation application using the same specification and claiming the gearing system part.
This helps small inventors by not requiring them to file dozens of applications on day one, when they might be short of cash, but file one big one, then later file additional continuations as they're able to raise capital. Importantly, all of the applications have the same effective priority date for prior art, and any patents coming from the applications expire on the same d
Re:Correct buried among incorrect bits in article (Score:3, Informative)
As a patent examiner I'll throw in here too. There has certainly been a palpable shift in how we are encouraged to process cases over the last few years. It has much less to do with president 'X' and more to do with the current director of the PTO. Each one that comes in has their own priorities. From 2005-2009 the heavy focus was on quality. For examiners this translated into avoiding errors from the QA department. Every allowance submitted by an examiner is reviewed by QA while only a random sample of rejections are reviewed, so rejections are less likely to result in an error for the examiner than allowances. Examiners generally refrained from suggesting claim amendments to the applicants that we believed would make the claims allowable over the prior art because this was seen as sticking your neck out for the applicant.
Recently the focus has changed more to backlog reduction. One of the problems with backlog as you noted, and was discussed in the article, is that a 'final' rejection does not dispose of the case. Rather an applicant can pay for continued examination (RCE) of the case on narrowed claims. For large companies the fee for an RCE relatively inconsequential (a few thousand dollars depending on the number of claims) so there is little impetus to abandon the case or to make real changes to the claim language. Some cases can kick around the PTO for 7 or 8 years in examination which ties up examiners from working on new cases. We can't force an applicant to abandon a case (even when they really, really should), and the PTO doesn't have the authority to set a limit on the number of RCE's (it was tried), so the only real alternative is to try to make the cases allowable, thus disposing of them.
Whether this has caused a significant drop in quality is debatable, in a lot of cases, as you also noted , its the examiner taking the initiative to suggest claim language that gets around the prior art which probably would have been arrived at eventually . But, I cant deny that the fear of QA has also diminished a bit so examiners may be less likely to apply the more risky obviousness rejections (4 or references, obvious to try, etc) than they might once have.
Management has also started to track the examiners 'actions per disposal' metric, which should ideally be at 3 (non-final rejection, final rejection, allowance/abandonment/RCE), but any instance in which we must issue a 2nd non-final or several after final communications skews this number above 3, so direct supervisors sometimes encourage their examiners to work toward allowances to get this number down.
As to your point on continuations, they don't necessarily let an inventor extend a single patent longer than 20 years, but they can be used to file a sort of submarine patent. This works by disclosing, in the original (usually very extensive) specification some invention that will eventually be, but is not initially, claimed. Several years of delay ensue while the parent is backlogged, examined, then eventually allowed. Immediately prior to allowance a continuation is filed, on some portion of the spec not originally claimed which goes through the same process. This can be continued as long as there are more 'independent' inventions in the specification. As a result some patents filed in this decade have priority extending back to the mid 90's, but are not subject to double patenting rejections. The claimed invention is technically 'new' according to its priority date, and parent cases, but the obvious problem is that the technology described in the original patent has been built upon and entered the market, and these patents start to surface 20 years later but none prior art published between the original filing and the present can be applied to them.
As a last point please include a drawing of the sexy animantronic mouse in your next application. We examiners can always use a good laugh.
Re:Disgusting! (Score:4, Informative)
It is sickening in the extreme to think that it's possible to deny other people access to information, simply because you thought of it first.
It's worse than that. Patents don't deny access to information, but they curtail our freedom to help each other. And those who register patents almost never think of it first. Did Apple think first of a rectangular device with rounded corners?
Of course not, but then, they never patented "a rectangular device with rounded corners". This whole bit of FUD is based on taking two words out of context in a list of features that Samsung could have changed to not infringe in the German design patent infringement case. Specifically, Apple said that their patent claimed A+B+C+D+E+F+G+H, etc. with one of those things being rounded corners. If Samsung changed any of them - not having rounded corners, but having every other bit; or having rounded corners, but not having a flush bezel - then they wouldn't infringe. And that latter one is exactly what they did with the Galaxy Tab 10.1N: a device with rounded corners that does not infringe Apple's patent... so therefore, the patent clearly cannot cover "a device with rounded corners".
As usual, Richard Stallman has a great solution:
We should legislate that developing, distributing, or running a program on generally used computing hardware does not constitute patent infringement.
This will work because a very similar law already works in the medical field. Just like surgeons, who can safely ignore procedural patents to save lives, programmers and distributors of free software deserve complete patent immunity because their work is entirely gratis, and benefits the whole world.
Note that Stallman's solution doesn't include either of those requirements, and therefore lacks the same moral justification.
Re:Correlation != Causation (Score:5, Informative)
Another reason the backlog of patents has declined is that the USPTO is better funded, and has hired many more patent examiners. In 2005, there were about 7300 examiners. Today, there are more than 9500.
It seems to me that we need to compare the numbers here [uspto.gov]. So looking at 2005 as you mentioned, there were 417,508 patents filed. That means a per-examiner average of 57.19 patents. Fast forward to today, in 2012 there were 576,763 patents filed which gives us a per-examiner average of 60.71 patents. Instead of showing a decrease in the average patents indicating a more acceptable workload for examiners, we instead find that the opposite is the case. Given even the budget increase that provided for additional examiners, the workload on each examiner has been increased by ~6.4% since 2005 which should indicate that the approval rate should be in the same ballpark as it has trended to in the past. I can see no way to account for the 20% increase in patent approvals when the workload has also increased.
I browse the USPTO approved patents occasionally just for fun because seeing what gets approved is laughable sometimes. Ultimately I haven't seen any difference in quality of those approved at all but I do not evaluate a significant enough subset of the total approved patents to say with certainty that this is the case. I find it very hard to believe, however, that in the face of increased workload the quality of overall submissions has improved so drastically in 3 years that 20% more patents are being approved. With such massive datasets, a huge trend change like this almost never occurs. You can't tell me that out of 276,788 approved patents 55,358 of them just suddenly got better than previous years when the ruling in KSR v Teleflex that should have spurred these changes happened ~4 years earlier. Again, there should be a trend to the changes that shows an upward swing in approvals rather than this:
2007 - 484,955 Total Patents / 182,899 Approved Patents / 37.71% approval
2008 - 485,312 Total Patents / 185,224 Approved Patents / 38.17% approval
2009 - 482,871 Total Patents / 191,927 Approved Patents / 39.75% approval
2010 - 520,277 Total Patents / 244,341 Approved Patents / 46.96% approval
2011 - 535,188 Total Patents / 247,713 Approved Patents / 46.29% approval
2012 - 576,763 Total Patents / 276,788 Approved Patents / 48.99% approval
Ultimately, just looking at the numbers comparisons, it is almost as though every additional patent to the total number of patents since 2009 has been approved. For example, in 2010 there were 37,306 more patents than 2009 but there are !!52,414!! more approved patents. Taking the same comparison but going 2011 vs 2009 (because this is the year that shows the marked trend change), we have 52,317 more patents and 55,786 more approved patents. Finally, in 2012 vs 2009, we have 93,982 more patents and 84,861 approved patents (the first time since 2009 there have been less approved patents than the increase of actual submissions). What intrigues me most about these numbers is that we see a huge increase in the number of submitted patents during 2010 and on...I have heard through the grapevine that this is because of the slackening in standards at the USPTO to bolster the agency's numbers which has spurred the patent trolls but IANAPL (patent lawyer) so I wouldn't know anything about whether this is reality or not.