Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Isaac]

USPTO PARTNERS WITH OPEN SOURCE COMMUNITY TO EXPAND
PATENT EXAMINER ACCESS TO SOFTWARE CODE

Alliance will focus on software-related patents


The Department of Commerce’s United States Patent and Trademark Office (USPTO) has created a partnership with the open source community to ensure that patent examiners have access to all available prior art relating to software code during the patent examination process.

more at...

http://www.uspto.gov/web/offices/com/speeches/06-02.htm

[JSonnabend]

I saw a blurb on this in the New York Times, and of course, the Slashdot folks are frothing at the mouth, but I'm not sure what it really means.  Has anyone taken the time to actually figure any of this initiative out?  I'd love to hear.

- Jeff

[JimIvey]

I've been sort of following the indirect discussions re software and patents and open source.  I'll try to read between the lines in the notice....

The group agreed to improve prior art resources available to the USPTO;

First, it's not clear whether "the group" is the entirety of the meeting attendees or just the representatives of the open source community.

I read this as providing better access to non-patent software prior art for use by examiners.  

This is one of the major concerns of the open source community.  The hope of OS is to create something and set it free to benefit the world generally.  Much of the focus on OS development is on independent creation, to avoid copyright infringement.  The licensing aspects of OS is to prevent the software from being corralled and privatized for exclusive business use (think of a corralled wild stallion -- that's the mental image of the software taken and privatized by a corporation).  Unfortunately, there is no independent creation defense to patent infringement (hence, the disdain for software patents, in my opinion).  The other problem is that prior public use of OS software generally doesn't prevent subsequent patenting of the same thing -- due largely to inadequate searching facilities (and to inadequate documentation of the OS software itself, from a patent examiner's likely perspective).  If such a patent issues, it has a strong presumption of validity and woe is the OS developing community.

This statement seeks to address at least the inadequacy of searching tools to cover a substantial portion of publicly used and/or distributed OS software, and may attempt to addressed the perceived inadequacy of OS software documentation (not how to use it, which is reasonably well documented, but how it does what it does and to some degree, what is does exactly).

to develop a system to alert the public when USPTO publishes certain software-related applications so that interested parties can submit related prior art in accordance with relevant rules and law;

My guess here is that the OS community wants to know when there's a new software patent application for which to mobilize the troops.    I don't know who's going to build the tool, but I could probably do it myself (some script to perform a weekly search of published applications in a given group and at least a front page download).  In fact, I've already done something like that for other types of searches (downloads the front pages of all results on a patent search page).  I'd happily share that with the OS community (or any other community) under OS terms.

I think combining this tool (mine or another's) with inexpensive means for third-party IDS submission could go a long way to overcoming any limitations in searching tools for OS software.  This would allow the OS community to voluntarily provide search services for the PTO.

to explore developing additional criteria for measuring the quality of software patents.

This one is scary to me.  For some in the OS community, they wouldn't know a quality patent (software or otherwise) if it came up and bit them on the behind.  There's no reason to single out software from all other sorts of technology within the patent system.  All criteria for a good software patent should be applicable to all other types of patents.  As much as I'm sympathetic to the OS community, asking them which software patents are good is like asking a canary which cats are good.  

Personally, I think that 70%-80% of the concerns of the OS community would be addressed by a good way to effectively consider publicly used OS software as prior art during patent examination.  The first two items quoted above would represent good progress in satisfying this need.  Hopefully, the last item just means the PTO and the OS community will keep an ongoing and open dialogue to facilitate providing high-quality examination of software patent applications.

Regards.

[JSonnabend]

I think combining this tool (mine or another's) with inexpensive means for third-party IDS submission could go a long way to overcoming any limitations in searching tools for OS software.

This, I believe, would help fix our broken system immeasurably.  Create an "opposition" system, of sorts, akin to the opposition mechanism for tm's.  Even third party IDS' would be a big step in the right direction.

- Jeff

[Cibola]

Of -->This one is scary to me.  For some in the OS community, they wouldn't know a quality patent (software or otherwise) if it came up and bit them on the behind.  There's no reason to single out software from all other sorts of technology within the patent system.  All criteria for a good software patent should be applicable to all other types of patents.  As much as I'm sympathetic to the OS community, asking them which software patents are good is like asking a canary which cats are good. <--

the work on the patent quality index is probably an outgrowth of the patent quality problem proposed by Quillen and Webster, wherein it was once alleged that the USPTO grants patents on 97% of all applications.  Although even Quillen and Webster dropped that, one notes that Professor Wagner cited the 97% number in a law review published in Nov. 2005, so he may be laboring under a false impression of how bad the "quality" problem is.  Separately, there seems to be a US patent very similar to the approach taken by Wagner (6,556,992).  Thus, ironically, the approach taken to alleviate problems caused by neglect of prior art seems to ignore prior art pertinent to the approach.  

[JimIvey]

... it was once alleged that the USPTO grants patents on 97% of all applications.

Of course, I'm not telling practitioners anything they don't already know, but this number is misleading.  My very unscientific estimate would be that more than half of those applications as issued have undergone significant modification and probably substantial narrowing during prosecution before issuing as any patent at all.  I've seen applications (prosecuted by others) with no less than 6 amendments before anything issues.  To suggest that the Patent Office is laying down and allowing anything with such a statistic is so misleading as to be intellectually dishonest (and, I understand Cibola was just quoting another).  

My unscientific guesstimate is that fewer than 25% of application are allowed without substantial narrowing during prosecution -- perhaps fewer than 15% or even 10%.  

The reality is that it's extremely rare that an applicant's specific implementation is so exactly duplicative of prior implementations that the applicant can't get some level of patent protection, albeit fairly narrow at times.  That "extremely rare" circumstance is probably at or below the 3% suggested by the statistic.  More often, I find applicants determine that, although they could get very narrow protection in view of circumstances revealed during prosecution, the narrowness of the available protection does not justify the expense and effort of continuing the application.  I would guess that such determinations make up the difference between the "extremely rare" case I mentioned and the quoted 3%.  The remainder accept the substantially narrowed protection to allow the case to issue.

In short, anyone who thinks that they can get any application issued should be in this business and doing it.  There would be much demand for someone with such talents.  And, the fact that talents are required belies the quoted statistic.

Regards.

[Cibola]

Of Quillen and Webster's "97%" patent grant rate number, the important point to note is that it is not accurate, even though Professor Wagner may have cited it in 2005.  Some interesting discussion may be found at http://talkaboutbusinesses.com/group/misc.int-property/messages/65029.html.

Of Wagner's approach, which incorporates information from re-exams and litigations, one notes that the Eolas patent, which has survived litigation and re-exam, would probably ring out as a high quality patent, perhaps not the result expected by the open source folks.  Separately, the search for claim terms not defined in the specification as an indicator of "low" quality might be a bit deceptive; as one example, the term "melting point" might not be defined in the spec but would not be harmful to the validity of the patent.

[Cibola]

Further discussion of the deficiencies of the method underlying the "97%" [and "85%"] patent grant rate allegations may be found in jip.kentlaw.edu/art/volume%204/ 4%20Chi-Kent%20J%20Intell%20Prop%20186.doc.

One notes that many academics, including not only Wagner but also Lemley and Moore bought into the Quillen/Webster approach without reviewing the compelling evidence against it.   In the present situation, with the patent quality index, one hopes that there will be a more serious analysis of a proposal to identify high and low quality patents.    Although there may be a lot of traction on the patent "quality" issue, one has to avoid being drawn into research that amounts to telling people what they want to hear, as happened vividly in the publication by the journal Science of two fraudulent papers on human embryonic stem cells by Hwang Woo Suk.

[JimIvey]

Thanks for the link.  While I have the utmost respect for at least one of the pundits named in the article (the others I don't know), I have two major problems with the theory that a large percentage of applications allowed/issued equates to too many patents covering known things.

First, as I mentioned above, none of this takes into account that claims are allowed, not applications -- and that claims are often amended in significant ways (or broadest ones canceled while narrower ones issue).  I think a much more interesting statistic would be the percentage of claims are issued as filed (without modification).  And, to be a valid statistic, all claims depending from any modified claims are to be considered modified (even if the actual text of the dependent claim was not modified).  Of course, cancelation counts as modification unless an identical claim was later filed in a child application (continuation, divisional, CIP) or even added back to the same application (that happens).

My very rough guess would be 10% or less of all filed claims are allowed without modification of any kind.  Perhaps 15% are allowed without any substantive modification (I know, a subjective criteria that's hard to evaluate -- but that's the nature of the patent business).

So, the presumption that allowance of many applications means the Patent Office isn't a meaningful filter is just not valid from the quoted statistics.

My second problem is the logical leap from many applications issued to many applications improperly issued.  It's simply not a valid assumption.  

The closest thing to any kind of comprehensive study on this is a rumored quality assurance program within the USPTO in which they privately re-examine cases selected somewhat randomly to evaluate the performance of examiners.  I don't know the results (don't even know the program really exists), but I seem to remember hearing rumors of a confirming conclusion (positive evaluation) in at least two-thirds or three-quarters of all evaluated cases.

I don't think one can reasonably believe (without substantial doubt) that the quoted statistic is anything approaching conclusive proof of a patent system run amok.  I think significant corraboration is required before such a belief (as opposed to unfounded opinion) is reasonable.

Regards.

[Cibola]

In this discussion, we have at least two themes:  

-->first, what Quillen and Webster put forward in two papers (Q1 and Q2), with Wagner having adopted the 97% grant rate of Q1 and Lemley and Moore (and others) having adopted the 85% number of Q2.  

-->second, what Wagner is proposing to do with a quality index in the context of the cooperative venture among open source people, IBM (and others), and the USPTO.

Of the first issue, I agree with your points that the claims define the invention (which is missed in both Q1 and Q2) and that the number of unchanged claims in the prosecution of applications is less than 50%.  Discussion of this number came out in the Festo case, wherein the parties discussed the scope of the impact of the absolute bar.   Additionally, the presumption that the patent grant rate is an indicator of patent quality can be questioned.  Looking only at grant rate does not address the scope of the allowed claims, which can be distinct from the scope of the submitted claims.  In the final report by the National Academy of Sciences, the panel did not rely on the numbers of Quillen and Webster, but did rely on a perceived higher grant rate in the US relative to Japan and Europe to infer the presence of a patent quality issue in the United States.  The NAS panel did not contemplate the impact of different rules among the US, Japan and Europe.  
The patent quality issue served as common ground to involve many players dissatisfied with the patent system to get traction for patent reform, in the body of HR 2795.   However, disagreement over the injunction issue among major players killed off 2795.  Of course, 2795 was disconnected from the fee diversion issue, which also went nowhere.  In the end, the USPTO pretty much ended up where it started.

Enter the open source initiative.  Here, the PTO gets access to "free" resources to aid in the evaluation of a problematic area.  Not a problem.  But tacked on is a lingering vestige of the quality discussion, spearheaded by someone who thinks the PTO grants patents on 97% of all applications.   In one irony, using the proposed metrics (including re-exam and litigation success), the Eolas patent, a poster child of evil to many in open source, will look like a high quality patent.  [As an aside, note that the re-exam of the Eolas patent which has now concluded was initiated by the USPTO director (with a strong push by W3C), and not by Microsoft; we now apparently have a different re-exam.]  In another irony, the quality examination initiative proposed by Wagner appears to fall within the scope of an issued, in-force US patent.  If the open source folks use a patented approach to evaluate patents, doesn't that undercut their underlying claim to fame a bit?

[JimIvey]

... the Eolas patent, a poster child of evil to many in open source

Do any reasonable people consider Eolas the poster child for improperly allowed patents (claims)?  It withstood challenges in court (all the way up to the Supreme Court) and in the USPTO (and, as you note, undergoing a second challenge in the USPTO).  Not many patents withstand that sort of scrutiny.  While I haven't read the Eolas patent, it seems that it's of sufficient quality at the very least -- perhaps relatively high quality.  I certainly hope my patents fare as well.

If the open source folks use a patented approach to evaluate patents, doesn't that undercut their underlying claim to fame a bit?

I'm not sure I follow here.  I don't think the OS community ever asserts that they can deftly avoid infringement of all patents (if any such assertions are made, it must be out of pure ignorance).  My understanding of their perspective is that it seems you can't do anything without running afoul of patent infringement liability.  I would think a patent on evaluating patents would support their claims, not contradict them.  Perhaps I misunderstood your point.

I think the biggest problem in this whole issue is the insistence in treating all patents like commodities.  Parent are, by definition, unique.  Each one must be considered on its unique merits.  

Of course, economists hate that.  I heard a joke at a recent IP CLE conference.  I'll shorten it considerably....  To improve efficiency of a dairy farm, a panel of expert economists are called in to study the operation.  As the collaboration begins, some assumptions are agreed upon.  First: "assume a spherical cow of uniform density...."

Like cows, patents are not spherical and do not have uniform density.  Assumptions to the contrary are ... "inefficient" (evil and an anathema to any economist).

Regards.

[Cibola]

Of opinions on U.S Patent No. 5,838,906 (Eolas), the W3C's HTML Patent Advisory Group put together prior art which it claimed would establish that the Eolas patent is invalid and "should therefore be re-examined in order to eliminate this unjustified impediment to the operation of the Web."  The Group stated the object embedding technology being claimed by Eolas has been a part of the HTML standard since the early days of the Web.  The Group warned that changes enforced by the Eolas patent would have a "permanent impact on millions of historically important Web pages," including pages with non-commercial content or older material that is not generating revenue.  In a letter to the USPTO Director, Berners-Lee wrote:  "Removing the improperly disruptive effect of this invalid patent is important not only for the future of the Web, but also for the past.  The '906 patent is a substantial setback for global interoperability and the success of the open Web."  Karsten Schneider wrote:  -->To me the Eolas patent is the poster child for software patent abuse. This so called invention has been independently “discovered” by everybody who has implemented a modern browser. The patent has done absolutely nothing to advance the art (one of the primary reasons for patents). <-- See http://kas.felinity.net/ndate/2005/09/

Of the quality index, one question might be whether or not Wagner and others seek to license the patented technology or simply just use it.  So far, there has not been much discussion by Wagner of the existence of the patent.   Similarly, there has not been much discussion by Wagner of why he believes the 97% patent grant rate is correct, in view of the substantial attack on the approach by Quillen.

[JimIvey]

Whether a particular patent advances technology isn't really the final determination of the quality of a patent.  I've seen many clients extoll the virtues of their technology only to have the VC community completely ignore it.  Of course, eventually some company that gets funded comes up with the similar solution.  And, the funded company, having independently created the same solution asserts that the patent added nothing to the technology.

I don't know how many presentations in front of VCs you've made, but many times it seems like you're casting pearls before swine.  If the swine don't appreciate the pearls, it doesn't mean the pearls weren't valuable.  Like a contract, contribution requires both offer and acceptance of the technical teachings.  The inventor has control over the offer only, not the acceptance.  If the acceptance comes later, it's no fault of the inventor's.

As for the spot-on prior art of the early HTML standard, are you suggesting that Microsoft, with half a billion dollars on the line, couldn't find something as basic and fundamental as an HTML standard?  Perhaps they couldn't afford good experts or didn't have access to qualified software engineers.  Perhaps they thought they could save 1% of a $3,000,000 litigation budget by avoiding having any experts look at HTML standards for patent on web browser technology.

The notion that the validity of the Eolas patent wasn't adequately challenged in court by Microsoft just seems absolutely ludicrous.  The only plausible explanation (for a lame defense) is that MSFT didn't care about the half-billion dollars given MSFT's size.  I find that very hard to imagine -- especially with all the other patentholders in line with hands extended, palms up.  To paraphrase a Senator whose name I can't remember, a half-billion here, a half-billion there... pretty soon you're talking about real money.

Lastly, I haven't read the articles, but just because someone has a by-line doesn't mean that they know what they're talking about.  That includes me, by the way.

Regards.

[Isaac]

There are various measurements of quality, but they boil down to an index related to patents that the statistician thinks should not be granted.   The PTO is likely to measure quality in terms of patents that an examiner reasonably allowed or rejected under the constraints faced by the examiner.   Someone else might measure quality based on whether granted patents meet the definitions of patentability under the law.

Poster Cibola is basically suggesting that the invention in the Eolas patent is obvious independently of whether the legal standard suggests otherwise.   He is suggesting that independent reinvention by a number of different implementers suggests that an invention is obvious.    

Inherent in his "definition" of obviousness is a fairly high standard for inventiveness that IMO has already been shown to be hostile to reaching the objectives of the patent system.   Unlike copyright law which requires creativity and not sweat of the brow, the patent system does reward hard work and being first; no flash of genius required just small improvements over the existing technology.  

The other issue is that a reimplementation test is highly subject to hindsight.   Once you know that someone has invented X, it becomes much easier for a motivated someone to invent X.  In hindsight everything is obvious.

[Cibola]

Isaac and JimIvey may have misinterpreted my earlier remarks.  I was responding to the statement  -->Do any reasonable people consider Eolas the poster child for improperly allowed patents (claims)? <-- by furnishing evidence of some (perhaps reasonable) people who do consider the Eolas patent a poster child for a low quality patent.  I did not say anything about what I thought, so extrapolations to  "my" view of obviousness (or anything else) are unjustified from the text presented.    

One notes that W3C employed Pennie & Edmonds to write a paper attacking the claims of the Eolas patent in the director-ordered re-exam under 102 (not 103) grounds.   The combination of the smart technical people of W3C and the smart legal people at P&E did not go far towards invalidating the claims.  The examiner made an initial rejection under 103 and then dropped the arguments made by W3C/P&E entirely.

Microsoft attempted to utilize Wei's Viola browser as prior art, but the district court (strangelyand wrongly) applied abandonment principles to 102(a) art.    Although the Viola browser is not an anticipation, the failure of Berkeley to cite the art to the PTO, after Wei asserted its relevance prior to Berkeley's filing, makes inequitable conduct loom large.

Could Microsoft have done a better job?  Could W3C have done a better job?  Probably yes.  These were big guys who maybe didn't get the job done.  Turn the fact pattern around to what has happened with the Kamil patent, wherein the patentee attacks small guys who can't put up much of a fight (or Smucker attacking Albie's over the peanut butter and jelly patent).  Do litigation victories really prove the quality of those patents?  No.  Wagner's approach in  "objectively" counting noses out of context does not bring much to the party of patent quality.

Returning to the background of Wagner's quality index, this is an attempt to look at the past, in terms of why other patents were/were not invalidated, to predict via identification of variables obtained through regression what will happen to other patents (or even applications).   This approach is covered in a US patent issued in 2003.    Wagner has announced that his algorithm will be made freely available.   Wagner also has adopted the "97% of all applications are granted" number of Quillen's first paper, even though Quillen himself dropped the 97% in a paper published one year later, and years before Wagner adopted the 97% number.  This is way more than a byline issue or nice guy issue.