12:01 Tuesday

Pruning Priority to Prolong the Patent

"The examiner is utterly confused about the lineage of this application..."

Thus begins the intriguing prosecution history of a patent that issued on Tuesday.  (Yes, even three feet of snow and the PTO being shut down could not stop the weekly arrivals of issued patents.)

Here are the basics:

U.S. Patent No. 7,658,935  Methods and compositions for stimulating the immune system Inventor: Samuel Bogoch Filed: May 15, 2001 Lineage as filed (and published in U.S. Pub. 2002/0045187): This application is a Continuation-in-Part of my application Ser. No. 07/744,649 filed Aug. 8, 1991, and the prior applications of which it is in turn a continuation-in-part, Ser. Nos. 941,940, 852,200, 621,112, 553,075, 550,432, 450,404, 385,451, and 922,799, which should be incorporated by reference in the present application.  Lineage as issued: This application is a continuation of and claims priority to application Ser. No. 08/031,562, filed Mar. 16, 1993; which is a continuation-in-part of application Ser. No. 07/744,649, filed Aug. 8, 1991, abandoned, to which the instant application also claims priority.              Application Ser. No. 07/744,649 is a continuation of application Ser. No. 07/227,621, filed Aug. 3, 1988, abandoned; which is a continuation of application Ser. No. 06/281,883, filed Jul. 9, 1981, now U.S. Pat. No. 4,976,957; which is a continuation-in-part of application Ser. No. 05/922,799, filed Jul. 7, 1978, now U.S. Pat. No. 4,298,590, and application Ser. No. 06/019,078, filed Mar. 9, 1979, abandoned; application Ser. No. 06/019,078 is a continuation-in-part of application Ser. No. 05/941,940, filed Sep. 13, 1978, abandoned; which is a continuation of application Ser. No. 05/852,200, filed Nov. 17, 1977, now U.S. Pat. No. 4,196,186; which is a continuation of application Ser. No. 04/621,112, filed Oct. 9, 1975, abandoned; which is a continuation-in-part of each of application Ser. No. 04/553,075, filed Feb. 25, 1975, abandoned, application Ser. No. 04/550,432, filed Feb. 18, 1975, abandoned, application Ser. No. 04/450,404, filed Mar. 12, 1974, abandoned, and application Ser. No. 04/385,451, filed Aug. 3, 1973, abandoned. All applications listed above are incorporated herein by reference.

Commentary: If you blinked while reading that, you may have missed an important grammatical mark: the period (emphasized with added space above) ending the first sentence and showing that this patent claims priority ONLY to the '562 application and the '649 application, and NOT to any of the other eleven applications listed there.  And if you've read this blog in the past few weeks, you know immediately why that matters: if the priority chain had not been severed, this patent would have have had an expiration date of August 1993, which is almost eight years before it was even filed. Instead, this patent has over a year of life, plus some extra time for PTO delays. Good move.

Of course, there's more to the story.  By cutting the chain, all the previous links theoretically become available as prior art under 102. Why only in theory?  Because the first severed-off application ("621) was abandoned and unpublished -- and thus is unavailable as a 102(b) reference.  And the '621's parent, which blossomed into the '957 patent, issued in December 1990 -- less than one year prior to the August 1991 priority date of this week's patent. (Surprisingly, the Applicant did not raise the unavailability of the '957 patent when it was cited in a 102(b) rejection along the way, instead choosing to argue on the merits.) And 102(e) does not apply since the inventive entity is the same (not "by another").  In other words, a reference filed twenty years earlier (July 1981) and not in the priority chain is unavailable as prior art. To this, my brother would say, "Only in America. (or was that Don King...)"

But wait.. there's more.  Late in the prosecution, the Applicant was hit with a 112 written description rejection.  The response?  Amend the specification to import language from the grandparent '649 application, thus giving the needed support to the claims. On the surface, that seems fine -- the '649 application is in the priority chain, after all, and it was incorporated by reference. [Note, however, that as of October 2004, such an incorporation of "essential" material from an unpublished application is not proper. 37 CFR 1.57(c).] The peculiarity here, however, is that the '649 application and the '957 patent--not in the priority chain--essentially share the same specification.

Talk about having your cake and eating it. In beating back the 112 rejection, the Applicant relied on language from a twenty-year old document to provide support for his claims; yet the Applicant distanced himself from that same document in order to avoid early expiration.

And it all appears to be legit. Fascinating.

Army Had Received Two Previous Dead Patents

It gets worse...

It turns out there is more to the story of U.S. Patent No. 7,604,811, the Army patent that issued this week with an expiration date in 2004.  Incredibly, the '811 patent is the third pre-expired patent to issue from this same chain.

As explained in the previous post, the '811 patent claims priority back to 06/590,308, which was filed March 16, 1984.  The '308 application served as a priority basis for many other issued patents, including:

• U.S. Patent No. 6,844,010, filed July 18, 2000; issued January 18, 2005.

• U.S. Patent No. RE40,786, issued June 23, 2009 as a reissue of U.S. Patent 5,762,965, which was filed February 9, 1996.
  

Thus, both of these patents also expired in March 2004, which was prior to their issuance by 10 months and 5 years, respectively.

The facts here show that the '811 fiasco was no fluke. As was done with the '811 prosecution, while prosecuting the '010 patent the Army similarly filed a terminal disclaimer to circumvent a double-patenting rejection. In April 2004.  A month after the patent's expiration date.  (Another note on '811 -- it almost was saved with 550 days of Patent Term Adjustment, except there was over 600 days of patentee delay that offset it)

And like the '811 prosecution, the '786 reissue proceedings included substantial argument after the March 2004 expiration date (i.e., nine amendments and an RCE).

If that's not enough, there is also this piece of bad news for the Army: contrary to what I supposed in the previous post, you cannot use a reissue patent to delete a priority claim in order increase the patent's term.  MPEP 1405 makes this quite clear ("[term] cannot be extended through the filing of a reissue"), as does the language of 35 USC 251, which only allows a reissued patent for "the unexpired part of the term of the original patent."  In other words, there does not appear to be any way to resurrect these patents, even if they don't rely on any subject matter disclosed in the 1984 application.

Simply unbelievable.

[10/22 7:44PM UPDATE: There's more: the Army paid the four-year maintenance fee on the '010 patent in February 2008.  Should the PTO give a refund?]

Be All You Can't Be: Army Patent Ambushed by Friendly Fire

You may recall the brouhaha back in the Reagan era when it was discovered that the Navy had purchased some very expensive ($600) toilet seats.

I remember comedian Emo Phillips around that time joking about his letter to the IRS, noting that the value of the paper clip he’d used to attach his check should cover any deficiency in his return, based on Pentagon spending figures.

Today, it appears the Army may have bought some much more pricey toilet paper to match the Navy’s purchase.  How?  By obtaining a patent that expired five years before its issuance.

Here's the relevant excerpts from the patent’s face:

U.S. Patent No. 7,604,811 Oral-intestinal vaccines against diseases caused by enteropathic organisms using antigens encapsulated within biodegradable-biocompatible microspheres Assignee: The United States of America as represented by the Secretary of the Army Issued: October 20, 2009  Filed: December 10, 1996 Continuation of application No. 08/396,986…which is a continuation in-part of application No. 06/590,308, filed on Mar. 16, 1984, now abandoned…

Because this application was filed after June 8, 1995, its term is 20 years from the filing date of the earliest reference specifically referred to for a priority claim: March 16, 1984.  In other words, this patent that issued today actually expired March 16, 2004 – five and half years ago.  35 U.S.C. Section 154(a)(2).

    "But maybe there is Patent Term Adjustment available…" Nope. The application was filed prior to May 29, 2000.  Nothing there.

    "But maybe there is Patent Term Extension available under 37 CFR 1.701…" Nope. There was no interference, secrecy order, or appellate review.  And even if it were available, it could only extend up to five years – to March 2009.

    "But maybe there is an extension pending FDA approval of the vaccine under Section 156…" Nope.  That would also be a maximum extension of five years under Section 156.

This patent is DOA.  If anyone has a good explanation as to why prosecution of this patent was pursued, please share.  (And note that if your answer is "to get the app published," there were other, less burdensome routes available, such as voluntary publication under 1.221.)

So let's see where your tax dollars went over the last five years in pursuing this stillborn:

• Four responses to non-final office actions
• Three responses to final rejections, including two RCEs
• Two Examiner interviews
• One terminal disclaimer filed
• One notice of appeal
• One appeal brief
• One petition to correct priority
• One issue fee payment

That's over $5,700 just in PTO fees.  In this case, those fees are effectively a shifting of government funds from one entity to another.  But it’s hardly an efficient use of resources.  And I wouldn't be at all surprised if the Army spent three times this amount for the attorney's corresponding time.

And of course, there are several interesting and crazy twists:

• First, you did read that list correctly.  The applicant filed a TERMINAL DISCLAIMER in May 2006 – a year after the patent had already expired –to obviate a double patenting rejection.  That's like a baseball team staging a press conference to announce it's not exercising the option to keep the player who retired the year before.

• Second, the Army filed a petition in October 2007 to correct the priority claim to get around an obviousness rejection.  The cited art was a PCT publication from overlapping inventors, and the petition corrected the priority claim to include the underlying 1991 U.S. application of that PCT app.  That is a different app from the fatal 1984 application. But the petition clearly includes the chain going all the way back to the 1984 application, which appeared as a priority document in today’s patent app as originally filed.  In other words, if it wasn’t apparent up to that point (hard to imagine), the Army filed a petition two years ago which on its face evidenced that any resulting patent would not be entitled to any term.

• Third, and perhaps most interesting, the priority claim to the 1984 reference may not have been even necessary.  In the last of the final rejections, the Examiner noted explicitly that the 1984 application "do[es] not teach the specific CFA/I peptides employed in the vaccine and method of the instant claims.  Accordingly, the priority date of the instant application is the filing date of its immediate parent, U.S. Patent Application 08/396,986, filed 3/01/95."

This raises a question: if an application is a continuation-in-part and none of its claims are supported by the ancestor application, should its term be determined by the date of that ancestor app? (Set aside that it makes no sense to file such a C-I-P…)  I don’t think so.  The language of the statute is clear that it is the “specific reference” to the earlier app that determines a patent’s term.  Indeed, the Army could have disclaimed priority to the 1984 application and removed that specific reference.  See MPEP 201.11 III(G) (which amazingly anticipated this type of post-GATT situation).  But it looks like now they’d need to get it reissued. And the Army would probably be estopped from disclaiming priority earlier than April 1991, the date of the underlying PCT application.  Would it take less than a year and a half to push a reissue through? And it doesn't feel right that the Army should serendipitously benefit from the examiner's unsolicited priority statement, either.

• Lastly, and true to form with common perception of government efficiency, care to guess what the Army did prior this patent's issuance?  That's right: it filed a continuation application.  

Sure, it's not Charmin.  But really, how abrasive do you think that official U.S. Patent is?  I'll trade you a couple paper clips for it...

Death is no obstacle: Another patent for Lemelson

Jerome Lemelson died almost exactly 12 years ago, and today he received yet another patent. This one is the fourth -- and last -- in a chain of continuations going back only to 1996.  That's forty years less priority than his immediately previous issuance, U.S. Patent No. 7,343,660. Notably, today's patent was submitted with a request for non-publication. Also notable is that Lemelson's patent attorney is a co-inventor.

According to his foundation's website, Lemelson submitted about 40 patent applications in the year before he died. If their progeny were submitted with similar NPRs, we still would have no idea if we've seen the last of them.  We do know that Lemelson now has 48 patents that were applied for after his death.

Of course, the value of these submarines is shrinking daily in a post-GATT world, where term is measured from filing rather than issue date.  And there is that doctrine of prosecution laches that Lemelson famously helped revive.

But today's patent still has about 6.5 years of term left. I'm not aware of any facial recognition systems like this in current vehicles, yet I'd think that six years of potential damages would be more than enough to make litigation worthwhile if another sub from Lemelson's final fleet were to surface in the right waters... We'll have to wait and see.

Claim 1 is presented below.

U.S. Patent No. 7,602,947 Facial-recognition vehicle security system Inventors: Jerome H. Lemelson & Louis J. Hoffman 1. A security system for a movable vehicle provided with an enabling system, the security system comprising:  (a) an electronic camera positioned to scan at least a portion of the face of a person located at the driver's station of the vehicle and structured to produce image signals;  (b) an electronic facial-recognition system responsive to the image signals and controllably coupled to the enabling system; and  (c) an electronic storage device coupled to the facial-recognition system to store a set of image signals, such that the facial-recognition system activates the enabling system only if signals representative of select characteristics of the scanned face of the person derived from the image signals produced on the camera's output match a member of a set of signals stored in the electronic storage device, each member representative of select characteristics of the scanned face of an authorized operator.

Jaywalking across the 101

Jay Walker knows patents.  The Connecticut entrepreneur, best known for creating priceline.com, is listed as an inventor on over 350 issued patents, and a few hundred more published applications.  A sizable portion of these are squarely within Class 705; if you actively prosecute e-commerce business method patents, there's a good chance one of his applications or patents has been cited as prior art against you.  Walker created an invention think-tank company, Walker Digital, to "incubate" these ideas and businesses based on them.  Apparently those businesses, and a couple hundred million dollars in licensing deals, have done well enough for Walker to amass a small fortune, including a pretty amazing library in his home.

Walker has been in the business method game from the start, and for that he's been the subject of controversy.  He was outspoken with excitement a decade ago after the State Street and AT&T decisions, and he has written the USPTO with his opinions on ways to improve the patent process.  Some have speculated as to how much value the Walker Digital portfolio may have lost in light of the Bilski decision last fall.  The extent to which this may be true is not clear, though priceline.com apparently felt it necessary to mention the case in the "Risks" section of its recently-filed 10-K.

What we do know, however, is that Walker Digital today was issued a patent that has a story worth telling.

Key terms are in bold.

U.S. Patent No. 7,523,045  Method and apparatus for a cryptographically assisted commercial network system designed to facilitate and support expert-based commerce Assignee: Walker Digital, LLC 1. A computer implemented expert matching method for managing communications between an expert and an end user, comprising:  receiving a first expert qualification corresponding to a first expert;  selecting a second expert that has a second expert qualification, the second expert qualification being higher than the first expert qualification;  transmitting the first expert qualification to the second expert; and  receiving a signal from the second expert that indicates approval of the first expert qualification.

Commentary: Take out the words "computer implemented" and the only possible Bilski machine is that "signal", which won't cut it (it's not even statutory subject matter itself, per Nuijten).  And there's no transformation, either.  The Notice of Allowance came in February 2009, well after Bilski had been soaked up by the examining corps, particularly in the 3600 TC and class 705.  Moreover, this claim actually made it through the BPAI in October 2008, two days before the Bilski decision. [Does anyone know why the decision is not available on the BPAI FOIA site? It's in the patent's IFW...]

So what happened here?  For starters, know that this patent has the longest prosecution of any patent to issue today, filed in July 1998.  First office action (of seven total) came nearly 10 years ago, in August 1999.  By about OA #3, the Examiner -- the same one throughout the entire prosecution, amazingly -- decided to get kind of cute.  Instead of finding a reference that taught the disclosed steps, he pointed to internal patent examination procedure.  For example, a prosecution may include a primary examiner who is a first expert, and who in turn consults his SPE (a more highly experienced expert), etc.  In later office actions, the Examiner actually cited the MPEP directly as codifying this "procedure."

And that's what went up to the BPAI in October 2008: an application with an examiner's cutesy rejection where the MPEP itself was the only cited prior art against the claims.  The Board wasted little time in calling the Examiner's scenario "no more than a speculative hypothetical situation, rather than evidence of anticipation."  It continued, "More critically, the Examiner's rejection is over the MPEP, not the business practices within which the MPEP is used."  Because the MPEP itself did not describe the claimed steps, there was no prima facie case of anticipation.  Reversed. (There is a 103 reversal, as well, with respect to other claims.) 

That still doesn't explain why there was no sua sponte 101 rejection, as we've seen many times in the last few months (101 was never raised during prosecution).  My guess is that the Board felt it was important to do some public smacking here (fresh in my mind is the Passover seder's story of the 'wicked son,' who asks a smart alec question for which we are instructed to "hit him in his teeth").  They didn't want to soften the blow with another rejection, turning the previous rejection into "harmless error".  Thus the rebuked Examiner, with tail between his legs, allowed the claims to issue in February. [Question: could he have raised Bilski at that point, even if he wanted to? Or were his hands tied with the current record?]

So Walker appears to be the recipient of a gift from the BPAI, who chose to remain silent on 101 in order to teach a lesson.  Had their decision come a week later, post-Bilski, this may not have been so easy for the Board to do.

The practitioner's takeaway practice tip: if faced with an examiner too smart for his own good, (and we've all faced them...), it's worth pushing up the ladder.

[4/23/09 Postscript: The parent of this patent, U.S. Patent No. 5,862,223, was one of several patents singled out by Judge Mayer in his Bilski dissent as examples that "Patents granted in the wake of State Street have ranged from the somewhat ridiculous to the truly absurd."  (Slip op. at 12).  Another Walker patent, 6,119,099, also appears in that list.]

More Bard's Tale

For those who were intrigued by my post last week on the 35 year patent battle between C.R. Bard and W.L. Gore, there is a concise, high-level summary on the AmLaw Daily from Thursday.  Sounds like my hunches about the case's complexity weren't even close.  Some highlights:

• Total damages will probably exceed $600 million
• File wrapper is 10 feet high
• 30-year old perjury confession 

You can read the rest for yourself.  I'll wait for the novel to get more details. 

A Wild Ride on the 101

Speaking of old patents, there were two more pre-GATT patents issued today: U.S. Patent Nos. 7,514,397 (Trimeris) and  7,514,083 (University of Melbourne).  Neither appear to have particularly interesting prosecution histories.

On the flip side, IBM again has several Accelerated Exam patents today, using their single method claim technique that we’ve seen before.  Two of them sound pretty interesting: U.S. Patent No. 7,516,220 “Method and system for detecting and deterring robot access of web-based interfaces by using minimum expected human response time” – perhaps the PTO can use this instead of that darned reCAPTCHA; and U.S. Patent No. 7,515,899 “Distributed grid computing method utilizing processing cycles of mobile phones”.  [Didn't Bale do something like this in the Dark Knight? Could Batman be prior art?  Filmed partly in my office building, btw.]

But there is one patent that issued today that is not quite old enough, from 1995.  It is our first triple loser:

U.S. Patent No. 7,516,172 Method for finding quotient in a digital system Assignee: United Microelectronics Corp. Filed: August 2, 1995 1. In a system for digital information processing, said system having a memory, a method for generating data representative of a quotient Q=a.sub.0a.sub.1a.sub.2 . . . a.sub.b from data representative of a divisor Y=y.sub.1y.sub.2 . . . y.sub.n and data representative of a dividend X=x.sub.1x.sub.2 . . . x.sub.a, comprising the steps of: (a) aligning the first non-zero bit of X with the first non-zero digit of Y; (b) defining a signed-digit partial remainder series R.sub.i where R.sub.0=Y, a first sign series of the partial remainder S.sub.i where S.sub.0=0, a second sign series of the partial remainder S.sub.ri, a quotient bit series a.sub.i, and a counter i beginning from zero; (c) subtracting X from R.sub.i which yields next signed-digit partial remainder R.sub.i+1; (d) setting the sign of R.sub.i+1 to S.sub.ri+1; (e) setting the result of exclusive-OR of S.sub.i and S.sub.ri+1 to the true sign of the next remainder S.sub.i+1; (f) setting a.sub.i to 1 if S.sub.i+1=0 or R.sub.i+1=0; (g) setting a.sub.i to 0 if S.sub.i+1=1; (h) inverting the signs of all digits of R.sub.i+1 if S.sub.i+1=1; (i) shift R.sub.i+1 left by one bit; (j) adding 1 to i; (k) repeating steps (c) to (j) until i reaches a predetermined value or R.sub.i+1=0; and  (l) storing in said memory as said data representative of a quotient, a quotient resulting from step (k).

Let’s begin with Loser Reason #1: the filing date.  Because this patent wasn’t filed 8 weeks earlier, it will  expire 20 years from its priority date -- 2014 -- rather than 17 years from today.  That’s 12 years of term that just vanished.  (Actually, the patent will get 4.5 years tacked on due to a BPAI decision... see below)  Worse, there’s no apparent reason why this application couldn’t have been filed by June 7 – it was a continuation from a January 1994 application.

Loser Reason #2:  The patent was allowed in 2002.  So even with the post-GATT filing date, this application sat an extra seven years after it was allowed because of an alleged error with the PTO not sending the NOA to the new attorneys.  Two petitions later, the abandonment was withdrawn and the patent issued.

Loser Reason #3: Bilski and the method roller coaster.  This is a slam dunk under the Machine or Transformation test.  But this patent's story is fascinating because it spans and discusses a good chunk of the law that the Federal Circuit discusses in Bilski.  Note that this was a pre-State Street case.  The Examiner wasted no time (only about 7 months) to reject the claims under 101 citing the Freeman-Walter-Abele test.  The briefing argues over Alappat.  By the time the case was eventually ruled upon by the BPAI, State Street had been decided.  The Board reversed the Examiner under that test (note: PAIR mistakenly says this 8-15-2001 affirmed the examiner). 

But wait, there's more.  The Examiner requested reconsideration of the BPAI decision.  While I know this practice of requesting reconsideration is common when an applicant loses an appeal (though I've never seen a successful request), I have to admit I was unaware that an Examiner could do this, too.  Who knew?  It's pretty rare, apparently.  This cost about nine months of unrecoverable patent term (because this app was filed before 2000, Patent Term Adjustment is not available; only Patent Term Extension.  Yes, for you non-practitioners, these terms are actually different things. Never confuse PTE with PTA...reminds me of this discussion).

How about that for a triple-whammy?

One more patent to discuss today, an unlikely candidate:

U.S. Patent No. 7,513,005 Washing method for a dishwasher, and dishwasher for carrying out such method Assignee: BSH Bosch und Siemens Hausgersete GmbH 1. A method for washing and drying items in a dishwasher, comprising:  executing a washing program during which a washing liquid is applied to the items in the dishwasher; and  following completion of the washing program, executing a drying program such that, in response to both an interruption of the drying program resulting in the dishwasher no longer being in a program run state and a thereafter following restoration of the dishwasher to its program run state, a parameter value is measured and, in a first given circumstance, the drying program is resumed if a selected one of a predetermined deviation and an absence of the predetermined deviation from a nominal value is detected and, in a second given circumstance, a fresh cycle of a washing program and a drying program is initiated if the other of the predetermined deviation and the absence of the predetermined deviation is detected.

Some questions for the crowd:  What if it is a "general purpose" dishwasher?  Would that be a "particular" machine?  Is it particular simply because it manipulates tangible articles from a statutory class? And does it even do that?  Isn't what's being claimed the program that controls the washing, and not the washing itself?  And do you think this passes the transformation test, too (from dirty to clean; wet to dry)?

Wait! Don't throw that thing out!

To anyone who has read my earlier posts, my fascination with "old" patents (i.e., patents with very long prosecution histories) won’t come as news.  Often these old inventions, despite getting a fresh 17 year monopoly when they issue, may have lost much of their latent value through obsolescence (if that secrecy order is eventually removed, the thing probably is not so important anymore...), or perhaps changes in law (think Bilski, or regulatory changes affecting applicability).  But this week we were given about 371 million reminders why some of these old inventions really do matter.  That’s the number of dollars W.L. Gore must pay C.R. Bard for infringing U.S. Patent No. 6,436,135, after the Court (D. Ariz.) doubled the jury’s award.

Take a close look at the filing date – it’s not a misprint. 1974.  And the patent issued in 2002 after an epic prosecution battle that included a 15 year interference.  (Only the transaction history is available on PAIR.  I imagine the file wrapper is ginormous)  So we’re already at 35 years, and I’d be shocked if it ended here.  Can you imagine taking a case as a first-year associate, and to have that case with you the rest of your career?  Without knowing much more than this (and I really do not know much more than this), I would bet that there’s a movie script or two lurking in there.  I'd be curious to hear some war stories from any of the presumably dozens of attorneys who must have been involved in this patent's 28 year prosecution.

There’s another factoid about the ‘135 patent, but I’ll save it for a future date.  And I should have another post coming later today reviewing some of today’s issuances.

Biggest. Week. Ever.

Although I haven't done a full analysis to back up the following statement, I would feel safe betting my kids' 529 plans on it: More patent applications were filed the first week of June 1995 than any other week in USPTO history.  How much more, you ask?  Consider the following:

• Number of utility patents issued for apps filed 6/1/1995 to 6/7/1995: 20,557
  
• Number of utility patents issued for apps filed the previous week: 2,476
• Number of utility patents issued for apps filed the following week: 1,564
• Number of utility patents issued for apps filed in all of 1995: 144,515

Staggering, no?  If you are looking at an issued patent that was filed in 1995, there is over a 14 percent likelihood it was filed in one particular week of that year.  But it gets better: half of those 20 thousand patents were filed on a single day: June 7, 1995.

Of course, this should not be shocking to most patent attorneys, who know that June 8, 1995 marked the GATT changes regarding the term of a patent: "17 years from issuance" went to "20 years from earliest claimed priority date".  The mad rush ensued, particularly for anyone with an application claiming priority back a few years or so.  (Jerome Lemelson has thirteen patents that were filed that week; Ronald A. Katz has nine).  I keep imagining what the post office looked like that evening...must have had an April 15th-like feel.  (if you have war stories, please share).

But what is more interesting is watching patents continue to issue that were filed that week nearly fourteen years ago.  So far in 2009, three such patents have issued.  The first,  U.S. Patent No. 7,491,205 issued on February 17, 2009, and will expire in 2026. Had it been filed a day later, it would have expired before it even issued -- it claims priority back to 1988. The other two, U.S. Patent Nos. 7,501,487 and 7,500,433, issued on March 10.

The chart below shows the very, very long tail of tricklers.

Fewer than five such patents have issued each quarter since Q3 2007.  One wonders when the tail will actually stop.

Obviously, when a patent takes this long in prosecution, there is almost certainly a story.  Typically, there is some up-and-down with the BPAI through appeals or interferences.  Sometimes there are just long, unexplained delays at the PTO (such as files that are mysteriously "lost").  But if any of those stories are particularly compelling, you'll be able to read them here.

Bilski Watch - Midnight Edition

Haven’t even done the “real” search and already have a winner.   From IBM, of course. 

U.S. Patent No. 7,487,425 “Low cost symbol error correction coding and decoding”

1. A method for encoding input digital electrical signals into blocks of n bits and subblocks of at most b symbol bits, said method comprising the steps of:

generating check bit signals from said input digital electrical signals so that said check bit signals satisfy parity check conditions determined by a binary parity check matrix derivable from a matrix of the form H=[H.sub.1 H.sub.2 . . . H.sub.N] wherein each symbol column, H.sub.j, j=1, 2, . . . , N, is a column of M disjoint binary submatrices having b rows and at most b columns and wherein at least one of said N symbol columns has been modified to provide a minimal number of ones in said symbol column; and

combining said check bit signals with said input signals so as to produce a block of n signals having redundancy for improved error correction and detection.

I don’t think there is much question that this kind of abstract mathematical matrix manipulation fails the machine or transformation test (or perhaps even Nuijten, for that matter).  But what makes this case so delightfully ironic is this:  it was filed in May 1993, claiming priority to 1989.  And here’s more: there was a first action allowance of 7/20 of the claims in April 1995, and 19/20 of the claims were allowed by December 1996.  IBM missed the first office action in the mail and fought the second one with an unsuccessful appeal.  Most likely they kept pushing this, rather than just cutting their losses, because the patent term would be a pre-GATT 17 years from issuance.  But they weren’t planning on Bilski…

So, instead of having a patent that could have been in force for 13 years already, by greedily trying to extend patent term they now have a patent with method claims that are probably worthless. (There is an apparatus claim, but it probably suffers the fate that the BPAI has been handing out lately as covering unpatentable subject matter).

Moral:  As they say in football, “Don’t take points off the board.”