A Simple Question about Patent Quality

    By Kevin E. Noonan --

Let's put our cards on the table.  The zeitgeist in the patent world these days is that the U.S. Patent and Trademark Office engaged in a frenzy of granting "bad" patents over the past few years, and that the issue of patent "quality" has caused a crisis in the Patent Office that cries out for "reform" (see "The IPO's Marc Adler on Patent 'Quality'").

Since "one man's ceiling is (frequently) another man's floor," there is the distinct possibility that "quality," like beauty is in the eye of the beholder.  "Everybody knows" that there are bad patents out there, but so far as I know, what "everybody knows" is in the sphere of anecdotal evidence (like many herbal cures, UFO sightings, and urban myths).

So we here at Patent Docs are posing a simple challenge for our readers.  We are inviting our readers to send us the Patent Number for any patent granted between U.S. Patent No. 7,296,299, which issued on November 13, 2007, and U.S. Patent No. 5,296,299, which issued on March 22, 1994, that you believe is a "bad" patent.  That's a range of two million and one patents.  It would help if the submitter told us why he or she thinks the patent is a "bad" patent, but we are most interested in obtaining as many allegations of "bad" patents as we can.

We do not want to ask for any similar number of "bad" applications for a number of reasons.  First, some bad applications are easy to spot, and to dismiss - perpetual motion machines, laws of nature, and clearly unpatentable subject matter for example.  Also, every American has the right to apply for a patent, even on the most simple device, and nothing in the statute permits the Office to prevent a patent application, however looney, from being filed.  (And, of course, many perfectly good inventions look a little looney when they are first proposed - talking through wires, indeed!).  Finally, bad applications aren't identified as the problem, bad patents are, and that's where we would like to keep the focus of these efforts.

We can't predict the number of responses we will get, but we would like to see what we get once our readers get past the "peanut butter and jelly sandwich" patent and the "leaf collection bag shaped like a pumpkin" patent and the "disembodied head" patent.  Also, if what "everybody knows" is correct, we should find more and more "bad" patents issuing between 1994 and 2004 or so, with a significant decline over the past two or three years (if the Patent Office's calculus is correct that the number of "bad" patents is disproportionately reduced as overall patent allowance rates fall).

We cannot, of course, be exhaustive.  But we have about 900 e-mail subscribers and get over 5,000 hits on some days, so there should be enough people reading these posts that we get a representative number of submissions.  Just to be "up-front" with our own bias, we remember a website called "Bounty Quest" that offered $10,000 for prior art sufficient to invalidate patents identified on their website, and as far as we know the proprietors didn't go broke giving out their bounties.  We suspect we will not get 2,000 "bad" patents from our survey (which would be 0.1%), and frankly if we do this number would be 30- to 50-fold better than the current Patent Office quality metrics for preventing "bad" patents from issuing.

Happy hunting.

Comments

[NIPRA Anonymous]

Unfortunately, information on patent validity/invalidity is usually client confidential, which prevents the public dissemination of such information.

[Kevin E. Noonan]

Dear NIPRA:

I think you missed the point. We want to know what patents in the two million and one target sample YOU think are "bad." We don't care if that is just your opinion (although in that case we'd like to know why you think so), or if the patent has been invalidated in whole or in part, or if it seems self-evident. We're trying to determine whether there really is a "bad" patent problem or not.

So we don't need client confidentiality here. Just let us know what you think.

Thanks for the comment.

[Kill Ugly Patent]

The poster child for the rotten biotech patent:

http://www.simonsjunkdna.com/dr__c__nottenburg.htm

5,612,179

A method for detection of at least one coding region allele of a
multi-allelic genetic locus comprising:

(a) amplifying genomic DNA with a primer pair that spans a
non-coding region sequence, said primer pair defining a DNA
sequence which is in genetic linkage with said genetic locus and
contains a sufficient number of non-coding region sequence
nucleotides to produce an amplified DNA sequence
characteristic of said allele; and

(b) analyzing the amplified DNA sequence to detect the allele.

Read the prosecution history and the claim construction briefs for the most vivid examples of goalpost moving and dust kicking you'll ever see.

[Kill Ugly Patent]

By the way, Kevin: setting this up as an exercise in percentages missing the point. It's the absolute number of bad patents that is the problem. And the thing is with patents like the '179 patent: they were examined very "carefully." There are equally cruddy patents that were reexamined one or more times.

Now consider all the patents that issued that were not examined repeatedly by Examiners who, for the most part, received Ph.D.s in a fairly rigorous discipline.

This exercise is not going to prove anything except that most people who read these sites don't have time to wade through the PTO to find garbage patents (and most practicioners are not going to send you the patent numbers of garbage they themselves prosecuted!).

http://www.freepatentsonline.com/crazy.html

Provides a short list.

[Kevin E. Noonan]

Dear Kill:

Thanks for the '179 patent. I'll look into it.

And I do think the percentages matter (if only because it's these types of statistics that the PTO and patent "reform" crowd use). I think it is inevitable that some objectively "bad" patents will be granted. It would surprise me if they didn't - this is a human endeavor and that's just the entropy of the system; mistakes happen. But I do think the numbers, and the specifics of the prosecution histories, will tell us something about how "bad" patents are granted. It's only when we understand the how that we can prevent or at least minimize them.

And I want to see what we get to answer some questions about that the submitters think are "bad" patents. I want to see if we can draw any conclusions from the material.

Thanks for the submission and the comments.

[Wondering]

Would you count all of the patents that have been held invalid or unenforceable in whole or in part as bad patents? those could be found in court records.

[A]

US6608779 http://www.google.com/patents?id=H3sOAAAAEBAJ

This has been done since at least the 80s and the claims are worded so as to read on every memory known to man.

[Kevin E. Noonan]

Dear Wondering:

That is certainly part of the plan, but not what we need our readers to supply.

Thanks for the comment.

[B]

I know your expertise is in biotech, but here is as family of patents that pretty much covers VOIP as it is known today:

US 5,604,737
US 6,240,085 (CON)
US 7,167,467 (CON)

All are issued to: Hitachi Ltd.

These patents are overly broad and cover VOIP at large: Vonage; Verizon; AT&T, etc. all are at mercy of these broadly issued patents.

Who examined these patents? Ajit Patel. He is a primary exmainer in this art unit, and has being doing this for over 15 years.

What is the prior art? US Pat. 5,341,374, among others, anticipates each and every claim of the Hitachi patents.

Ironically and sadly, the '374 priot art patent, was also examined by the same Mr. Patel several years earlier.

There's simply no excuse for the issuance of these patents.

Mr. Patel should be fired.

[cm]

While it is admirable that you are looking into patents that were granted and should not have, please not forget that there is the whole body of "patents that should have been granted, but did not."
Many inventors, when repeatedly receiving outrageous rejections, run out of time, money, patience and give up.

[NIPRA Anonymous]

"Who examined these patents? Ajit Patel. He is a primary exmainer in this art unit, and has being doing this for over 15 years."

B, did he examine all these patents or just the last '467 CON? The '737 and '085 patents list him as neither the primary nor the assistant examiner. Also, did 13 years elapse between when he examined the '374 and '467 patents? (Your memory must be better than mine then to judge him so, and the draconian response you suggest sounds a little like corrective measures that the present USPTO management would see fit to undertake to eviscerate the corps.)

Having been an examiner and having issued a patent in which all claims were later found to be invalid by reexamination (based on Japanese art), absent additional evidence, I disagree with your conclusion about what should happen to Mr. Patel.

By the way, the stark reality of today's USPTO is that when CONs of allowed applications come in, the CONs are considered "gravy" (i.e. for counts) and they are not typically re-searched and re-examined. To the extent that the claim scope reasonably corresponds to that of the issued patent, they are cursorily allowed (otherwise the USPTO would be effectively rendering an opinion that the previous patent was invalid, which it is not permitted to do absent a reexamination... though a rejection of claims which correspond in scope to claims in an issued patent may be made with a Group Director's approval).

P.S. I'm sure I made many more errors than the one I mentioned when I was an examiner - that doesn't mean I was doing a poor job in the time allotted.

[KR]

I only see one kind of "bad patent," and it's the kind that is enforced knowing it will obtain nusiance value in a settlement.

[JimD]

How about 6,561,976
Method of avoiding giving thalidomide to a pregnant woman ... by checking to see if she's pregnant. (A computerized database, evidently, is not the obvious way to do this.)

[David Safran]

Claims were found patentable in 6,006,164 on the basis of an amendment making a wired connection to a computer into a wireless one. This patent is currently undergoing reexamination.

In 6,430,019, the applicant inappropriately contended that 35 USC § 103(c) required withdrawal of a commonly owned patent that was cited under 35 USC § 102(e) in a §103 rejection, and the Examiner incorrectly withdrew the patent as a reference on that basis and approved the application. However, the Examiner evidently was unaware of the fact that the application fell into the category of cases for which § 102(e) prior art is not disqualified under § 103(c),i.e., where the patent issued before December 10,2004 on an application that was filed prior to November 29, 1999.

[Rajiv]

Perhaps a better way would be to find out patents that have been granted in the US but have been rejected at EPO and/or JPO for fundamental issues like obviousness etc.

Simple % terms will not yield the desired "bad" patents. It will merely be a collection of patents that people think are "bad" in their field.

[Kill Ugly Patent]

"please not forget that there is the whole body of "patents that should have been granted, but did not."

Right. And all the incredible patents that were never even filed! Let's not forgot those, either.

[B]

NIPRA Anonymous:

My apologies about the oversight regarding the examiner of the '737 and '085 patents.

BTW, Hitachi is aggresively pursuing additionaly continuations (have about 6 or 7 pending).

Mr. Patel was about to allow an additional CON appl. 11/319,737 (subject to a terminal disclaimer), which included far broader claims, than any of the earlier patents. It is only after the applicant filed an RCE with an IDS disclosing the '374 patent (I guess they got alerted to this piece of prior art), that Mr. Patel issued a 102 rejection for all the pending claims.

The applicant has now entered an amendment with a new limitation, seemingly trying to overcome the rejection. If Mr. Patel carefully reads the entire '374 patent, he will notice that this new limitation is also disclosed a number of times in the '374 patent.

I will be watching this case, to see how Mr. Patel responds.

If the examiners are not guilty, in view of their narrow time allotments, then it is the higher-ups that are guilty.

In any event, I think an examiner should be applying heightend scrutiny when he notices an application with far overreaching claims that are about to be issued.

Mr. Patel - a person skilled in the art - should readily recognize that he is about to issue a patent that will esesentially render almost every VOIP phone company an infringer of such patent. There's so much he can do to ensure that such patent passes the minimum test of not being anticipated by a U.S. patent in the same art unit - especially one that he himself examined.

He shouldn't be sitting back, relaxing, and taking in another "count".

I stand by what I said - that he should be fired.

[NIPRA Anonymous]

"If the examiners are not guilty, in view of their narrow time allotments, then it is the higher-ups that are guilty."

B, the Examination at the USPTO is not supposed to be exhaustive - we couldn't afford a patent system if it were. Some patentability errors must be expected (3.5% is fine, 50% is not fine). Please read this Patently-O comment posted by Ender (and the following one too):

http://www.patentlyo.com/patent/2007/11/pto-rule-challe.html#comment-88617576

[Anon]

6,331,415 currently under reexam

[B]

NIPRA:

We need a patent system that readily recognizes when a it is about to issue a patent that would be embracing an entire industry. When such patent is about to issue they should be quadrupling their efforts in ensuring that it passes the minimum prior art searches.

When you are about to issue a patent 13 years after its priority date - when the industry has fully developed - with the patent potentially embracing the entire industry - shouldn't PTO alertness be heightened?

It is not the silly "pride" patents - the ones not worth the paper they are written on - that are of concern (which includes 90% of issued patents). It is the submarine time-bombs that should be concerning.

We need a system that can readily distinguish between the two and allocate resources accordingly.

Vonage has within the past six months parted with over $300 million in patent settlements which included Verizon, Sprint, AT&T, and others.

Watch out, Vonage! here comes Hitachi for their piece of cake..

[NIPRA Anonymous]

"We need a system that can readily distinguish between the two and allocate resources accordingly.

B, agreed completely. And it shouldn't be that hard to do with the observations you've already made....

[NIPRA Anonymous]

"Vonage has within the past six months parted with over $300 million in patent settlements which included Verizon, Sprint, AT&T, and others."

P.S. B, that doesn't tell you something... or do you really think Vonage is an innocent victim?

[j]

Your idea is a good one. However, the percentages will not be useful because of one main problem. Assuming you have 2000 submitted "bad" patents, you cannot conclude that .1 percent of the 2 million patents in the sample are bad. The reason is that it is doubtful that all 2 million were examined. For an accurate percentage you would have to compare the number of patents examined by your readers to the number of patents submitted. If I look at 50 patents and think 10 of them are bad, the relevant percentage of bad patents is 20% because to compare those 10 to the total 2 million is to assume that every patent I didn't look at is "good"

[Kevin E. Noonan]

Dear J:

You have a point, but I think the analogy I'm trying to draw is the difference between going to your refrigerator and thinking you detect a funny smell, and living next to Jeffrey Dahlmer. If the patent system is in real trouble, we should get lots of "bad" patents cited at us and when we look we will agree these should never have been granted. Plus there are other metrics we are developing.

But I thought it would be interesting to get a group of motivated readers finding the "bad" patents in their technology area (particularly those the Docs wouldn't necessarily think about), and then step back and see what we think.

So, the bottom line is that there is more to come.

Thanks for the comment.

[B]

NIPRA:

While I haven't looked into the merits of all Vonage cases, I did follow the Verizon v. Vonage case quite closely.

In the Verizon lawsuit, Vonage was the victim of: (1) incompetent lawyers (that focused on claim construction instead of invalidity); (2) E.D. Va. rocket docket (9 months from the filing of the complaint to permanent injunction); (3) pre KSR mindframe; (4) Incompetent counsel that did not timely take advantage of the reexam proceedings.

I personally have found ample prior art that would have invalidate the Verizon claims at issue. So have many others.

This is a link to the CAFC decision in the Verizon v. Vonage case.
http://www.cafc.uscourts.gov/opinions/07-1240.pdf

Note that Vonage did not even appeal the subtance of the "no invalidity" findings of the jury. They only objected to the "jury instructions" on obviousness as being pre-KSR.

Vonage spent millions of dollars defeding themselves, but they haplessly fell victim to inexperienced patent attorneys and a rushed VAED docket system.

The primary patents at issue in the Verizon v. Vonage case cover the mundane concept of intelligent routing of phone calls, albeit in the context of VOIP. e.g. if condition A is TRUE then route to destination X, if condition B is TRUE then route to destination Y. Very mundane and obvious with much prior art to support.

[your mother's an astronaut]

6278006--see claims 19-20 and 24-25.

[anonexam]

How about this one posted on your site a while back:

7,157,418 issued from U.S. Application No. 09/360,242 and claims the benefit of International Application PCT/CA99/00659, filed July 21, 1999, and U.S. Provisional Application No. 60/155,186, filed July 22, 1998. Representative independent claims 1 recites:

1. A method for inhibiting proliferation or migration of activated immune effector cells, comprising administering a conjugate to an animal, whereby proliferation or migration of the immune effector cells is inhibited, wherein:
the conjugate comprises a targeted agent or a portion thereof and a chemokine receptor targeting agent or a portion thereof sufficient to bind to a chemokine receptor on immune effector cells and facilitate internalization of the conjugate;
the chemokine receptor targeting agent is a chemokine, an antibody that specifically binds to a chemokine receptor or a fragment of the chemokine or antibody, wherein the chemokine, antibody or fragment thereof binds to the receptor and internalizes the targeted agent in a cell;
the targeted agent or portion thereof, when internalized in a cell, alters metabolism or gene expression in the cell, regulates or alters protein synthesis in the cell, inhibits proliferation of the cell or kills the cell; and
the conjugate binds to a chemokine receptor resulting in internalization of the targeted agent in cells bearing the receptor.

There are alot of "or" alternatives in issued claim 1. Let’s see - I am to assume the examiner who issued this POS couldn't find 102 or very strong 103 art teaching these limitations:

A method for inhibiting proliferation...of activated immune effector cells comprising administering a conjugate to an animal...whereby proliferation…of the immune effector cells is inhibited, wherein:
the conjugate comprises a targeted agent…and a chemokine receptor targeting agent…sufficient to bind to a chemokine receptor on immune effector cells and facilitate internalization of the conjugate;
the chemokine receptor targeting agent is a chemokine, an antibody that specifically binds to a chemokine receptor or a fragment of the chemokine or antibody, wherein the chemokine, antibody or fragment thereof binds to the receptor and internalizes the targeted agent in a cell;
the targeted agent or portion thereof, when internalized in a cell…inhibits proliferation of the cell or kills the cell; and
the conjugate binds to a chemokine receptor resulting in internalization of the targeted agent in cells bearing the receptor.

102 - hard to believe the genomics patent deluge from 1994-2000 didn’t have at least one if not many novel CCR with boilerplate for anti-CCR antibody conjugates for inhibiting the proliferation of autoimmune cells (i.e., activated immune effector cell).

103 - based solely on my personal knowledge gained from examining in related arts (I have not examiner CCRs but don’t think I need to – that is why this patent is a POS):

Activated immune effector cell – how about your good ‘ole “autoimmune cell;”

Chemokine receptors – well known before 1998 – for example, I think CCR5 was isolated @ 1995, but this is just one of the most famous example;

Chemokine receptor internalization – an “evidentiary type” reference could be used to show that this is intrinsic (inherent) to antibody or ligand conjugate binding (FYI also a very commonly studied phenomenon in receptors of all kinds, don’t think CCRs are an exception);

Inhibiting proliferation or killing immune cells with a targeted conjugate – duh? Off the top of my head I would think of radiolabeled anti-CD20 (now marketed as Bexxar I believe)(of course good for 103 only)

You be the judge.