Friday, March 20, 2015

What the PTO should do about Alice

Last week I had the very good fortune to be a speaker at a Bloomberg BNA conference in San Francisco on the Impact of the Supreme Court on Patent Enforcement. I was invited by Wiley & Rein partners Brian Pandya, a former student of mine, and James A. Wallace. Along with Bloomberg BNA's Steffan Welch, they put together an excellent program.

I was on the panel on "Ineligibility of Abstract Ideas, Business Methods, and Naturally Occurring Substances" moderated by Jim Wallace. The slides I prepared to accompany some brief remarks are here.

I am not a fan of 101 or of the latest phrase associated with it, patent ineligibility. Ineligibility is invalidity without bothering with messy things like proof.

The only sensible part of 101 is the final phrase "subject to the conditions and requirements of this title." Those conditions and requirements are found in 102, 103, and 112. Like Annie Oakley when she compared herself to Frank Butler, those sections can sing to 101, "Anything you can do, I can do better."

Likewise, 101's judicially-created implicit exceptions -- "laws of nature, natural phenomena, and abstract ideas" -- do not need to be exceptions, implicit or otherwise. That's because there has never yet been a patent claim that passes the tests of 102, 103 and 112 that is nevertheless offensive for attempting to patent one of those exceptions. Or anyway, nobody has been able to show me such a claim. No, not even the Supreme Court, not in the most recent 101 cases and not in any earlier ones.

Whenever these judicially-created implicit exceptions have been invoked, it has been to combat advances in technology that patent examiners and judges find scary. These are cases where unfamiliarity breeds fear which is then masked as contempt.

But none of that matters if my suggestion* for the Patent Office is put into effect. Yes, I know, the PTO has ignored my advice about 101 law in the past. (See, e.g., Urgent Letter and Time-Sensitive Proposal to then-new PTO Director David Kappos, posted right before the oral argument at the Supreme Court in Bilski v. Kappos.) But the PTO/DOJ brief in Mayo gave me hope (Argument II: "The disputed claims ... are likely invalid under 35 U.S.C. 102 or 103") and regime change gives me more.
*  Because of time constraints at the San Francisco conference, I was unable to present my "What would you do about it" slides (32-53). The idea for this post is on slide 34 item C and slide 35 item A.
Many at the conference seemed happy with 101 especially because it has become a weapon to resolve cases quickly, in some cases without discovery. My suggestion should be even faster and less costly for the parties and the taxpayers, without knuckling under to the 101-first wrong-headedness. Here's how it would work:

Step 1. As soon as an Accused Infringer (AI) asserts a 101 defense in court, whether by a Rule 12 motion to dismiss or a Rule 56 motion for summary judgment, or perhaps even after an early status conference, the court notifies the PTO.
There is an alternative that would not involve direct communication between the trial judge and the PTO. By statute (35 USC 290), the clerks of the district courts must notify the PTO when a patent lawsuit is filed. The PTO could monitor all such dockets so that it would know when an AI filed a pleading with a 101 defense.
Step 2. The PTO, sua sponte, throws the patent into reexam.

The Director of the PTO has the statutory right to begin reexamination proceedings. 35 USC 303; see also 37 CFR 1.520. True, the Director has to identify relevant prior art. That, however, should not be hard if the AI has argued that the claims-in-suit cover one of the implicit exceptions combined only with well-understood, routine, conventional activities.

As mentioned in slide 12 of my talk, that phrase from Mayo, quoted with approval in Alice, has the acronym WURCA, a plural noun. The associated adjective is WURC. Finding prior art should be easy: WURCA -- if truly WURC -- will be documented in patents, patent applications, journal and media articles, websites, product advertisements and specifications, tweets, etc.

WURCA and obviousness prior art are very similar. The main difference is that obviousness prior art is identified with specificity and considered in a careful manner with intellectual rigor and honesty. WURCA are whatever a decision-maker -- judge or patent examiner -- believes they are, no evidence required, and intellectual rigor impossible by the very nature of the legal standard.

Step 3. As soon as the reexamination is declared, the court entertains a motion to stay proceedings until the reexam is completed. Absent unusual circumstances -- so unusual I can't imagine any at the moment -- the court grants the motion.

Benefits: This suggestion benefits the parties, the courts and the public because it:

a. frees up court time. Judges would not have to write Rule 12 opinions or hold status conferences or anything else until the reexamination is completed and the time has run out for all possible appeals;

b. gets rid of the baloney patents so wittily described by panelist Morgan Chu. Claims that are invalid under 103 are just as dead as claims that are patent ineligible under 101;

c. brings the matter back to the PTO, whose search the first time may have been nonexistent or inadequate;

d. gives the patent owner (PO) a chance to narrow the claims. That might save some claims from attacks under both 103 and 101;

e. discourages use of 101 as a delaying tactic or legal boilerplate with only a minor relationship to the substance of the controversy;

f. helps everyone -- courts, PTO, inventors, competitors, the public -- return to an appreciation of 103 and its powerful, intelligent standard, the HOA-TA (hypothetical ordinary artisan, time appropriate, pron. WHAT-uh), which inherently keeps pace with technological change. See slides 23-24 and, e.g., slides 20-21 in my 2012 talk to the Oregon Patent Association; and

g. saves taxpayer money because, with respect to pubic funding, the PTO is probably a much less expensive forum than is a federal district court.

Advantages over AI-initiated reexamination: Having the PTO Director start the reexamination rather than putting the onus on the AI has additional advantages in terms of fairness, efficiency and justice:

a. No waiting: The Director acts predictably and early in the litigation;

b. the PO interacts with the PTO without the AI having first characterized the prior art and the claims;

c. the AI saves the filing fee and the lawyer bills associated with the reexamination, possibly a life-saving savings for young, financially-strapped AIs;

d. the PTO gets a chance to do its job right in the specific case and to collect data concerning searches in all the cases of Director-initiated reexamination so that improvements can be made; and

e. examiners have no incentive to whitewash these patents. Compare: when the AI initiates the reexam, examiners may want to make the original issuance look good; when the Director initiates the reexam, examiners may want to do it to get it right, not just to get it done (see this post for more about the getting it right/getting it done continuum).

Now, how can I convince the PTO and courts to work together on this?
March 20, 2015; rev 0321-27, 0402

Wednesday, July 23, 2014

Federico's Commentary - Follow-up

In 2009 I wrote that the USPTO should post Federico's Commentary on line. I woke up this morning, July 23, 2014, to find this comment by Anonymous:
I know this blog post is from a while ago, but this appears to be an online copy of Federico's commentary:
July 22, 2014 at 10:55 PM
I responded with thanks and confirmed for myself that Federico's Commentary is still not available on

As I noted in the original post, the Commentary was re-published in the Journal of the Patent and Trademark Office Society (JPTOS, a USPTO publication), 75 JPTOS 161 (1993). Later JPTOS printed a special issue containing only Federico's Commentary and a speech by Judge Giles Rich, long viewed as the best "We were there" resources on the language of the 1952 Patent Act. (I guess you would call them "legislative STAFF history" rather than "legislative history"?) Surely the Patent Office has an archive, electronic or not, of all the issues of JPTOS? And there must be at least one employee not yet retired who has a paper copy of 75 JPTOS in nis office? Is there so little need -- by Patent Office employees or others -- to understanding the 1952 Act's original language? It still lingers though buried under the tax-code-like verbosity of the last many years of amendments and there are still questions, or Anonymous would not have found nis way to my blog...

Anyway: here is what I found today when I searched for "federico." The search box says "Powered by Bing" by the way but I don't blame the absence of the Commentary on Microsoft. Of the first 20 hits, all but 3 relate to other people named Federico, mostly inventors. Two of the three hits for the Federico of interest (PJ Federico, a Patent Office Examiner-in-Chief) are citations and rather recent ones at that: a 2010 BPAI decision, Cal Tech v. Enzo, Interference No. 105,496, and the Government's brief to the Supreme Court in Bilski v. Kappos (08-964), filed in September 2009. The third occurrence of the right Federico is in the remarks by Todd Dickinson in his memorial tribute to Judge Rich, which I had mentioned in my original post. If there had been a thirty years ago, all 20 of the top 20 hits would have been to Federico's Commentary, would they not? Losing our history because of the internet is a problem, solvable but not yet solved.

In today's search I also noticed the absence of occurrences of federico's -- federico apostrophe s -- in the search results. That is a difference between powered by Bing and powered by Google. (I searched all of Google for federico and did, ultimately, confirm that federico's would be found: Hit number 47 was for federico's pizza.) With federico apostrophe s in the search box, Bing found exactly one (1) occurrence: the 2007 OG Notice on the final rules for continuation applications, also mentioned in the original post.

One last search, this time on Google and for federico's commentary. The first two hits are to The third is to heinonline. And the fourth, amazingly, is to my 2009 post. Which by the way has a grand total of 1320 hits in its entire life. A sad commentary (npi) in several respects, eh?

But the silver lining was the fifth hit. It was, a twenty-year-old post by Robert M. Storwick of a patchwork quilt "based on the 26th order, 500-unit compound perfect square discovered by Pasquale J. Federico, mathematician and a long-time [PTO] employee." Storwick gives math journal citations, too. Way cool.

Thank you, ipmall, for posting Federico's Commentary, and thanks also to ipmall's sponsor, the University of New Hampshire School of Law, and last but not least to Anonymous, for sending me on this delightful chase.
Alas, one more search, this time for Storwick, may have been one too many. He was disbarred by his home state in 2007. Maybe he had some major mishap because he did not respond to the state bar at all and the latest financial transactions mentioned in their decision were from 2006. But in 2013 somebody updated the registration and paid for it through 2018, even though the homepage says it was last updated in 1995. A mystery.
Links fixed and minor rev 7/27/14 rjm

Thursday, June 12, 2014

Braintree Labs and "A" - the indefinite article excludes the singular?

Most of what I know about the Federal Circuit's recent Braintree Labs v. Novel Labs decision is what I read in Greg Aharonian's newsletter commenting on Hal Wegner's post. According to Aharonian and Wegner, the Federal Circuit in Braintree ruled that the word "a" (yes, the indefinite article) "excludes the singular" when used in a patent claim.
Note that when I write "a patent claim" I use "a" in the singular sense, in particular referring to the one claim I know about. There may in fact be more than one claim in this patent that uses the word "a" in the very phrase before the court. But there is no principle of patent law that would require the inventor to write more than one claim. Therefore, as a matter of law, the phrase "a patent claim" cannot exclude "one patent claim."
The patent claim recites:
    A composition for inducing purgation
    of the colon of a patient ..."

At the risk of breaking a* cardinal rule of law, writing, and common sense, I would like to add the following idea to the discussion of this case, even though I have done no more than skim the decision for the word "colon."

A* principle of claim interpretation - or statutory interpretation or any other kind of "construction" in the legal sense - is to look at the words surrounding the disputed language. Context may provide meaning that removes ambiguity. Here, I notice that "a patient" is the object of the preposition "of" which in turn is preceded by the words "THE COLON."

"THE COLON OF" can only belong to one patient at a* time. It certainly belongs to a SINGLE patient. That it could belong to a* group of patients (that is, a number of patients that exceeds "the singular") is -- absent some pretty spectacular surgery -- unlikely.

I will now read more fully to see whether the words "the colon of" gave a* moment's pause to any of the judges or lawyers arguing for this singular (in the figurative sense) meaning for the word "a," or whether the dissenting judge or lawyers on the other side thought at all about colons in the anatomical, not the grammatical, sense, and the 1:1 correspondence between (colon) singular and patient (singular).

Generations, nay centuries, of patent lawyers have understood and used the indefinite article to mean AT LEAST ONE or, in an even more common formulation, ONE OR MORE. As has been taught for years, including both TO me and BY me, an* inventor could claim a* "chair having a leg ..." and the claim would read on a chair with one, two, three, four, or any number of legs. Beyond question, it would read on a* chair with a SINGLE leg.
* I intend here to give the words "a" and "an" their ordinary meaning, and the meaning that was generally accepted among patent lawyers until Braintree. That meaning is "one" or, to be polysyllabic and grammatical about it, "the singular."

Wednesday, April 2, 2014

Three (3) Ways the patent system promotes progress. Not 1. Not 2.

The patent system promotes progress in three ways.  At least three.  But certainly not one. And not two. They are: the carrot, the stick, and the "it's not all about you" factor: enriching the public domain.

Let me count the ways.

The first way patents promote progress is the one that comes to mind from the phrase "incentivizing the innovators" (see [n1] below).  The carrot of possible patent protection motivates people to come up with ideas, sometimes all on their own, sometimes because their knowledge and training convinces people with money to hire them to come up with ideas.  Then people (the same or different) spend some money on developing, commercializing, and marketing those ideas.  Obtaining a patent lets the people who own the patent exclude competition, albeit for a relatively short time (see [n2] below).  Those people may make more money and then they may decide to spend it on additional research and development.  Innovation happens.  Society benefits.

But that is not the end of the story about how the patent system promotes progress. We have to count past one to two and three.

Numbers two and three on the list of benefits are less obvious.  They also tend to be ignored by journalists, academics, judges, justices, and others who do not know all that much about the patent system.  In fact, this post was prompted by a talk I heard last May about patents as incentives to innovation. The talk was half over before the speaker mentioned the third benefit and ne never mentioned the second at all.

The second way patents promote progress is with a stick.  Other people's patents can stop you from doing what you want because of the threat of suit, or the owner's refusal to license you at a rate you can afford, or your refusal to deal with the owner because it is a fierce competitor or, what gets too much attention these days, a troll. But the stick of patent protection is more in the nature of a prod. You are always at liberty to come up with a better idea: in patent lingo, to "design around" the patent.  Patents, during their limited life, don't bar everything in the field, they only bar what is claimed. Alas, in at least the years since the 1982 creation of the Federal Circuit (the appeals court for patent cases), the Supreme Court has never used the phrase design around to describe one of the ways that patents "promote Progress in Useful Arts," the justification for patents given in the Constitution (Art. I, sec. 8, cl. 8). Why not? Everyone says "saying it does not make it so" (attribution unknown) but this is a case where
NOT saying it does not make it NOT so.
"Designing around" may lead to another patent because of that carrot, and then there will be another stick (prod) as well as benefit number three.

The third way patents promote progress is that they enrich the public domain, the public store of knowledge about how to do things.  In order to obtain a patent, the inventor must describe the invention so that other people can (in patent lingo:  are enabled to) make and use it.  "Enablement" has long been referred to as the quid pro quo for the patent "monopoly."  Anyone can read a patent and learn from it. That was true even before the internet. That rich store of free, accessible and relatively easy-to-search information is a huge benefit that we, the people, receive because we have a patent system. Yet the term public domain has been used only rarely by the Supreme Court in the last forty years in decisions about patent law, and then only to focus is on how a wrongly issued (invalid, overly broad, etc.) patent steals from the public domain. The only exception to that anti-patent bias is found in Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 US 141, 151 (1989). Bonito Boats addressed a challenge not to a patent but to a state statute against copying boat hulls, and the opinion contains several references to the theft view of patents before acknowledging that:
"[T]he ultimate goal of the patent system is to bring new designs and technologies into the public domain through disclosure."

I think the patent system -- and the statutes, the case law, the proposals for reform -- would be improved if people understood that it has more progress-promoting capability than just offering the promise of a time-limited monopoly.

(n1)  Jim Alexander, a student in my IP Survey course in 2007, was the first person I heard use this phrase.  I believe it originated with Prof. Katherine E. White. See her 2006 law review article: Preserving the Patent Process to Incentivize Innovation in Global Economy, 2006 Syracuse Sci. & Tech. L. Rep. (2006).

(n2) It is true that patents give patent owners' some rights but there are limits.
           A.    A patent gives the owner the right to exclude but only for a limited time.  Patents have expiration dates.  Before the expiration date, the owner must pay maintenance fees.  Once the patent expires, or once the fee hasn't been paid, the patent owner has no right to exclude. The patent has risen into (see [n3] below) the public domain.
           B.  A patent is a right to exclude not a right to do. That's because there may exist a patent #1, which gives owner #1 a right to exclude what patent owner #2 does, even though owner #2 has a patent. During the time that patent #1 is in effect, owner #2 has no absolute right to practice nis own patent but ne can: (1) wait until patent #1 rises into the public domain; (2) design around patent #1 (and maybe obtain patent #3), or (3) hope that owner #1 won't do anything. If that hope is dashed, owner #2 can choose to (a) change its business; (b) fight; or (c) take a license. Depending on how much #1 and #2 hate each other (and how good patent #3, if there is one, is as an alternative), owner #1 and owner #2 may decide to cross-license or merge. More innovation may emerge from that transaction and often does.

(n3)  At a conference on digitization at the University of Michigan in 2006, President Mary Sue Coleman gave excellent opening remarks about copyright and education and libraries.  I felt obliged to comment, however, when she said something about works "falling into" the public domain.  I pointed out that we ought to say that they RISE into the public domain.  When a piece of intellectual property passes from private hands to the public domain, the change should be likened to going to Heaven, not the other place, or to standing up and being counted, not to collapsing and being buried.  (I was unable to attend the afternoon session at that conference, but I was told that some speakers had changed their slides where they mentioned the public domain, crossing out the word "falling" and putting in the word "rising.")  Yes.

Drafted, but not posted, in May of 2013.
Minor revisions and posting: April 2, 2014; rev 20140523

Saturday, April 13, 2013

Damages and Trolls - a loophole in the marking statute - 01

Questions:  Can it be that there is NO NOTICE REQUIREMENT at all as a limit on damages for
non-practicing entity patent owners (NPE POs aka trolls)? Can it be that patent owners can always collect from the start of infringement as long as they do not practice the patent?  Answer to both:  Under the current law, YES.

In all the anti-troll fervor that pushed for the America Invents Act of 2011 (112 PL 29), did anybody propose to close the loophole?  The answer seems to be NO, but I admit I haven't searched much yet.

Consider the relevant statute, 35 USC § 287(a)(as last amended effective Dec. 16, 2011)).  Section 287, often referred to as the marking statute, is entitled
Limitation on damages and other remedies;
marking and notice."
(Full text provided at bottom of post; numerals have been added before each sentence and the whole passage has been reformatted for ease and speed of comprehension.)

I propose that the provision concerning "Limitation on Damages"  be separated from the provision describing proper marking.   That is, I would leave the first sentence of 287(a) by itself, and replace the second two sentences with a new subsection that focuses on limitation of damages.  Marking and damages are related but different.  The loophole may well have arisen because the motivation to include the limitation on damages was to encourage people to mark their products with patent numbers, rather than to legislate a fair starting date for calculating damages.  A sensible Limitation on Damages provision might say something like this:
In an action for infringement, a patentee is entitled to damages from the date on which it first provided notice of the infringement.  Notice can be actual or constructive.  Marking the patented product in accordance with 35 USC section 287(a) constitutes constructive notice. 
I plan to write another post or two to discuss 
   - the historical cases (Wine Railway (Sup. Ct. 1936), American Standard, Amsted, Texas Digital), 
   - the sub-loophole related to method patents (partially closed by Amsted, a 1994 case I taught when I taught basic patent law), and 
   - the very recent opinion that got me thinking about this problem now (SSI v. TEK, summary judgment decision of 3/7/2013). - RJM

35 USC 287(a) (reformatted, numerals inserted for each sentence)
and persons
       offering for sale, or
       selling within the United States
   any patented article for or under them,
       or importing any patented article into the United States,
may give notice to the public that the same is patented,
   by fixing thereon the word “patent” or the abbreviation “pat.”,
        together with the number of the patent,
   or by fixing thereon the word “patent” or the abbreviation “pat.”
        together with an address of a posting on the Internet,
        accessible to the public without charge for accessing the address,
        that associates the patented article with the number of the patent,
        when, from the character of the article, this can not be done,
   by fixing to it, or to the package wherein one or more of them is contained,
        a label containing a like notice.
In the event of failure so to mark,
no damages shall be recovered by the patentee in any action for infringement,
except on proof that
   the infringer was notified of the infringement
   and continued to infringe thereafter,
in which event damages may be recovered
only for infringement occurring after such notice.
Filing of an action for infringement shall constitute such notice.  

NB:  I see no reason to include sentence [3] in my proposed replacement provision.  If a patentee files an action but does not serve the Complaint, then it has not given notice, it has avoided doing so.  If the patentee serves the Complaint, then it has given actual notice.   Am I missing something?

Thursday, February 7, 2013

A Great Idea That's Never Been Tried - The Patent System

I've collected some quotes about great ideas that have never been tried.  None of the quotes was about the patent system but all of them could have been. Click here for the powerpoint or just enjoy the individual screen shots:

I had used the quotes in some recent talks (Oregon Patent Bar 2012, slide 12; and Stanford faculty lunch half-talk, slide 22). Today Greg Aharonian's newsletter inspired me to put the quotes together in a separate document and post them here.  Greg wrote about the Federal Reserve economists who are proposing abolition of the patent system. They undoubtedly did it to get attention, and will succeed, and to avoid the real work of determining how to help the economy in ways the Federal Reserve can accomplish.[n1] Like too many employees at _____ [fill in the name of your current employer or other appropriate entity], these economists may have forgotten why they were hired.  Or maybe they never knew.
[n1] A Constitutional amendment would be required to abolish patents [see Art. I, sec. 8, cl. 8], and under the Constitution neither the power to propose amendments nor to ratify them is given to the Federal Reserve [see Art. V].) 

I have long said - well, not exactly in these words -
"Let's let the PTO put all its effort into making sure applications are valid over the prior art. [n2]   That would utilize the examiners' supposed expertise.  And then let's wait,say, 5 years before we decide that the patent system is irretrievably dreadful.  That is, let's TRY having a patent system instead of fussing around with irrelevancies like patentable subject matter."
[n2] A little amplification for those who know something about patent law and the patent system:  What I mean is that examiners should issue rejections *ONLY* on 102 and 103 [35 USC 102 and 103, requiring patents to be valid over the prior art, aka the novelty and obviousness requirements].   OK, for claims to perpetual motion machines for which there is no prior art (ha), they can reject under 112  [35 USC 112, requiring patents to be written so that an ordinary artisan in the field is enabled to make and use them, aka the enablement requirement], too.

Really, if the PTO does not give a high priority to examining patent claims for validity over the prior art, it should not require a technical background of its examiners.  There are many smart liberal arts majors who would be just as good as engineers, maybe better, at deciding if a claim is drawn to "patentable subject matter," or anyway writing at length about it, especially if they have already considered how many angels can dance on the head of a pin.

For more of my views on the PTO's expertise and the role of prior art, see among other things my In re Bilski amicus brief, Microsoft v i4i amicus brief, blog post about Bilski when it was pending before the Supreme Court (see [n24] and accompanying text for a discussion of angels dancing on the head of a pin), and my comments on the PTO's Interim Guidance after Bilski.
Broken link fixed and typo corrected 20140402

Sunday, November 6, 2011

Advice to a Young Would-Be Patent Law Scholar - 06 (MERELY and CLEARLY)

In  Advice 05 - Those NEVER and ALWAYS Rules of Writing, I mentioned my own rule about MERELY and CLEARLY:
Never use merely or clearly more than once in every 25 pages.
I also pointed out that the purpose of a NEVER/ALWAYS rule is not so much to prevent or mandate anything, but to make the writer self-conscious whenever ne starts to violate the rule, which in turn will lead to less and more of the NEVER and ALWAYS behavior, respectively.

The main reasons I do not like words like MERELY and CLEARLY are:
  1. They allow the writer to be sloppy and lazy.  
  2. They insult the reader.
If the referenced information is so MERE, the author's previous discussion of it in comparison to other factors ought to have demonstrated its mereness.  If the referenced conclusion is so CLEAR, the author's development of the arguments ought to have persuaded us already.

An author who writes MERELY or CLEARLY, either thinks nis readers are stupid or hasn't done nis job (or hopes they are because ne hasn't).

Another reason I dislike such words is that
3.  They are used as code for "I am a member of the club." 
The lawyer's club, in this case, with no disrespect meant to the beautiful dormitories at the University of Michigan.  They are words that ought to be on a law school bingo card.
See if you are unfamiliar with the concept. (Unfortunately that site does not have a lawschool version yet.) Other candidates for the bingo card are paradigm, cabined, conversation, any word formed by combining where or there with a preposition (therefrom, whereby, and so on), and just about anything in Latin (the one I dislike the most is vel non). Want to add to the list? Please write a comment!

When I read or hear those clubby bingo words, I immediately look for other signs of a failure to think things through, to  marshal the facts carefully, or to structure the argument convincingly, and I almost always find them.

Last, but maybe first,
4. or 1.  Using "merely" or "clearly" violates my ALMOST NEVER/ALMOST ALWAYS rule:  Use adjectives and adverbs sparingly.  Instead, color your writing with the nouns and verbs.
Another post, amplifying my reasons to favor  nouns and verbs over adjectives and adverbs is coming soon.
typo in link corrected 11-14-2012