..What Algorithm Patents Can Do For You
...(or: Sky? What Sky?)

I have finally been convinced (by the Herculean efforts of those
supporters of algorithm patents) that detractors from algorithm
patents are a bunch of whinging no hopers, who, like Chicken
Little, have been deluded by their fears into the belief that the
sky is about to fall in upon them.

Poor misguided fools.

On the contrary, patents on algorithms can only serve to encourage
innovation in the field of software, and, indeed, in all the
multifarious walks of life along which we are led. 

Allow me to elaborate, by reference to the beauty of algorithm
patents and how they are helping the progress of mankind. In
particular, I refer to the _Application of Pardo and Landau_, Decided
on August 5, 1982, by the CCPA. (citation: 214 USPQ 673).

I begin with a history of the patent:
Some time prior to 12 August 1970, (the date of application)
two gentlepeople Rene Pardo, and Remy Landau (they
are Canadians) devised a most ingenious method of dealing with
many formulas, which might be given in a random order. Permit me
to quote the description of claim 30 from the judgement (at 674):

.30. A process of operating a general purpose data processor
.of known type to enable the data processor to execute formulas
.in an object program comprising a plurality of formulas,
.such that the same results will be produced when using the
.same given data regardless of the sequence in which said
.formulas are presented in said object program, comprising
.the steps of:

.(a) examining each of said formulas in a storage area of the
.data processor to determine which formulas can be designated
.as defined;

.(b) executing, in the sequence in which each formula is
.designated as defined, said formulas designated as defined;

.(c) repeating steps (a) and (b) for at least undefined formulas
.as many times as required until all said formulas have been
.designated as defined and have been executed.;

.whereby to produce the same results upon execution of the
.formulas in the sequence recited in step (b) when using the
.same given data, regardless of the order in which said formulas
.were presented in the object program prior to said process. 

A truly diabolical invention, do you not agree? I myself, still marvel
at the subtlety of the mechanism of the algorithm. How they ever hit
upon the idea of doing first things first, and second things second
is completely beyond my feeble intellectual capacities. I am but a
poor hyperbolist, after all... 

Well, the Commissioner got the frightful idea into his head that this
expression of mental fortitude and mathematical insight was undeserving
of patent protection. "My! My!" I thought to myself, "for such an
astounding invention, surely those sterling chaps Pardo and Landau
deserve not only patent protection, but popular acclaim!" I was
understandably taken aback by this contention of the Commissioner,
as no doubt, you are too, dear reader. 

He argued that the invention was (1) a mathematical algorithm under
section 101, and (2) obvious under sec. 103. Fortunately, in a most
enlightened judgement I might add, the Court of Customs and Patents
Appeals rid the Commissioner of all basis for his scurrilous
suggestions.

The court argued that a mathematical algorithm was neither directly
nor indirectly recited (applying the first step of the Freeman test,
reasoning to be found at 676). Indeed, they note, at 677, "That a
computer controlled according to the invention is capable of handling
mathematics is irrelevant to the question of whether a mathematical
algorithm is recited by the claims."

Similarly they made short shrift of the obviousness contention of the
Commissioner. Noting that the Board had cited no references in support
of this contention, and were relying on nothing more than their naked
common sense, the Court correctly castigated them for their poor form.

If I may be permitted but one more, short quotation  (@ 678):
.The board did not determine the scope and content of the prior
.art or even indicate that there was any prior art; it did not
.determine the differences between the prior art and the claimed
.invention; and it did not ascertain the level of ordinary skill
.in the art at the time the invention was made. 

Moreover, the court observed, with a fiendish sagacity, that computers
had been in general use for 18 years prior to the filing date, and yet
the PTO had been unable to cite a single reference for their claims.

I can but laud the Court its insight.

Does this not demonstrate, gentle reader, the supreme utility of the
application of patent law to algorithmic inventions? How is the area
of Computer Science to progress without fundamental breakthroughs of
the likes of that described above? It is a verity without need of
citation that the whole of society will benefit from the extension of
patent law into this virgin territory.

And now, a short word of apology. I had intended to release this
article several weeks earlier. Unfortunately, upon
consultation with my attorney, I discovered that, via the doctrine
of equivalents, I may have been infringing the patent had I followed
the normal process I do in the preparation of these documents.(1) For
example, I did research before beginning to write, I got a new page
out prior to writing upon it with my pen. I even went to the lengths
of turning the computer on before I entered the article into its
memory.

"Very bad," he said, "very bad."

So, in order to avoid a possible prosecution I carried out
those steps in an order not dictated by the realms of possibilty.
It was not easy, but I have finally done it. You should look for
the patent I have lodged with the office: "A
Method for Subverting Causality".

See, that original patent has already spawned innovation!


(1) Okay, I'm playing a little bit hard and fast with the doctrine
of equivalents here.


                                                 Brendan Scott
                                                 Canberra 1992
