Print

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

[smgsmc]

Hi.  Is the phrase "tied to a particular machine or apparatus ..."  literally interpreted to mean tied to a *single* machine or apparatus?  I'm concerned about processes operating in a system or network.  For example, transmitting data from a transmitter in a first network element, routing the data through a data network, and receiving the data at a receiver in a second network element.  If different steps in the process claim are tied to different machines, am I OK?  Or do I need to cover myself in the spec by defining an entire communications system as a single machine?  Thanks.

[Isaac]

I'm concerned about processes operating in a system or network.

Isn't the network of elements collectively a system or apparatus?   Probably not a Bilski issue.

[JimIvey]

I agree with Isaac, except that I'd say it's clearly not a Bilski problem.  Of course, the USPTO often sees things differently that I do.

The gist of Bilski is what to do in the complete absence of a machine or transformation of physical matter (except for physical matter found in a computer, which is presumably dark matter or anti-matter or figments of our collective imagination).  Adding more machines should not, under any reasonable rationale, be less "technical/physical" than a single machine.

Regards.

[Robert K S]

What about when an applicant has a claim like:

1. A method wherein a doodad does
(a) this,
(b) that, and
(c) the other.

"Doodad" is defined in the spec as being any one of a list of unquestionably well-known particular machines "or similar"/"or the like".

Examiner issues Bilski rejection.

I say, this is improper.  An applicant is entitled to be his/her own lexicographer, and the applicant's spec defines "doodad" as being any one of these listed particular machines or the like.

But what if examiner argues--"Exactly.  Thus a doodad is not a particular machine!"?

[JimIvey]

Well, I'd use "machine" rather than "doodad."  I'm sure the examiner will argue that "machine" isn't a machine, but I think I'd prevail, ultimately.

I'd say it would be even better if the "this", "that", and "the other" involved the machine in some way.  Someone recently posted about a decision of the BPAI in which steps of something like monitoring memory usage of a computer process did not require a machine and therefore ran afoul of Bilski.  I think the other steps somehow involved detecting a condition in the memory usage and then doing something about it.

How can one monitor memory usage by a computer process without a computer?  Apparently, some panelists of the BPAI do it all the time.

Regards.

[Robert K S]

Hilarious.  Post the link to that BPAI decision if you find it again, would you?

[JimIvey]

Here's my post on the decision.

A PDF of the decision is available here.

Regards.

[Robert K S]

I really don't mind the Bilski decision as written, but when I see it being interpreted this way, I think it's time for the Supreme Court to blast it out of the water.  My jaw dropped repeatedly reading this BPAI decision.  Especially when I read sentence like "That a claim may appear on its face to be directed to § 101 subject matter (e.g., a “machine”) does not end the analysis."

IT DOESN'T?

As support for the statement, the opinion cites to In re Alappat, which only begins a string of further references, next to In re Johnson which only quotes Gottschalk v. Benson.

Benson says that a claim cannot preempt a mathematical formula by for all practical purposes patenting an algorithm itself.  I fail to see how this logic carries through to the case at hand.  The BPAI decision appears to hinge on a falsity hidden behind a chain of references.

And the decision's logic that a "data structure" doesn't inherently imply a memory would confound a person having ordinary skill in the art.

[smgsmc]

Hi.  Let me rephrase my original question.  Right now there are mixed rulings on whether software processes are patentable if you tie them to a general purpose computer (according to a web seminar on Bilski that I just took).  Let's assume for the sake of argument that by hook or by crook, you can.  Now consider software processes running in a client-server architecture.  There would be a set of method claims in which all the steps are performed by the server machine.  Similarly, there would be a set of method claims in which all the steps are performed by the client machine.  Now what about the method claims in which some of the steps are performed by the server machine and some of the steps are performed by the client machine?  Does the requirement of "a particular machine or apparatus" create a problem with the third set of method claims?  In the spec, do I need to provide explicit support for [a system (comprising a server machine and a client machine) = a particular machine or apparatus]?  Or is a method claim OK as long as each step is individually tied to some particular machine or apparatus, even though more than one particular machine or apparatus is involved?  Thanks.

[JustAnotherExaminer]

Hi.  Let me rephrase my original question.  Right now there are mixed rulings on whether software processes are patentable if you tie them to a general purpose computer (according to a web seminar on Bilski that I just took).  Let's assume for the sake of argument that by hook or by crook, you can.  Now consider software processes running in a client-server architecture.  There would be a set of method claims in which all the steps are performed by the server machine.  Similarly, there would be a set of method claims in which all the steps are performed by the client machine.  Now what about the method claims in which some of the steps are performed by the server machine and some of the steps are performed by the client machine?  Does the requirement of "a particular machine or apparatus" create a problem with the third set of method claims?  In the spec, do I need to provide explicit support for [a system (comprising a server machine and a client machine) = a particular machine or apparatus]?  Or is a method claim OK as long as each step is individually tied to some particular machine or apparatus, even though more than one particular machine or apparatus is involved?  Thanks.

I examine networking stuff, so you may appreciate my opinion on the issue. You should get a Bilski rejection if the claimed method covers one or more embodiments that can be performed completely mentally/verbally/physically by or between humans.  99% of Bilski rejections can be overcome by adding minor language to the preamble or the "things" in the method.  It must be tied to at least one particular machine. It can be tied to more than one. The particular machine does not have to be the thing performing the method.

If you claim the following...

A method performed at a client, comprising:
Constructing a message;
Storing the constructed message in memory;
Transmitting the message to a server;
Waiting for a reply from the server; and
If a reply is not received from the server within a predetermined amount of time, retrieving the stored constructed message from memory and transmitting the message to a second server.

Barring explicit specification definitions, you will get a Bilski rejection.  Client = Person A; Server = Person B; Memory = Person A's memory; message = verbal sentence construct; transmitting = talking; waiting for reply = listening; retrieving from memory = recalling from Person A's memory.  Notice I emphasized "definition".  Examiners' definition for "definition" may be different than your attorney's definition for "definition".

You could, for instance, amend the above claim to limit the "a second server" to being a particular machine and do nothing with regard to "a client" or "a server" and still overcome the Bilski rejection.  For example, adding the limitation "wherein the second server is a SIP gateway server" would overcome the Bilski rejection.

There is no requirement that each step be tied to a particular machine.  Only one of the steps, OR, more commonly, recite enough structure in the preamble.

Hope that helps.  Also note that in the networking realm, if you actually claim what your invention is, you will never run into a Bilski rejection.

[JimIvey]

Interesting.  Thanks for the analysis.  It's helpful.

Would something like this help:

Transmitting the message through a communications network to a server;

I would argue that, even if that covers a telephone call, it requires a particular machine -- in fact, several machines that make up any communications network.

I believe this overcomes any Bilski rejection.  Now, that still leaves the claim subject to very broad interpretation and so may be anticipated or obvious, but I think utility is satisfied here.

As for definitions in the Spec, would this help?

As used here, a client is all or part of one or more computer processes executing in one or more computers ... .

Thanks.

[JustAnotherExaminer]

Interesting.  Thanks for the analysis.  It's helpful.

Would something like this help:

Transmitting the message through a communications network to a server;

I would argue that, even if that covers a telephone call, it requires a particular machine -- in fact, several machines that make up any communications network.

I believe this overcomes any Bilski rejection.  Now, that still leaves the claim subject to very broad interpretation and so may be anticipated or obvious, but I think utility is satisfied here.

As for definitions in the Spec, would this help?

As used here, a client is all or part of one or more computer processes executing in one or more computers ... .

Thanks.

Again, barring explicit specification definitions, I interpret "communication network" as two or more things communicating.  Which could be humans talking to one another and therefore not tied to a particular machine.

As for the spec definition, that's a tough one.  I think it was specifically said by BPAI in one of their recent add-a-Bilski-rejection-extravaganza cases that sole use of the term "computer" is not a particular machine.  Is there any way to interpret that definition as a client still being a human? Doubtful.  I think you'd be fine.

[JimIvey]

Again, barring explicit specification definitions, I interpret "communication network" as two or more things communicating.  Which could be humans talking to one another and therefore not tied to a particular machine.

I can see that as an interpretation of a "social network" or just "network", but not a "communications network."  Would "electronic network" be better?

As for the spec definition, that's a tough one.  I think it was specifically said by BPAI in one of their recent add-a-Bilski-rejection-extravaganza cases that sole use of the term "computer" is not a particular machine.  Is there any way to interpret that definition as a client still being a human? Doubtful.  I think you'd be fine.

Can you give a cite for such an opinion?  I find it utterly ridiculous that a computer is not a particular machine.  In essence, it sounds as if the PTO believes that a computer is functionally equivalent to a human brain.  It simply isn't.

Regards.

[JustAnotherExaminer]

I can see that as an interpretation of a "social network" or just "network", but not a "communications network."  Would "electronic network" be better?

Yep, social networking is one interpretation. "Electronic" would probably overcome.  I'd personally prefer seeing what makes up the network though, which is typically in the preamble. Something like "wherein the network comprises two or more devices" or whatever.

Can you give a cite for such an opinion?  I find it utterly ridiculous that a computer is not a particular machine.  In essence, it sounds as if the PTO believes that a computer is functionally equivalent to a human brain.  It simply isn't.

Regards.

I may have been thinking of this case:
Ex parte Cornea-Hasegan
http://des.uspto.gov/Foia/ReterivePdf?system=BPAI&flNm=fd20084742-01-13-2009-1
http://www.bitlaw.com/source/cases/patent/Cornea-Hasegan.html

The term "processor" does not impose being tied to a particular machine.  I'd imagine one could make the same argument for "computer"?

Or this one:
Ex Parte Nawathe et al
http://des.uspto.gov/Foia/ReterivePdf?system=BPAI&flNm=fd20073360-02-09-2009-1
http://www.uslaw.com/library/Intellectual_Property_Law/Bilski_Decision_BPAI.php?item=378001

"[T]he recited method, while being computerized, is not tied to a particular machine for executing the claimed steps."
So, the sole use of the term "computer" does not impose being tied to a particular machine.  Also, I don't agree with uslaw's interpretation of the apparatus analysis of the Ex Parte Nawathe decision.  I think the apparatus was fine because it used "means for" language, invoked 112 6th, and therefore had enough structure to be a particular machine.

Or this one:
Ex Parte Gutta
http://des.uspto.gov/Foia/ReterivePdf?system=BPAI&flNm=fd20083000-01-14-2009-1

Further evidence that nominal use of "computer" isn't enough in the BPAI's opinion.

Or any one of these:
http://www.patentlyo.com/patent/2009/01/the-bpai-and-the-machine-or-transformation-test-of-bilski.html

[JimIvey]

Thanks.  I'll take a look at these and see if I can divine from where the idea that computers are too ethereal for patent protection comes.

Regards.