If we're going to have software patents, let's at least have a level playing field | Opensource.com

If we're going to have software patents, let's at least have a level playing field

Posted 11 Oct 2011 by 

Rating: 
(12 votes)
If we must have software patents, at least use the same rules
Image by : 

opensource.com

submit to reddit

Last week, I participated in a panel discussion at the Eastern District of Texas/Federal Circuit Joint Bench-Bar Conference in Dallas.  The Court of Appeals for the Federal Circuit is the specialized appellate court for patent cases.  My panel was on corporate counsel opinions of patent litigation and recent judicial and legislative patent reform.  The discussion was moderated by Judge Richard Linn of the Federal Circuit.  It was a great opportunity to present some views of the patent system, and to provide options for improvement to the very people who can enact judicial change.

In my remarks, I pointed out that while there have been some significant judicial changes over the last five years (regarding damages, injunctions, obviousness, indirect infringement, and willfulness), much is still needed.  As my colleague Rob Tiller has repeatedly discussed software patents exact considerable costs to innovation in this country.  Although we are still waging the war on patent coverage for software, other battles are also in play, which I present here.

The points I made in Dallas focused on the basic business need for certainty, or at least some degree of predictability, since making efficient and effective business decisions requires a reasonable understanding of the future.  The patent system, especially as it relates to software, does not provide this predictability.  Software patents are frequently particularly vague in scope, whereas chemical patents will often recite a particular structure which is objectively described.  For example, any potential infringer should know exactly what molecule is in play when a patent recites "pyridine-3-carboxylic acid" (niacin or vitamin B3).

Similarly, everyone knows that "a writing device comprising a cylindrical lead, and a tubular wooden body having an end-to-end axial channel, said body contiguously surrounding said lead" is a pencil.   Even electrical patents have a reasonable degree of specificity since one can "follow the electrons."  However, none of this is true with software patents, which often cover abstract concepts rather than specific physical devices or processes.

This is not merely an academic argument because unclear or unknown boundaries do not provide notice of what such patents cover.  To fulfill their notice function, which is one of the basic tenets of our patent system, patent claims must delineate the scope of patent rights with sufficient clarity so that a person skilled in the relevant art can reliably determine whether planned activities would infringe.  However, a creative plaintiffs attorney is often the only person who knows whether a product might be accused of infringement.  Software patentees are often not required to fully describe their ideas or how to achieve them, and are rewarded with broad “flexible” patents that can metamorphose into different identities throughout litigation.  For example, patents on now-obsolete ideas are resurrected to read on technologies developed years later.  The claims in litigation often bear little, if any, resemblance to the invention described by the inventor. 

So what can we do to solve this uncertainty problem?  One answer is to simply apply existing rules.  For example, the inventor must fulfill the "written description" requirement by describing the claimed invention in sufficient detail that someone of relevant technical expertise can reasonably conclude that the inventor had possession of the claimed invention.  This requirement serves both to disclose the technical knowledge upon which the patent is based and to demonstrate that the patentee was in possession of the claimed invention.  Further, the "enablement" requirement mandates that the inventor describe the invention so that that someone of relevant technical expertise can make and use the claimed invention.  Yet many, if not most, software patents do little to assist a skilled technician to actually recreate the alleged invention, and many do not disclose enough to prove that the patentee is actually in possession of any invention, but rather reveal an unsupported and impracticable idea.

We should apply these same rules to all patents.  Currently, biotechnology and pharmaceutical patents are more closely examined by both the Patent Office and by the courts to make certain that the inventor has an invention and has sufficiently described it.  This is a significant part of biotech and pharma patent litigation.  However, the written description and enablement requirements have been largely glossed over, if not ignored, in the case of software patents.  If a plaintiff is going to twist a patent away from the inventor's original meaning, the patent should provide the public – or at least those who might be accused of infringement via such a creative read – notice of this "new meaning" in some manner.  Forcing patentees to describe in detail how to practice the invention would give later-accused parties some predictability so that they can avoid infringement.

Until we have some reasonably certain boundaries on software patents, innovators will continue to face risks that they are unable to fully evaluate.

submit to reddit

10 Comments

Shawn H Corey

There is no such thing as a software patent. All software are algorithms and, by law, algorithms can NOT be patented. The same for business practices. If it manipulates information, it is, by definition, and algorithm. The only things that can be patented are inventions, that is, physical machines and real-world processes, that is, things that manipulate real-world material.

If it manipulates information, then it is, by definition, an algorithm and, by law, algorithms can NOT be patented. It's time to start holding judges criminally responsible for not enforcing the law.

Vote up!
4
Vote down!
0
Laurel L. Russwurm
Open Enthusiast

I can just see it... the patent police breaking into math classrooms around the world and arresting teachers for teaching patented equations...

Vote up!
4
Vote down!
0
Scott Dunn

While we're on the topic of a level playing field, how about lowering the "clear and convincing" evidence standard for evidence challenging patents to something more reasonable like a preponderance of evidence? The current evidence standard makes patents almost invincible in court, even really bad patents can survive most challenges of validity based on this standard.

Vote up!
6
Vote down!
0
James K.

I want the patent on adding two numbers please in any language of choice for the programmer. For adding like objects such as thing-a-majig and unlike objects such as a thng-a-majig, and a doo-hicky.

Vote up!
0
Vote down!
0
ClaimDude

Sure, here you go:

The claimed invention is:

1- A computer program code comprised in a computer readable medium, and loadable in a computer memory, said computer program code comprising instructions to cause a computer to apply an operation to at least two objects, said computer program code comprising:
computer program code instructions to cause a computer to obtain a first object, wherein said first object is a like object comprising at least one thing-a-majig;
computer program code instructions to cause a computer to obtain a second object, wherein said second object is an unlike object comprising at least one thng-a-jmajig; and
computer program code instructions to cause a computer to apply an operation to said first object and said second object, wherein said operation further comprising an adding operation.

Vote up!
0
Vote down!
0
James K.

I'm not advocating patents. I already don't like them. Someone already has a patent for a device that calculates the checksum of IP packets.... which totally amazes me that they actually accepted the patent.

Vote up!
0
Vote down!
0
Jim (JR)

Will someone please explain to me why copyright law is not sufficient? You cannot patent a musical performance, a movie, or a book, but you can copyright them which provides significant protection in and of itself.

The current "patent" laws allow someone to patent a piece of software that "takes family, income, and expense data - performs some mysterious "black-box" functions on this data, and returns a completed tax-return"

So, who is the patent infringer here? Intuit? H&R Block, QuickTax? (etc)

For all the reasons specified above, I believe that "patents" on software should not be valid.

What say ye?

Jim (JR)

Vote up!
0
Vote down!
0
Pierre MacKay

Shawn Corey's response says it all, but I offer a footnote to the main (and very timely) article. Does anyone remember zoomraks? (Surely, from the name, it will be obvious what is claimed.) What is on my display screen as I write this is an extreme violation of the zoomraks patent.

A computer program is an ordered subset of the rules of mathematical logic. In what way can the rules of mathematical logic be patented in part, if they cannot be patented in the whole?

Vote up!
0
Vote down!
0
Jim (JR)

Let's take a look at this from another angle:

Assume: Software is patentable.

Therefore: It is, (should be), subject to the same rules and burdens-of-proof that any other patent is required to show.

Therefore: In order to embody and comply with the standards of specificity and enablement - at the very least the human-readable source code for the specific invention claimed should be published as a part of the patent application. (Here I am assuming that the underlying libraries and API's need not be published, in the same way that a patentee does not need to publish the inner workings of a lathe or steel foundary to patent an object made of metal.)

However: It has already been established judicially that source-code is "speach" and is therefore protected by the First Ammendment of the Constitution.

Viz.: Software as Free Speech - DVD Copy Control Association v. Bunner, No. H021153 (Cal.App. Dist.6 11/01/2001)

http://biotech.law.lsu.edu/cases/IP/conlaw/dvdcca.htm

Likewise, it is "hornbook law" that mere speach cannot be patented.

Therefore Source-code cannot be patented.

However: In order to get a software patent, the applicant must (shouold) provide a complete listing of the source-code embodying his invention.

Therefore: Since a patent application woudl involve the publishing of something already regarded as "speach", software source code cannot be patented and therefore software itself cannot be patented.

What say ye?

Jim (JR)

Vote up!
0
Vote down!
0