Section 112 as a limit on claim breadth

[JimIvey]

Another limit to the two limits you mention, is the applicant's spec, and if the claims are broader than the disclosure enough to cover something that is not enabled, that's trouble.  It's happened !!   Those are 112 issues.  They limit the claim scope not based on what is contained in the prior art, but rather, what has been taught/enabled by the Applicant.

I'll respectfully disagree with this.

Consider this:  I describe a bubble sort in the specification.  Trust me; those of skill in the art can make and use a bubble sort.  I claim something that comprises, inter alia, "sorting the data" or "wherein the data is sorted."

My claim covers sorting by bubble sort, shell sort, or any of a vast number of types of sorts -- even those not yet invented.  Have I enabled any of these non-bubble sorts?  (I believe I haven't -- except for perhaps known sorts.)  Is my claim broader than my specification?  (I believe it is.)  Does my claim violate Section 112 for being too broad?  (I don't believe so.)

The law is quite clear that Section 112, paragraph 1, only requires enablement of one embodiment of the claimed invention.  I'm not aware of anything authoritative that says that the second paragraph of Section 112 negates that.  And, by "authoritative", I mean Federal Circuit or Supreme Court.  While the MPEP can be very informative, it is not authoritative.

Regards.

[ChrisWhewell]

I still have doubts there are practitioners that don't consider 112 as having an impact on the breadth of the language they choose in writing the claims. 

I see what you've mentioned about different sorters as an example of establishing claim element equivalency, supporting what I mentioned to another writer above referencing how to broaden the allowable scope of the claims by broadening the contents of the disclosure.     



______
In Wang v. AOL one of the reasons for the court's narrow interpretation of the claims was that the only embodiment described in the 669 spec was the character-based protocol,  and the court so limited the claims.   If that's not section 112 in action, then what is ?
Summary judgement that the defendant's system utilizing a bit-mapped protocol did not infringe was upheld. 
____
Watts v. XL - The court narrowly interpreted claim language because the spec only described one method in which tapered threads were dimensioned to achieve a seal using particular alignment of angles in the setup.  The spec did not discuss any embodiment without the particular angle alignment, else broader scope the claims would have had.  The reason for their being limited by the court, was based on the contents of the spec.

 

 

[JimIvey]

I still have doubts there are practitioners that don't consider 112 as having an impact on the breadth of the language they choose in writing the claims. 

Oh, I don't doubt that there are many practitioners who practice differently than I do.  I'm fully aware that many of my opinions, perspectives, and techniques are in the minority.  For an example, just search in here for my thoughts on provisional patent applications -- though I might not be in the minority there anymore.

I see what you've mentioned about different sorters as an example of establishing claim element equivalency, supporting what I mentioned to another writer above referencing how to broaden the allowable scope of the claims by broadening the contents of the disclosure.     

I don't mean to beat this issue to death, but this is important.  It's not equivalents.  Not at all.  The claims says "sorting" and covers "sorting" of any kind -- literally and directly, and not by the doctrine of equivalents.  Only if the claim recited "sorting by bubble sort" would the doctrine of equivalents be needed to cover sorting by other methods.

It is absolutely essential to practicing patent law to understand this fundamental distinction:  the claims (and the claims alone) define the invention and the specification is for no other purpose than to provide at least one enabling example of that invention.

In my example, the claim says "sorting" and is silent on the particular method of sorting.  So, to determine literal infringement, you simply need to determine whether he accused device/method includes "sorting."  It is entirely immaterial whether the accused device/method includes bubble sorting.  Bubble sorting makes no appearance in my example claim.  It has no role in literal infringement analysis.

So, the literal breadth of my claim is broader than my specification.  And, I believe the same is true for each every claim that does not include each and every implementation detail described in the corresponding specification -- i.e., every claim ever, other than claims written entirely in means-plus-function language.

The specification is not a limit on the breadth of claims.  Period.  At least that's how I write my applications.

Regards.

[ChrisWhewell]

Always appreciate your perspectives.  What threw me a little was the "sorting by...." language in the earlier writing.   Indeed different strategies apply to processes, vs. machine type claims.  I'm with you on the provisionals, penned a warning letter myself in an article back in 97.  Its probably still on the internet somewhere.

[JimIvey]

Sorry if I went off a bit re equivalents.  The doctrine of equivalents (DoE) is withering and, despite the reprieve by the US Supreme Court, is not very much alive and well.  In addition, my approach to writing applications is to avoid any reliance whatsoever on the DoE -- get the literal breadth of the claims right <-- priority #1.

So, equivalents is a bit of a hot button for me.  So is importing limitations into my claims that I intentionally left out.  And, it seemed like you were importing limitations into my claim and then using the DoE to capture breadth I had intended to be there originally -- albeit only in a hypothetical.

And, while we're on the topic of our tenuous hold on sanity, a word of wisdom:  Friends don't let friends go off meds.

Regards.

[khazzah]

Jim Ivey said:
>I don't mean to beat this issue to death, but this is important.  It's not
>equivalents.  Not at all.  The claims says "sorting" and covers "sorting" of any
>kind -- literally and directly, and not by the doctrine of equivalents.

I absolutely agree. Perhaps this is one of those topics that varies depending on your technology area? I don't think those of us who practice in the EE/CS area rely on DoE much, though the doctrine seems to be more important in other areas.

Perhaps those in the mechanical arts have a need for words like screw or rivet and so DoE helps to broaden. Whereas my software apps lend themselves naturally to broad words like sorting, so DoE is less important.

Jim Ivey also said:
>My claim covers sorting by bubble sort, shell sort, or any of a vast number of
>types of sorts -- even those not yet invented.  Have I enabled any of these
>non-bubble sorts?  (I believe I haven't -- except for perhaps known sorts.) 

I disagree, though maybe this is all a matter of semantics. I say that you have enabled -- just by mentioning SORT -- any species of SORT that is understood to a POSITA to be a SORT.
 
>Is my claim broader than my specification?  (I believe it is.) 

Well, your claim is broader than the explicit teachings of the specification. But not "broader than the spec" as understood by a POSITA.

>Does my claim violate Section 112 for being too broad?  (I don't believe so.)

Is this another way of stating the point you made earlier: that Section 112 does not limit claim breadth?

I think I disagree with this too, though maybe again we're disagreeing on semantics. Do you agree with this statement: claims are invalid for lack of enablement when the construed claim is broader than the spec, as understood by a posita. If so, this leads me to the conclusion that 112 does limit claim breadth.

>The law is quite clear that Section 112, paragraph 1, only requires
>enablement of one embodiment of the claimed invention.

I don't disagree with that, but isn't there an important corollary? Namely, that if an embodiment covered by the construed claims is not enabled, that claim is invalid. 

As in Sitrick vs. Dreamworks, where the asserted claims were construed as covering both movies and video games, and the claim was found invalid because the spec didn't enable the movie embodiment. In my view, the spec in Sitrick did limit the claims to a video game embodiment.

Jim, I apologize if I'm misunderstanding you or going down an irrelevant path.

[ChrisWhewell]

Sorry if I went off a bit re equivalents. 

Its an important topic.  I was thinking along the lines of a machine, or a system in which its important to disclose multiple embodiments, multiple alternatives for each element in the embodiments, multiple features, when possible.  Reason is that Fed Circ. has invalidated broad claims or ruled as non-infringed patents in which the spec only contained enabling description of one embodiment.  Gentry Gallery, Tronzo, and Toro cases.

[Isaac]

I don't disagree with that, but isn't there an important corollary? Namely, that if an embodiment covered by the construed claims is not enabled, that claim is invalid.

Your statement is true for correct as far as it goes and for some definition of covered.  But enablement does not mean literally describing every possible embodiment within the scope of the claims.  For one thing, you are not required to disclose the details of things known in the art.  So giving example an example of a sort and allowing that others can be used is probably enough to enable claims that encompass other known sorting techniques.  No undue experimentation required to substitute one known technique for another.   On the other hand, if you try to add a new claim during prosecution reciting a bubble sort, when you've only described a quick sort embodiment, you may have a written description problem but probably not an enablement problem.

As in Sitrick vs. Dreamworks, where the asserted claims were construed as covering both movies and video games, and the claim was found invalid because the spec didn't enable the movie embodiment. In my view, the spec in Sitrick did limit the claims to a video game embodiment.

That's quite a bit different from not describing every possible sorting algorithm.  There was no way to extend the video game embodiment which involved identifying which character is speaking in a video game, likely a trivial task, to identifying which on-screen character in a movie is speaking, something decidedly more complex.   Enablement is a far more complex analysis than written description.

[JimIvey]

So, if I claim two or more servers cooperating to achieve some really nifty result, and I don't explain how 2,341,576,980 servers can cooperate, my claim is not enabled?  Note that "two or more servers" covers 2,341,576,980 servers.  Note also that, since the different numbers of servers covered by "two or more" is infinite, it seems to be that no claim reciting "n or more" of anything could ever enable each and every embodiment of that claim and all such claims ought to be rejected under Section 112, p1.  Same for "at least n" and any claim using "comprising".

Honestly, I've never known how to reconcile the line of cases stating unequivocally that enablement of each and every embodiment is required by Section 112 with those stating that "the full scope of the claims" must be enabled under Section 112.

I've only briefly reviewed that opinion, so I may be off-base here.  It looks like the claimed invention related to "skinning" in video games with user-supplied images.  I don't see how it is not enabled for movies.  Movies using CGI use essentially the same technology and tools used by video games, except that processing parameters are tweaked for performance in video games and for realism in movies.

Perhaps Dreamworks so modified the CGI engine that the technique described in the patent in question really didn't apply.  In that case, it should have been decided using the reverse doctrine of equivalents rather than invalidating the claim under 112 for not enabling each and every embodiment that the claim covers.

So, which is true?  Section 112 does or does not require enablement of each and every embodiment of a claim?  If enablement is only required for some (more than just one) but not all embodiments, which embodiments must be enabled?

Regards.

[dablueman]

It looks like the claimed invention related to "skinning" in video games with user-supplied images.  I don't see how it is not enabled for movies.  Movies using CGI use essentially the same technology and tools used by video games, except that processing parameters are tweaked for performance in video games and for realism in movies.

Rendering movies typically uses ray tracing which gives you very high quality images but requires much more time to render, while video games use rasterization to give you real-time images. 

That being said, I don't believe generally that overly broad claims raise 112 questions. If a species of the claim is enabled in the spec, I don't understand how the claim isn't enabled.

[JimIvey]

Rendering movies typically uses ray tracing which gives you very high quality images but requires much more time to render, while video games use rasterization to give you real-time images. 

Well, this is getting a bit off-topic, but....

I'll admit ignorance, but don't movies use some skinning, for example to show dirt and oil stains on the hull of ship?  Or do they model oil stains individually? 

And, don't video games use some ray tracing?  Yeah, they do it once for each polygon to get general luminance information before applying the skin, but they do it that once, don't they?, for each polygon on each frame.

And, there are also varying degrees of each in various instances.  My daughter still loves The Backyardigans, and that is very obviously skinned -- and not nearly as well as even simple video games these days.  Wii games look more realistic.

My experience in developing in this area is very, very old.  Left it for law school in 1988.

One last observation, my very quick scan of the opinion didn't reveal much about infringement, only invalidity.  Is it possible that Dreamworks didn't have strong arguments for non-infringement?  Is it possible that Dreamworks was, in fact, applying the technique(s) claimed and enabled, albeit in a different context? 

Regards.

[ChrisWhewell]

I think in Ariad Pharma v. Eli Lilly, a court declared a genus invalid because the spec didn't enable all the species.

[JimIvey]

Time again for my periodic disclaimer.  I understand that chemical/life science practice has many special rules and such and I don't practice in those technologies.  Genus/species stuff makes my head hurt.  I almost never see that stuff in my practice.

Regards.

[ChrisWhewell]

There's currently an unresolved issue looming, whether written description and enablement are one and the same, or two different requirements.    Amgen v. Hoechst in 2003 said sec 112 mandated two separate and independent requirements.   A petition for rehearing in banc was filed in August in Ariad v. Lilly relating to the question.  My guess is that the eventuality of it may impact all areas of practice.   One source of confusion is that "to enable" appears as the object of the subject "written description" in the statute, yet many use enablement as a noun.  Written description is question of fact, enablement question of law based on facts.   If the same exact set of facts are necessarily used in determining sufficiency of each, can they be separate and independent ?

[JimIvey]

Yeah, the "written description" requirement thing is just too vague to really grok in any meaningful way.  The "written description" shows that the inventor "got it" -- understood what she was inventing.  The "full scope" of what is claimed appears to be related to the heart, gist, essence of the invention, and it seems that every case that notes that must remind us all that there is no protectable heart, gist, or essence of an invention in any patent.  That goes to my question of what the "full scope" of the claimed invention is if it's not each and every embodiment of the claimed invention. 

So, going back to the Dreamworks case, the heart, gist, essence of the invention applied to both video games and movies and the inventor didn't realize it applied to movies, so the claim is invalid.  So, the claims is invalid and therefor covers neither movies nor video games, all because the inventor didn't fully realize that the same innovation could be applied to another technology.

Have I completely misrepresented this area of the law?  It seems completely ridiculous and absurd.

One thought on this:  on occasion, I see a patent asserted where it really shouldn't be.  Please don't ask me to elaborate.  I won't.  Perhaps this is just one of those harsh rules to spank patent owners who over-reach.  Honestly, I still think the reverse doctrine of equivalents is a better tool for this.

As someone who writes patent applications, there's really not much I can do to avoid this particular pitfall.  On the other hand, I wonder if a single sentence in the Dreamworks patent would have made all the difference:  "To the extent movies generated using CGI techniques [have this or that quality], images can be interjected into movies in the manner described herein."

There's an age-old policy debate about what to do with benefits of an innovation that the inventor(s) fail(s) to realize is/are there.  Do you say "gotcha!" and donate the unrealized value of the inventor's contribution to the public?  Or do you give to the inventor the full value of her contribution even if she doesn't realize that full value herself?  I've always pushed for the latter.

Regards.