Print

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

[khazzah]

Splitting off a new thread from what started as 112 Guidelines, but became an interesting discussion about obviousness...

It's DRM for digital music (or media of any type I suppose).  The music was bound to a key and the player displayed part of the key during playback.

Obvious?  It's just high-level display of text during music playback.  It's gotta be obvious, right? 

Under the KSR mantra of "Combining prior art elements according to known methods to yield predictable results", then yeah, this invention would probably be obvious.

Music playback + display of text = predictable result of text display during music playback. Assuming there's nothing special about how you're combining the two features, ie, you're doing it according to known methods.

This rationale seems to set a very low threshold for a prima facie case, at least in the computer/electronics arts. Seems to boil down to nothing more than since two elements *could* be combined, it's obvious to do so. That just seems wrong. It seems like this KSR rationale does away not only with motivation in the references themselves, but a showing of motivation of any kind.

Well, the kicker was that the part of the key that was displayed included the credit card information used to purchase the song.  The idea was to tie playback of the media to display of information that the media owner would prefer not to share.

Ah, your facts include a reason why one would *not* combine the features, ie the media owner doesn't want his credit card info displayed.

I wonder how this countervailing factor plays into the KSR analysis.

The way I see it, it's still a predictable result using a known method. So is this countervailing factor best viewed as a strong rebuttal of a prima facie case based on the predictable results rationale? IOW, there is a prima facie case, but you beat it by showing a reason not to combine? 

[JimIvey]

Under the KSR mantra of "Combining prior art elements according to known methods to yield predictable results", then yeah, this invention would probably be obvious.

You've left out an important part of KSR (emphasis supplied):

The combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results.

Sure.  If I claim "displaying text during playback of media content", I've probably recited the "combination of familiar elements according to known methods [that] does no more than yield predictable results."

On the other hand, if I claim "displaying payment authorization information of the user during playback of media content", the results include something other than display of information during media content playback.  The results include a co-opting the desire for privacy of a certain type of information to leverage a desire for privacy of media content the user would otherwise be happy to share with the whole world.

The "no more than" is very important.  In my opinion, not only should that phrase be included in every quote and paraphrase of KSR of that text, but that phrase should always be highlighted to remind people that it's there.

It's not just my own perspective as an advocate for patent owners and applicants that makes that phrase important in my opinion.  The phrase is generally intrinsically important, like "not" and "only".  To illustrate, consider the following two sentences (imagine yourself as the speaker).  "Go ahead and take $5 out of my wallet."  "Go ahead and take no more than $5 out of my wallet."  Think the difference is important?

Regards.

[khazzah]

You've left out an important part of KSR (emphasis supplied):

The combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results.

You're right: my shorthand referral to the rationale left that part out. I just got sloppy. I do realize that piece is there too.

Sure.  If I claim "displaying text during playback of media content", I've probably recited the "combination of familiar elements according to known methods [that] does no more than yield predictable results."

On the other hand, if I claim "displaying payment authorization information of the user during playback of media content", the results include something other than display of information during media content playback. 

Ah, I didn't realize that you *claimed* the feature which explains why a POSITA would not combine. Didn't think it through, I guess.

The "no more than" is very important.  In my opinion, not only should that phrase be included in every quote and paraphrase of KSR of that text, but that phrase should always be highlighted to remind people that it's there.

Good point. And now that I see that you claimed the unpredictable result, I think the rationale does not fit. That is, if the Examiner rejects under the rationale, you explain that you have *more* than predictable results, and you should win.

So I've answered my own question. This example is not a rebuttal of a good prima facie case. Instead, the prima facie case is faulty.scenario.

[khazzah]

If I claim "displaying text during playback of media content", I've probably recited the "combination of familiar elements according to known methods [that] does no more than yield predictable results."

On the other hand, if I claim "displaying payment authorization information of the user during playback of media content", the results include something other than display of information during media content playback.  The results include a co-opting the desire for privacy of a certain type of information to leverage a desire for privacy of media content the user would otherwise be happy to share with the whole world.

OK, I agree that displaying credit card info during media playback is "something other than display of information during playback".

But why is "displaying credit card info during media playback" an unpredictable result?

Why is it not the predictable result of combining the known feature (a) playing media and the known feature (b) displaying credit card info? (Assuming that displaying credit card info is known.)

Seems to me that only-predictable-result vs. other-than-predictable-results turns on how you frame which features are known.

I'm still convinced by your argument that the combination is non-obvious, for the reasons you suggest: persons generally want to keep this info private.

But I'm still struggling with what we mean by "no more than predictable results."

[klaviernista]

Ah, your facts include a reason why one would *not* combine the features, ie the media owner doesn't want his credit card info displayed.

I wonder how this countervailing factor plays into the KSR analysis.

I might be understanding your question incorrectly, but I think the USPTO addressed this issue in its KSR guidelines, which referenced Crocs Inc. v. US ITC.  Arguments that the prior art teaches away from the claimed invention are still available post KSR.
 
Guidelines available here: http://edocket.access.gpo.gov/2010/pdf/2010-21646.pdf

See, e.g., example 4.2 (excerpt below).

Example 4.2. Crocs, Inc. v. U.S. International Trade Commission, 598 F.3d 1294 (Fed. Cir. 2010). Teaching point: A claimed combination of prior art elements may be nonobvious where
the prior art teaches away from the claimed combination and the combination yields more than predictable results.

[khazzah]

Ah, your facts include a reason why one would *not* combine the features, ie the media owner doesn't want his credit card info displayed.

I wonder how this countervailing factor plays into the KSR analysis.

I might be understanding your question incorrectly, but I think the USPTO addressed this issue in its KSR guidelines, which referenced Crocs Inc. v. US ITC.  Arguments that the prior art teaches away from the claimed invention are still available post KSR.

You did understand my question correctly. And it's helpful to the forum that you pointed out the Crocs decision.

But reading your comment made me realize that my question wasn't well-phrased. I didn't mean how does reason-not-to-combine fit into the KSR obviousness analysis in general. I meant how does reason-not-to-combine fit into the framework of the various KSR *rationales*.

In particular, the "predictable results" rationale. That is, is offering a reason not to combine an attack on an insufficient prima facie case, or is it a rebuttal to a sufficient prima facie case? [Perhaps the point is academic. I'm not sure.]

I view "teaching away" as one species of the genus "reason not to combine". I'm not sure whether Jim's argument that "a POSITA wouldn't display credit card info while playing music because that info is usually kept private" constitutes teaching away. The reason not to combine we're talking about is a general understanding rather than a specific teaching in the references.

But an important point of this thread, and your contribution to it, is that a credible reason-not-to-combine should go a long way to beating an obviousness rejection.

[ManOfManyBadIdeas]

I have to say I share Karen's concern. Among the rationales outlined in KSR, the "known methods, predictable results" rationale makes me a bit uneasy, because motivation isn't one of its elements. It's not required to be addressed, at all. To that end, I feel a "teaches away" argument alone would be largely irrelevant to combat that particular rationale, at least at the USPTO level. Crocs cite here is not clear cut universally applicable, because I think the "known methods" rationale wasn't properly applied. The "teach away" argument was directed at the foam straps, which were not known, which challenges the applicability of "known elements" component of the rationale. In that context prior art teaches away not from the combination, but from modifying a known element to use a new material. I think the only way to combat this rationale, like any other, is to point out its deficiencies. Lack of motivation to combine isn't one of this rationale's deficiencies. Producing more than predictable results, however, is. The reason motivation did pop up in the Crocs case IMO, is that "known elements with predictable results" rationale wasn't properly applied, since foam strap wasn't a known element. To draw a parallel to the DRM example here, applying the "known elements" rationale to it isn't proper if CC info display isn't known, only general info display is known. One way to argue it is lack of motivation to combine (display CC info is undesirable). Another way to argue it is directly pointing out that CC info display isn't known, and shows more than predictable results compared to general info display (aversion to sharing).

Bottom line - I think "known method predictable results" rationale should be rarely applicable (any element's modification, or combination method modification, or additional specific utility kills it). And I think any response to it must point out such fatal deficiencies explicitly. I do not consider "teaching away" as a fatal flaw of "known method predictable results" rationale.

[JimIvey]

But I'm still struggling with what we mean by "no more than predictable results."

Really?  I think it's synonymous with "predictable results and nothing else" or "predictable results and absolutely no unpredictable results".

So, displaying confidential information (such as a credit card number) during audio playback has a number of results.  A result of the display of the confidential information is predictable.  So, a predictable result.  Obvious?  No. 

We only need to show one unpredictable result to show "more than predictable results".  The unpredictable result here is that the purchaser of a copy of someone else's audio data diligently guards the audio data (or at least the key needed to play it back) from unauthorized copying.

All you need is one unpredictable result to get beyond "no more than predictable results".

Regards.

[ManOfManyBadIdeas]

So, displaying confidential information (such as a credit card number) during audio playback has a number of results.  A result of the display of the confidential information is predictable.  So, a predictable result.  Obvious?  No. 

I think there is an important distinction between predictable results of known elements/methods and predictable results of new elements/methods. I think the latter doesn't belong in a "known methods, predictable results" rationale. So, if displaying info during playback is known, but CC info display is not, then sharing aversion is not a predictable result of what's known (general info display). The latter case, the expectation that CC info display will result in sharing aversion, is not what one should call predictable result under the rationale, but rather the expectation that the invention would be useful in the way it is described without having to demonstrate it.

[JimIvey]

I think there is an important distinction between predictable results of known elements/methods and predictable results of new elements/methods.

Yeah, I think that's a good point. 

Before you even get to whether display of confidential information during playback is a predictable result, you have to produce a number of prior art references that, collectively, teach all limitations in the claims.  So, without some reference teaching some sort of display of confidential information as sharing aversion, the combination doesn't collectively teach all limitations and whether the combination is obvious for yadda yadda yadda producing no more than predictable results is immaterial.

First, you have to find references that teach each and every limitation of the claim. 

Only then, even without a teaching, suggestion, or motivation in the prior art to combine the references in the manner they're combined to show obviousness, the combination can nonetheless be proper if the references are combined using known methods (also in the prior art) to produce no more than predictable results (i.e., results shown to be predictable in the prior art).

If the combination produce an unpredictable result, the combination is not obvious.  But, in cases in which the prior art collectively fails to teach a claim limitation, we never get to that part of the analysis.

I sense a general inclination of those having recently read KSR to start with the claim language and then look for unpredictable results.  That's not the proper inquiry.  You never start with the claims.  That's clearly impermissible hind-sight reasoning.

Getting back to the confidential information display during playback, suppose a prior art reference was found that taught the display of confidential information in some other context to motivate someone in the other context to carefully guard the privacy of something else in that context, then the "predictable results" might come into play.  In other words, the "predictable results" could support combination of that reference with another reference teaching display of textual information of some other sort during playback of audio data. 

I hope that makes sense.

Regards.

[khazzah]

We only need to show one unpredictable result to show "more than predictable results".  The unpredictable result here is that the purchaser of a copy of someone else's audio data diligently guards the audio data (or at least the key needed to play it back) from unauthorized copying.

I'm just not seeing this combination of claimed features as an unpredictable result.  Sure, one result of the combination is that confidential information is displayed in a manner which allows the info be viewed by others. This result is undesirable. And therefore a reason not to combine the display of confidential info with media playback. And therefore non-obvious.

But to me the result itself seems perfectly predictable: if I combine a) media playback with b) displaying confidential information in a manner which allows the info to be viewed by others, I get ... display of confidential information during media playback. How is this unpredictable?

It seems like you and I are reading "unpredictable" in very different ways. 

[JimIvey]

It seems like you and I are reading "unpredictable" in very different ways. 

Yes.  I agree.  Well, sort of.

First, you only get to predictable results if there's a collection of references that collectively teach the limitations of the claim.  In that particular patent (DRM), I don't believe there was any reference teaching a binding of publication of private information to something such that the owner of the private information would also protect privacy of the bound thing. 

As a result, there were inadequate "familiar elements" to combine "according to known methods" to do "no more than yield predictable results."

So, perhaps this key binding claim isn't a good example as it never even gets to the "predictable results" analysis due to inadequate "familiar elements". 

Of course, what I'd expect examiners trained under the Dudas regime to argue is that "confidential information" is just text and to examine the claim as if "confidential information" had been amended to recite "text".  Of course, the easy rebuttal is to simply ask for some text and then to ask for the text on their credit card(s).  If there's no meaningful distinction, they shouldn't mind sharing all text -- since text is text.  (Can you tell I argue against this sort of nonsense all the time?)

Getting back to "predictable results", I think you're looking at the claim and looking for something unpredictable.  That's not the inquiry.  That's clearly impermissible hindsight reasoning.  And, if that's how that passage was implemented, you'd be absolutely right -- any claim not reciting something spectacular and unexpected would be obvious.

Here's how I see that passage of KSR.  Let's say the prior art of record teaches all the limitations of a claim (familiar elements).  Let's say that the methods by which the teachings of that prior art collectively can be combined is known (combined using known methods).  Lastly, let's say that the results of the combination would be predictable, generally speaking (producing no more than predictable results).  Without showing all 3 of those elements, the quoted passage of KSR re predictable results does not come into play.

Once those 3 things are shown, I would expect that the burden shifts to the applicant to show that something extraordinary happens with the combination, some results that are not predictable.  If we can't allow the PTO to force us to prove a negative, we should expect them to prove the negative of no unpredictable results.

Let's just say that, as I think over recent arguments I've made regarding non-obviousness, this passage of KSR does not affect the way I amend/argue in any way. 

I still argue that the combination does not collectively teach each and every limitation of the claim (inadequate familiar elements).

I still argue no motivation to combine or that the combination would be inoperable or any of the other improper combination arguments (not combined using known methods).

I still argue unpredictable results (no more than predictable results), if I ever get that far and really need to.  That can involve things like teaching away, or just counter-intuitive results.

Regards.

[khazzah]

So, perhaps this key binding claim isn't a good example as it never even gets to the "predictable results" analysis due to inadequate "familiar elements". 

You've convinced me that the predictable results rationale is inappropriate for your DRM key binding example. That is, you should be able to win by showing the Examiner did not meet a prima facie case.

Getting back to "predictable results", I think you're looking at the claim and looking for something unpredictable.  That's not the inquiry.  That's clearly impermissible hindsight reasoning. 

I can see why you said my analysis was hindsight. I still can't see how you can apply the rationale to your DRM example without assuming that the combination exists, which would see to be impermissible hindsight.

But, maybe that is yet another indicator that your DRM example isn't appropriate for this rationale.

Without showing all 3 of those elements, the quoted passage of KSR re predictable results does not come into play.
Once those 3 things are shown, I would expect that the burden shifts to the applicant to show that something extraordinary happens with the combination, some results that are not predictable. 

So once the PF case of predictable-results-rationale is made, you *do* assume the combination exists. And the Applicant shows why in this combination, "something extraordinary happens ... some results that are not predictable." 

Let's just say that, as I think over recent arguments I've made regarding non-obviousness, this passage of KSR does not affect the way I amend/argue in any way. 

I still argue that the combination does not collectively teach each and every limitation of the claim (inadequate familiar elements).

I still argue no motivation to combine or that the combination would be inoperable or any of the other improper combination arguments (not combined using known methods).

I still argue unpredictable results (no more than predictable results), if I ever get that far and really need to.

Thanks for the explanation. Now I have a better understanding of the framework for the predictable results rationale.

Strangely, the Examiners still use teaching-suggestion-motivation in most of the cases I prosecute. So I haven't had too much practical experience with addressing the predictable results rationale.

[ManOfManyBadIdeas]

I think there is an important distinction between predictable results of known elements/methods and predictable results of new elements/methods.

Yeah, I think that's a good point. 

Before you even get to whether display of confidential information during playback is a predictable result, you have to produce a number of prior art references that, collectively, teach all limitations in the claims.  So, without some reference teaching some sort of display of confidential information as sharing aversion, the combination doesn't collectively teach all limitations and whether the combination is obvious for yadda yadda yadda producing no more than predictable results is immaterial.

First, you have to find references that teach each and every limitation of the claim. 

Only then, even without a teaching, suggestion, or motivation in the prior art to combine the references in the manner they're combined to show obviousness, the combination can nonetheless be proper if the references are combined using known methods (also in the prior art) to produce no more than predictable results (i.e., results shown to be predictable in the prior art).

If the combination produce an unpredictable result, the combination is not obvious.  But, in cases in which the prior art collectively fails to teach a claim limitation, we never get to that part of the analysis.

I sense a general inclination of those having recently read KSR to start with the claim language and then look for unpredictable results.  That's not the proper inquiry.  You never start with the claims.  That's clearly impermissible hind-sight reasoning.

Getting back to the confidential information display during playback, suppose a prior art reference was found that taught the display of confidential information in some other context to motivate someone in the other context to carefully guard the privacy of something else in that context, then the "predictable results" might come into play.  In other words, the "predictable results" could support combination of that reference with another reference teaching display of textual information of some other sort during playback of audio data. 

I hope that makes sense.

Regards.

It certainly does. Karen's remark about not seeing much of this particular rationale reinforces my impression that it's not supposed to be a typical, mainstream line of reasoning. To me, it looks like a shortcut that the examiner can use in the most clear cut, apparent cases of obviousness. It's a time saver, sort of like a "Official Notice" should be. I say should be because it apparently is being abused, and I see that potential in the "predictable results" rationale. I think you have outlined the right approach to demonstrating improper application of this rationale, such as the Crocs, or your largely parallel DRM case. I do expect the properly applied "predictable results" rationale to be just as hard to argue as the properly applied Official Notice of Fact. So, to me, as long as there is a shred of doubt in the obviousness of the combination, the "predictable results" rationale most likely is applied improperly, and one just has to carefully go through all of its required elements to demonstrate lack of PF case.

[JimIvey]

Strangely, the Examiners still use teaching-suggestion-motivation in most of the cases I prosecute.

Same here, though I don't find it strange.  Though, I still find that such rationales have very weak, apparently completely made-up, motivations to combine -- without even the weakest assertion that the proffered motivation would have been known to an ordinary artisan at the appropriate time.

I believe that TSM was a formulation that was intended to make obviousness rejections easier for the Office.  If you can find a TSM in the prior art, you don't need to show the level of ordinary skill in the art at the relevant time.  You just need an art in which the ordinary artisan would be literate.  Once the Office tried so hard to escape the frying pan to the safe confines of the fire, I really pushed them on failing to show the level of ordinary skill in the relevant art(s) at the relevant time.

The post-KSR trend of trying to push obviousness rejections without any sort of TSM nexus appears to be waning.  My sense of this is that, while KSR made it clear that TSM is not the only way to show obviousness, TSM turns out to be a convenient way to show obviousness.  And, if done well, a good TSM argument for obviousness is difficult to overcome without amendment.

Getting back to predictable results briefly, this discussion has clarified some of the earlier comments about computer-implemented innovations.  It may well be that, once a computer-implemented invention gets to the predictable results part of the analysis, it's game-over for that claim (excepting amendment, of course).  In other words, it may well be that the results of combining computer-implemented elements using known methods generally produces only predictable results.  So, perhaps the fight is really about whether all elements of the claim are "familiar elements" and whether the combination is strictly by "known methods".  Though, I'm not yet ready to completely give up on the possibility that a computer-implemented method can produce unexpected, unpredictable results.

Regards.