[Macworld]

Post your comments for Judge cancels Motorola, Apple patent trial here

[pcharles]

Does Posner get the 2012 sanity award for this decision?

[bastion]

pcharles, on 08 June 2012 - 05:27 AM, said:

Does Posner get the 2012 sanity award for this decision?

Only if he's actually right, and was correct in introducing the previous limitations. This bears a certain resemblance to Apple's case against Microsoft in the late 1980s. Microsoft had compensated Apple for a license to copy certain elements of the Mac user experience in Windows, and they ended up going to court over a dispute regarding the termination conditions of the license - specifically to how many releases of Windows it applied. In that case, the judge pretty much ignored the suit in front of him *and* the gestalt user experience under discussion. Instead he chose to step through atomic components of the interface and declare them ineligible for protection one by one, after which he simply ruled the license to be invalid in that there was nothing to license.

This is less obviously egregious but certainly could stand some scrutiny.

[zarmanto]

bastion, on 08 June 2012 - 06:36 AM, said:

This is less obviously egregious but certainly could stand some scrutiny.

Agreed. It sounds to me like Posner is basically trying to say, "Get out of my kitchen (court), and go play nice with the other kids! And stop beating up on (suing) each other all the blasted time!" As a parent, I generally agree with such an approach in principle, but the problem as I see it (and probably as many other Apple aficionados see it) is that Apple has a legitimate beef with Motorola stealing their IP, whereas Motorola was just trying to deflect Apple's suit by fabricating a suit of their own.

So in the end, if Posner actually does dismiss this case without awarding "damages" to either side, as he seems to be suggesting, then he is effectively granting the "win" to Motorola, and by extension, to all Android manufacturers.

I wonder what kind of phone Posner uses?

This post has been edited by zarmanto: 08 June 2012 - 07:17 AM

[Stewsburntmonkey]

zarmanto, on 08 June 2012 - 07:12 AM, said:

bastion, on 08 June 2012 - 06:36 AM, said:

This is less obviously egregious but certainly could stand some scrutiny.

Agreed. It sounds to me like Posner is basically trying to say, "Get out of my kitchen (court), and go play nice with the other kids! And stop beating up on (suing) each other all the blasted time!" As a parent, I generally agree with such an approach in principle, but the problem as I see it (and probably as many other Apple aficionados see it) is that Apple has a legitimate beef with Motorola stealing their IP, whereas Motorola was just trying to deflect Apple's suit by fabricating a suit of their own.

So in the end, if Posner actually does dismiss this case without awarding "damages" to either side, as he seems to be suggesting, then he is effectively granting the "win" to Motorola, and by extension, to all Android manufacturers.

I wonder what kind of phone Posner uses?

It will be interesting to see the full ruling, but as an Apple aficionado, I don't feel either of the patents involved should ever have been granted.

One deals with automatically detecting "structures" like phone numbers in text and then allowing users to perform functions with them, like tap on a phone number to place a call. This sort of thing is completely obvious (as other structures like URLs have been treated this way for a long time).

The other deals with APIs used for realtime processing. It is such a broad claim I see no reason for it to have been granted any protection.

In the end while I think Apple does have some legitimate complaints with regard to Android and Android devices, I think most of this demonstrating how broken the patent system is with regard to this area of technology. I would love to see more legal scrutiny of the validity of many of these sorts of patents.

[bastion]

Stewsburntmonkey, on 08 June 2012 - 07:36 AM, said:

It will be interesting to see the full ruling, but as an Apple aficionado, I don't feel either of the patents involved should ever have been granted.

One deals with automatically detecting "structures" like phone numbers in text and then allowing users to perform functions with them, like tap on a phone number to place a call. This sort of thing is completely obvious (as other structures like URLs have been treated this way for a long time).

Completely different. URIs are self-identifying and conform to a formal, simple, standardized structure. Apple's data detector feature is doing much more complex idiomatic pattern matching. And keep in mind patent eligibility doesn't require the *idea* to be inobvious, but rather the solution.

[Stewsburntmonkey]

bastion, on 08 June 2012 - 08:05 AM, said:

Stewsburntmonkey, on 08 June 2012 - 07:36 AM, said:

It will be interesting to see the full ruling, but as an Apple aficionado, I don't feel either of the patents involved should ever have been granted.

One deals with automatically detecting "structures" like phone numbers in text and then allowing users to perform functions with them, like tap on a phone number to place a call. This sort of thing is completely obvious (as other structures like URLs have been treated this way for a long time).

Completely different. URIs are self-identifying and conform to a formal, simple, standardized structure. Apple's data detector feature is doing much more complex idiomatic pattern matching. And keep in mind patent eligibility doesn't require the *idea* to be inobvious, but rather the solution.

A URL is actually much harder to detect reliably than a phone number (which is the example Apple used in the suit), but simply making better regular expressions or pattern recognition rules isn't grounds for a valid patent.

In looking through the patent it doesn't specify the specific solution, beyond the basic components, which are just standard programming components (API, server, etc.).

[zarmanto]

Stewsburntmonkey, on 08 June 2012 - 08:24 AM, said:

bastion, on 08 June 2012 - 08:05 AM, said:

Completely different. ...

A URL is actually much harder ...

Alright now, boys, settle down. Don't make me come over there and blister your hind-sides with my lawsuit!

[bastion]

Stewsburntmonkey, on 08 June 2012 - 08:24 AM, said:

A URL is actually much harder to detect reliably than a phone number ....

I'm sorry, but that's a ridiculous claim. Someone with a copy of the IETF RFC in-hand (http://www.ietf.org/rfc/rfc1738.txt) can write a perfect URL scanner from scratch in an hour. It's a large part of the point behind the 'U' in URL. Phone numbers, in contrast, are nothing like "uniform." They have different "standard" presentations world-wide and local and personal deviations in presentation that make identifying and isolating them a much harder problem.

It's also irrelevant. Your post - and thus my response - were talking about obviousness which is a consideration for patentability. Complexity is an orthogonal concept and *not* a consideration for patentability.

And that's leaving aside the fact that data detectors work on a broader range of information types than just phone numbers. I think you've drastically underestimated just how much the data detectors engine is actually doing. It's better at identifying interesting pieces of information and reasonable operations to perform on them than many people I've met.

Quote

but simply making better regular expressions or pattern recognition rules isn't grounds for a valid patent.

Well, yes. Actually better pattern recognition is certainly protectable. You may not *like* that, but it is reality. The biggest part of the trick is determining with reasonable reliability what you're look at in the first place. It can, for example, identify that a block of text describes a complete event, including date, starting and ending times, location and potential description. That's not just picking up a string of digits that have the right length to be a phone number for the current locale and taking a stab that that's what it might be. It's doing substantial processing of the surrounding context as part of it's evaluation.

[Stewsburntmonkey]

bastion, on 08 June 2012 - 08:58 AM, said:

I'm sorry, but that's a ridiculous claim. Someone with a copy of the IETF RFC in-hand (http://www.ietf.org/rfc/rfc1738.txt) can write a perfect URL scanner from scratch in an hour. It's a large part of the point behind the 'U' in URL. Phone numbers, in contrast, are nothing like "uniform." They have different "standard" presentations world-wide and local and personal deviations in presentation that make identifying and isolating them a much harder problem.

You are assuming URLs always follow the RFC exactly, but even if they did the RCF specification is much more complex than one might naively think. That said, one can write a phone number detector in a matter of minutes which catches the vast majority of numbers accurately.

bastion, on 08 June 2012 - 08:58 AM, said:

It's also irrelevant. Your post - and thus my response - were talking about obviousness which is a consideration for patentability. Complexity is an orthogonal concept and *not* a consideration for patentability.

You are the one who brought up complexity…

bastion, on 08 June 2012 - 08:58 AM, said:

And that's leaving aside the fact that data detectors work on a broader range of information types than just phone numbers. I think you've drastically underestimated just how much the data detectors engine is actually doing. It's better at identifying interesting pieces of information and reasonable operations to perform on them than many people I've met.

That's fine, but it doesn't make it any less obvious as far as patentable concept is concerned.

bastion, on 08 June 2012 - 08:58 AM, said:

Well, yes. Actually better pattern recognition is certainly protectable. You may not *like* that, but it is reality. The biggest part of the trick is determining with reasonable reliability what you're look at in the first place. It can, for example, identify that a block of text describes a complete event, including date, starting and ending times, location and potential description. That's not just picking up a string of digits that have the right length to be a phone number for the current locale and taking a stab that that's what it might be. It's doing substantial processing of the surrounding context as part of it's evaluation.

I would argue whether a better pattern matcher is in general patentable, but the patent doesn't actually describe the details of the pattern matching.

My previous job involved writing crawlers to extract event information from webpages, maybe that's why I'm less than impressed with the patent. It is a difficult problem, but just because something is hard doesn't mean the solution is patentable (nor should it). To me this patent describes a very standard software engineering methodology for solving a specific problem. That is not a novel, non-obvious invention or discovery.

This post has been edited by Stewsburntmonkey: 08 June 2012 - 09:38 AM

[bastion]

Stewsburntmonkey, on 08 June 2012 - 09:37 AM, said:

bastion, on 08 June 2012 - 08:58 AM, said:

I'm sorry, but that's a ridiculous claim. Someone with a copy of the IETF RFC in-hand (http://www.ietf.org/rfc/rfc1738.txt) can write a perfect URL scanner from scratch in an hour. It's a large part of the point behind the 'U' in URL. Phone numbers, in contrast, are nothing like "uniform." They have different "standard" presentations world-wide and local and personal deviations in presentation that make identifying and isolating them a much harder problem.

You are assuming URLs always follow the RFC exactly, but even if they did the RCF specification is much more complex than one might naively think. That said, one can write a phone number detector in a matter of minutes which catches the vast majority of numbers accurately.

Surreal. First of all, my familiarity with the RFC for URLs is not naive. As RFCs go, it's not particularly complex. There are many optional elements but the structure itself is really quite simple and rigorously defined. Yes, I am assuming that a URL will conform to the RFC. I think I'm quite justified in saying that if it doesn't conform to the RFC it is by definition not a URL. That's the whole point of a standard.

Your phone number counter-argument is spurious, not least because it's simply incorrect. Or, I should say, very provincial.

Quote

bastion, on 08 June 2012 - 08:58 AM, said:

And that's leaving aside the fact that data detectors work on a broader range of information types than just phone numbers. I think you've drastically underestimated just how much the data detectors engine is actually doing. It's better at identifying interesting pieces of information and reasonable operations to perform on them than many people I've met.

That's fine, but it doesn't make it any less obvious as far as patentable concept is concerned.

There's your problem: You don't patent concepts. You patent mechanisms (physical or otherwise). The concept "scan a stream of data looking for substreams that appear to represent certain kinds of information and present them to the user with options to operate on them in useful ways" is, relatively, obvious. But that's not what the patent is on. The patent is on a specific technique for achieving that goal.

Quote

My previous job involved writing crawlers to extract event information from webpages, maybe that's why I'm less than impressed with the patent. It is a difficult problem, but just because something is hard doesn't mean the solution is patentable (nor should it). To me this patent describes a very standard software engineering methodology for solving a specific problem. That is not a novel, non-obvious invention or discovery.

So you can cite prior art then? And not just something that performs the same task but does so in the same way? Excellent. I look forward to seeing it.

[Stewsburntmonkey]

bastion, on 08 June 2012 - 12:53 PM, said:

Surreal. First of all, my familiarity with the RFC for URLs is not naive. As RFCs go, it's not particularly complex. There are many optional elements but the structure itself is really quite simple and rigorously defined. Yes, I am assuming that a URL will conform to the RFC. I think I'm quite justified in saying that if it doesn't conform to the RFC it is by definition not a URL. That's the whole point of a standard.

Your phone number counter-argument is spurious, not least because it's simply incorrect. Or, I should say, very provincial.

If you think so, but forgive me if I don't agree with you given that you provided absolutely no evidence to support any of these assertions. As for the non-triviality of parsing URLs all anyone need do is look around the web for URL parsers to see how many are rather more complex than one would think and generally have plenty of caveats and edge cases they fail to cover.

bastion, on 08 June 2012 - 12:53 PM, said:

There's your problem: You don't patent concepts. You patent mechanisms (physical or otherwise). The concept "scan a stream of data looking for substreams that appear to represent certain kinds of information and present them to the user with options to operate on them in useful ways" is, relatively, obvious. But that's not what the patent is on. The patent is on a specific technique for achieving that goal.

That's a semantic complaint. I know you like to assume anyone who disagrees with you doesn't know what they are talking about, but I know full well what a patent is. I stand by what I've said, feel free to offer any substantive rebuttals you wish, but I don't feel this complaint can be accorded the label substantive.

I will just caution you that in the end my complaint is meant in a wider sense than just what is allowed under the current patent system. I believe the patent system is serious broken in certain areas and that many patents which are perfectly valid under law should not be granted (i.e. I believe the system should be reformed).

bastion, on 08 June 2012 - 12:53 PM, said:

So you can cite prior art then? And not just something that performs the same task but does so in the same way? Excellent. I look forward to seeing it.

According to this you seem to believe than simply because a solution hasn't been done exactly in the same way means it reaches the threshold for patentability. That's not the current legal threshold in the US and is certainly not a threshold I would support.

What I can say is that I was doing a similar thing well before the iPhone was launched. It used almost exactly the same structure for parsing as this patent outlines. The primary difference is that my system was concerned with extracting the identified structures, not providing any actions based on the parsing. The providing of actions to structures clearly has precedence in the auto-linking of URLs.

I have no real desire to continue this rather unproductive back and forth. All I wanted to say was that I don't feel these patents are productive. I feel such patents seriously hamper innovation (both in the short and long term).

This post has been edited by Stewsburntmonkey: 08 June 2012 - 01:30 PM

[bastion]

Stewsburntmonkey, on 08 June 2012 - 01:28 PM, said:

If you think so, but forgive me if I don't agree with you given that you provided absolutely no evidence to support any of these assertions.

I provided the RFC that defines what it means for something to be a URL. It's trivial, as a consequence of being a rigorously defined and well-documented structure. Sorry; I spent a noticeable portion of my life dealing with ISO 8583 message streams. Those aren't trivial, but still not too bad. My wife, for what it's worth, is the lead developer for the web site of a company with a substantial international market. Dealing with phone numbers is one of the worst tasks she's had to deal with in that capacity, and that's just taking a string that she's been told is a phone number and verifying that structurally it has any prayer of being legitimate for the locale whence it ostensibly came.

Quote

That's a semantic complaint. I know you like to assume anyone who disagrees with you doesn't know what they are talking about, but I know full well what a patent is.

Then, if I may offer some constructive criticism, you do a terrible job of expressing yourself. It would probably be a worthwhile exercise to re-read what you've written at least once before hitting the button to post, because this is not the first time I've had a debate with you that's hinged in part on a disconnect between what you've said and what it eventually turns out you meant.

Quote

According to this you seem to believe than simply because a solution hasn't been done exactly in the same way means it reaches the threshold for patentability. That's not the current legal threshold in the US and is certainly not a threshold I would support.

It's also not what I said. A piece of advice to help avoid future unproductive back and forths: If you start off a sentence saying that I "seem to" mean or believe something, you're almost guaranteed to be wrong.

Quote

What I can say is that I was doing a similar thing well before the iPhone was launched.

And what *I* can say is: "So was Apple." Did you happen to look at the date on that patent that you think failed the novelty test? The mechanism it describes made its first appearance as a shipping product in January of 1997.

And that'll wrap up all I have to say as well.

[TriumphNut]

bastion, on 08 June 2012 - 06:36 AM, said:

pcharles, on 08 June 2012 - 05:27 AM, said:

Does Posner get the 2012 sanity award for this decision?

Only if he's actually right, and was correct in introducing the previous limitations. This bears a certain resemblance to Apple's case against Microsoft in the late 1980s. Microsoft had compensated Apple for a license to copy certain elements of the Mac user experience in Windows, and they ended up going to court over a dispute regarding the termination conditions of the license - specifically to how many releases of Windows it applied. In that case, the judge pretty much ignored the suit in front of him *and* the gestalt user experience under discussion. Instead he chose to step through atomic components of the interface and declare them ineligible for protection one by one, after which he simply ruled the license to be invalid in that there was nothing to license.

This is less obviously egregious but certainly could stand some scrutiny.

This like the results of the monopoly suit against MicroSuck, just confirmed to me that Judges and Congressmen can be purchased.
As the late Texas Judge Hal Latimore once said to me, "When you go to court, you can expect anything except justice.