Have you ever heard of type inference? Have you ever taken a course in formal logic or semantics?
One of us has written a non-trivial compiler, methinks.
Yes. Yes. And both of us apparently. So I will give you this.... type inference is one area that compilers does in fact attempt to infer something from the syntax and context. But the rules are still explicit as to when you have to annotate with casts and such to disambiguate. This is certainly one of the problems in C++ where you have multiple cast operators on a class and you call a function that does not take the type of the object you pass but takes both of the other types that you can cast to. The inference of the polymorphism breaks down and you get an error at compile time regarding the ambiguity.
I will retract from "compilers do not infer anything", but I still believe that everything you are doing is intentional and is fundamentally explicit from the developers point of view. There are strict rules about what the compiler can and cannot infer based on the syntax and context you give it.
The reason I was so adamant about it is because when somebody emails free-form text without markup (e.g.: HTML) they are not explicitly designating which parts of the texts are links to websites and which part are phone numbers or dates or tracking numbers. There is not even a grammar that defines a structure for that email. It is literally free-form text. However, parts of that free-form text may have structures that match specific regular expressions and are therefore identifiable with some degree of confidence. So somebody with no knowledge or expertise in markup language can construct a message that will automatically get marked up. TO me that is true "inference". When I code something I fully expect the executable to do whatever I told it to according to the rules of the programming language I am using -- no more, no less.
Bear in mind that I do NOT think that Apple has a patent on inferring links from emails -- since their patent, as it is written, would exclude server-side CGI-type applications from infringement for simply rewriting something that looks like a link as a link. The reason for exclusion is because that server-side code is not providing the handler for that link which is one of the elements of their claim nor is it making the input path from the user available -- it is simply processing text. I don't even think Apple could go after Google because they have server-side code + Google Chrome browser since the server-side code will talk to any browser via HTML and Chrome can be used to point at non-Google websites. This would, however, be applicable in
desktop mail clients that picked up on email addresses and URLs in the body of an email message though and provided handlers for them.
That part of the argument was only to establish that inferring something from a piece of text is not in any way novel.
I fully agree with you -- inferring something is in now way novel in and of itself -- it is only one aspect that is a
requirement of the infringement, but it is not
sufficient for infringement. The thing about these claims is that you can do any one part of them or any combination of parts of them you like so long as you are not doing all of them in the specified way (or something that does them and more). For example, I could take Apple's claim and add something non-obvious to it and go patent that. That would just mean that I would have to license Apple's patent to implement my own, and that Apple in turn would have to license my patent if they wanted to implement my enhancement. Anybody else who wanted to implement my hypothetical patent would have to license from both parties.
If they did, that might make the patent truly innovative. It's not really about prior art here, but the below is. What Google does is much more sophisticated. It is the basis of Google Translate, BTW.
I'm darn sure that Google Translate is far more sophisticated and far more specialized than what Apple is doing with this patent. But the "value" in this patent is its wide applicability. Sometimes the most valuable patents are fairly straightforward -- especially once you see them (they just make sense).
Full disclosure: I haven't actually used Lotus Agenda. I've only read about it and the product that was directly based upon it (Chandler).
Full disclosure: neither have I. I just read about it. But it seems very different than the goal and implementation of Apple's patent.
I'm confident they'll be able to get these patents thrown out on their own.
Maybe. Maybe not. Certainly Motorola feels strongly about it since they have filed lawsuits to invalidate these patents. I am just not inclined to so readily dismiss patents as obvious with 15 years of experience and hindsight working in my favor. Motorola/HTC needs some prior art that is the same or very closely analogous to Apple's claims to defeat it. Screaming "it's obvious!" fifteen years after the fact is not going to do it since it already past the non-obvious test with a patent examiner over a decade ago.
I fully believe that Android did not set out to "steal" Apple's IP. I'm sure some developer saw this feature on a plethora of apps and mobile phones for years from Apple and others (after all Apple has either licensed this one or has not enforced it). Then they implemented this same thing and just happened to step on the mine that was buried over a decade ago.
There is no better analogy for patents than the "mine field". Personally, I think it sucks. The stakes in this whole patent race just keep going up and it reminds me more and more of the nuclear arms race -- with patent trolls representing the radical terrorist regimes that have nothing to lose by launching a nuke. Some folks are calling for total disarmament -- maybe that is the answer, but then I just see the little startups getting squashed by the big boys who just steal all their innovations.
