|
Authored by: Ian Al on Tuesday, May 08 2012 @ 05:52 AM EDT |
The 520 patent was all about Oracle's rubbish compiler. I think 104 is all about
evaluating symbolic references at run-time. Oracle's definition of run-time is
any time from 1997 to this day. Google point out that it means during or just
leading up to the execution of the code in the Java Runtime Environment (or
Dalvik, in their case) on a Java phone. I think the judge still has a question
outstanding from the claims construction phase about what the term 'runtime'
means in the patent.
I think that '104 is what Mitchell was claiming
that dexopt does. Just in case you forgot, Mitchell cites the two Java programs
that call dexopt. Thing is, Dalvik does not run Java. Those programs have to be
run under the JDK on the developer's machine. On the Android website, Google
prohibit the use of dexopt in a production Android device.
Anyway, on a
production 'phone it would not be the dexopt that Mitchell accused because any
dexopt on an Android 'phone only runs on the Dalvik VM. Did I mention that
Dalvik is not Java?
Google incite the 'method and apparatus' patents to
be used, if at all, by the developers in the Sun Java Development Kit. The
apparatus is the developer's computer once the JDK, dx and dexopt have been
installed on the computer (Microsoft v. AT&T). None of the accused phones
has the cited apparatus invention installed.
Oracle have to show that
the developer's infringing apparatus is in the US or was imported into the US.
Have any developers' computers been accused in this case?
Similarly,
the method is only employed on the developers' computers. Have any developers in
the US jurisdiction been cited in the case?
And, of course, in their
opening statements Oracle have explained why Google do not infringe on either of
the patents. Not that I think Oracle understand that.
So, what about
all those Android registrations that Oracle complained about? Why is that
relevant to developers' computers? Well, looking at the Diehr patent for a
process to make cured, precision, rubber components, supposing someone used the
unlicensed process to make millions of rubber ducks. Would the infringement
entitle Diehr to a proportion of the profits on the rubber ducks? On the other
hand, Google's Android profits were not from selling Android.
So, if
Google had infringed, would Oracle have been entitled to profits from things
done with things made using the inventions? And, what is the proven link between
the inventions, the number of Android phones registered and the revenue Google
make on advertising via internet browsers on all mobile phones and other mobile
devices?
I don't think the patent case is anywhere near as strong as
the copyright case. Perhaps Google ought to concede infringement of at least one
patent to get a level playing field. No, it doesn't help, does it?
Of
course, the final clincher is that Oracle say this:
Compare with
Android Platform Components: Java App source code gets compiled by a java
compiler. "Go to the Oracle website and download a Java compiler"
dx
tool is bolted onto to bottom of the Java compiler. The dx tool converts Java
bytecode to "dexcode." Load onto Dalvik VM. Dexopt (optimization), and bytecode
interpreter.
They are absolutely correct: that is what Google
require on the Android website as the prerequisite for development.
104 patent, 520 patent, developed by Java developers back in the
1990's. Solved the same problems, speed and memory utilization.
As
Oracle say, the two inventions are essential for the Java platform to run at all
well on mobile devices. The Java versions must be included in the JDK. That's
why the marking is such an issue because infringement can only be from 2010 when
Oracle threatened to sue.
Before the JDK can be downloaded, developers
have to get an Oracle licence for the use of the JDK with any of the software
they develop.
If the two inventions are Java inventions and solve a
major drawback with the Java platform, they must be licenced along with the rest
of the JDK for use on developers' software. That would include Android
developers getting a JDK licence and downloading the JDK just as Oracle explain
they must do. Oracle don't prohibit the use of third party tools. They are an
essential part of the ecosystem. Over the last few days, many developers have
highlighted just how important they are to the ecosystem.
So, the
developers have a licence to use the two patents with Android. Not that they
need them. Dalvik is not Java and Google don't need the two
inventions.
--- Regards
Ian Al
Software Patents: It's the disclosed functions in the patent, stupid! [ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Tuesday, May 08 2012 @ 09:29 AM EDT |
Microsoft's C++ compiler, for at least 4 years (and probably more like 7)
has been able to simulate execution of constructors for global variables
and if there are no side effects other than initializing the fields, it can
elide
them and emit an already-initialized object into the data segment.
I wonder if this is prior art? Or infringing?
...the simple fact is this is an OBVIOUS IDEA to anyone skilled in the art
of compiler construction. I had the idea myself before I ever noticed that
Microsoft's compiler can do it.
The U.S. patent system is hopelessly screwed up if it allows companies
to patent such an OBVIOUS IDEA and then sue each other over it.[ Reply to This | Parent | # ]
|
|
|
|
|