|
Authored by: Anonymous on Friday, April 27 2012 @ 01:04 PM EDT |
I'd completely misread their backing off of "Java is not a
collective work".
From a logical perspective then, if they claim these 37 APIs
are the whole of the work, then they can't be in the Java
copyright - or it would be a collective copyright (by
containing two separate copyrighted items - these 37, and
everything else). If these 37 are a work as a whole, they
must be disjoint from the rest of Java, which, not being
collective, must be a single copyright... etc.
If Oracle gets the work as a whole their way, Google just
needs to bring in a mathematician to slaughter them with set
theory.[ Reply to This | Parent | # ]
|
|
Authored by: hardmath on Saturday, April 28 2012 @ 06:15 PM EDT |
I think Judge Alsup contemplated using differing scopes of
"the work as a whole" for different questions in the jury's
verdict form.
Oracle backed off the business about a compilation work, and
that led to some simplification as far as using the same
scope for both code and documentation. However there is
only minimal evidence for copying of the documentation, and
the 9 lines of rangeCheck code are arguably minimal copying
of code.
Bearing in mind Judge Alsup will instruct the jury in some
form that the APIs enjoy copyright protection, the main
thrust is what scope will the "work as a whole" have for a
question about the accused APIs, considered perhaps as the
"stub" component of the implementing code. There was some
testimony about what fraction of the library code the
declarations form, and if one limits scope the 37 accused
APIs, that works out to 3% or so.
However what I thought Judge Alsup was proposing was to use
all the Core APIs as the relevant scope for "the work as a
whole". This omits the JVM/Dalvik contrast, which witnesses
widely admitted would dilute the fraction of copying if
included in scope, but does not give Oracle the narrow "just
the 37 accused APIs" that they wanted.
Because the accused APIs are for the most basic and probably
shorter code libraries, I'd expect the inclusion of all the
Core APIs (the figure 166 was given for Java SE 1.5) would
disproportionately reduce the numerical fraction. On the
other hand Oracle may argue the selection of 37 "copied"
APIs (which Google admits are substantially similar to what
is included in Java 1.5) signifies a higher importance as to
representativeness of the whole.
---
"Prolog is an efficient programming language because it is a very stupid theorem
prover." -- Richard O'Keefe[ Reply to This | Parent | # ]
|
|
|
|
|