|
Authored by: Anonymous on Thursday, May 10 2012 @ 10:43 PM EDT |
In the end though, the timing test is meaningless in terms of showing
infringement. Let's go back a few decades and assume that software patents were
recognized. Everybody was using a bubblesort to sort data, but then John Von
Neumann came up with mergesort and patented it. A few decades later Tony Hoare
invents quicksort and uses it in an application. John Von Neumann's employer
sues the publishers of the new application on the grounds that the speedy sorts
displayed by that application must use mergesort because the sorting is
happening too fast to be getting done by bubblesort. Unless the mergesort patent
was written broadly enough that it could be stretched to apply to quicksort, the
mergesort patent shouldn't apply to the quicksort-using app. The duration of the
sort is completely irrelevant.
Oracle's performance numbers show that Android/Dalvik doesn't use the slow naive
approach against which they are comparing their patented operation. That's a
long long way from demonstrating that their patented approach is what is used by
Dalvik. The real question is, does Dalvik use the patented technique/algorithm?
That should be straightforward to determine by an examination of the Dalvik
virtual machine's implementation. The whole timing issue is at best irrelevant
and a waste of the court's time, and at worst a deliberate attempt at confusing
the jury.[ Reply to This | Parent | # ]
|
|
|
|
|