Subject: Re: US Software Patents Hit Record High (fwd)
From: Thomas Lord <lord@emf.net>
Date: Wed, 27 Sep 2006 11:47:55 -0700

Russ Nelson wrote:
> simo writes:
>  > b) get out of it any useful teaching that is not yet widely available or
>  > trivial to come with up given the problem space
>
> Indeed.  Unfortunately, patents are being granted to people who ask an
> unobvious problem and then solve it using the obvious solution.
>
>   

I'm not opposed to creating incentives for people who anticipate unobvious
questions and then, when the question becomes shared, just "happen to have"
the right sized wrench.

We should take bets, in the case at hand, about what the actual patent says.
I'm wondering if it tries to lay claim to a variation of bit-matrix 
transposition
by-recursive bitblit.    I've never done any SSE programming but, what with
the optimized loads and stores and the various shuffling instructions, 
I'll bet
that would be pretty fast.   You'd have to special case the deepest 
levels of
the recursion, of course.

-t