question about patent

Eric Springer erikina at gmail.com
Fri Mar 27 16:54:10 UTC 2009


On Wed, Mar 25, 2009 at 12:40 AM, robert song
<robertsong.linux at gmail.com> wrote:
> Hello, everyone.
> Now I am using Pettis-Hansen method as follows to reorder functions.
> http://www.cs.virginia.edu/kim/courses/cs771/papers/pettis90profile.pdf
>
> But I found that the algorithm has its patent as below.
> http://www.freepatentsonline.com/EP0459192.html
>
> So if I use this algorithm in my codes, it will infringe the HP patent.
> Does it mean that this algorithm can not be used in codes ?
> Is there any method to deal with this problem ?
>
> Best wishes,
> robert
>

I agree with Ric on this one. You're far better off to never look at
any patents and not discuss them either. For a variety of reasons:

a) Knowingly infringing has x3 the penalty.
b) You're drawing attention where none should be.
c) Almost all of these patents are untenable


And such is the unfair nature of patents. I bet some of my code
accidentally "infringes" a dozen patents. Completely ignore them
completely unless you're doing something like a clone of some
commercial software, where you'll need to tread carefully .


P.S. I'm not a patent lawyer, but I read through the patent. And
people have been doing optimising like this for well over 30 years in
assembly. And considering the patent was filed for a compiler method
of doing it, in 1991 -- I have no doubt it was in every major compiler
by then. Worthless patent as far as I can tell.

> --
> fedora-devel-list mailing list
> fedora-devel-list at redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>



-- 
Eric Springer,
PGP Fingerprint: 097D E98D 9278 FE86 2659 2959 DA9E 90BD F183 2F88




More information about the fedora-devel-list mailing list