Scott Dorsey wrote:
>> Today, we have the additional issue that the patent office does not
>> have enough inspectors with actual familiarity with software
>> technology or with algorithms. This is how Microsoft can get away
>> with patenting the ring buffer, a data structure used at least as
>> early as the CDC 6000.
> Why didn't CDC patent it?
Because software patents didn't exist at the time. Had they existed, I am sure CDC would have.
> We currently have a situation where huge numbers of obviously
>> invalid patents are being issued, and there is no way for the
>> patents to be declared so without going to court. And once it comes
>> time to go to court, sadly it tends to be a situation of the person
>> with the most money winning.
> I won't deny money helps a lot, but let's be clear Vonage is not a
> single guy working out of his garage; it's a large ongoing concern
> with plenty of resources. And let's not forget the huge AT&T lost in
> court to pipsqueak MCI over cream skimming.
In this particular decision, the patent exists. It now becomes AT&T's job to prove that prior art exists and that therefore the patent should be invalidated. To do this, they need to find prior art.
Finding previous attempts to run realtime data over packet-switched networks is going to be interesting, and it'll be fun to see how their legal staff does it.
At the same time, supposedly patented IBM technology was easily
> circumvented. IBM invested a heck of a lot of money into development
> high speed disk packs and disk drives for its System/360 and other
> companies, helped by ex-IBM employees, copied them and sold them,
> undercutting IBM. That wa easy to do since they didn't have to cover
> the R&D. RCA's Spectra was a clone of IBM's System/360. How the heck
> did they get away with that? They obviously did.
Lots of folks cloned the System/360, from Amdahl to National Advanced Systems. Most of them did much better jobs than RCA did. (RCA used the S/360 instruction set but very different channel controller and peripheral design, including the use of ASCII). They could do that, because you couldn't patent or copyright an instruction set.
Now, I think the idea of a finite automaton COULD be patentable, but Turing never filled out the paperwork...
--scott
"C'est un Nagra. C'est suisse, et tres, tres precis."