Re: Shared game-data
Date: Sun, 30 Apr 2006 22:18:28 GMT
Message-ID: <UIa5g.1030$A26.29343_at_ursa-nb00s0.nbnet.nb.ca>
Dmitry A. Kazakov wrote:
> On 30 Apr 2006 11:55:31 -0700, Neo wrote:
>
>>What is also evident is that Bob, Marshall, Alfredo and Jay Dee are >>clue-less about the extent for performance and flexibility requirements >>of some gamming applications.
>
> Performance is an inadequate word here. It is like to say that differential
> equations don't have good acceleration and petrol consumption. RM isn't a
> solution, it is a description language. Translation from this language to
> the machine one is as non-trivial as solving differential equations.
Given the lack of theoretical work on physical independence and the abundance of work on solving differential equations, I suggest that at this point in time solving differential equations is more trivial than translating predicate calculus to machine code.
> Talking about AI, RM is perfectly suitable to describe supervised learning
> problems. But it gives absolutely no clue how to solve them.
One could say the same for any other implementation language.
It could be
> disputed if gaming applications, which rely not only on AI, but also on
> simulation, HMI, computational geometry, translation, real-time, concurrent
> and distributed computing could be described (not yet implemented!) in RM.
> As for implementation, well, well...
Proper support for domains and adequate physical independence are all that's needed. Received on Mon May 01 2006 - 00:18:28 CEST