Don Dailey wrote:
Can you dig out the textbook where you got this list from
Writing it down as if it were a formal definition was a joke, of course.
I have made up the list myself but it has strong reasons: It is the
essence of my study of theoretical informatics at university. Whichever
good proof might have been made, it did apply all my points. Exception:
it was a summary style explanation in some maths journal where the
readers (mathematicians themselves) are expected to think 30 minutes per
text line to verify and understand completely what they are just reading.
> and be more precise about what they are trying to define?
>> a) the underlying theory is published,
Privately stored text won't do. You need to publish it to everybody.
Explain the underlying theory, unless it is standard knowledge for the
typical readers. E.g., "alpha-beta" may be well known but if you make
inventions that change your specific alpha-beta search, then explain
your new theory.
>> b) the underlying theory is proven mathematically,
You are going to see endless lists of propositions and their formal
mathematical proofs.
>> c) the algorithm is published,
Not the source code of your implementation is interesting here but the
algorithm independent of programming language, software, and hardware.
>> d) the algorithm is proven mathematically,
Prove that it well-defined, complete, does what it is said to do, always
does so, never does anything else, prove what the computational
complexities of time and space are.
>> e) the used computer environment is stated,
Mainboard, processor, operating system, etc.
>> f) there is a statement that the computation has been done
successfully and
Trivial, but you should add the running time taken.
>> g) it is possible to repeat the computation independently.
Anybody with a different computer / different set of computers must be
able to take the algorithm, implement it on a different computer, and
invariably produce the same result.
For instance if I solve a Rubiks cube in private, is it sudenly not solved
Nobody won't believe you until you publish your used algorithm, etc.
> it just means that I cannot credibly claim to have solved it.
Indeed. Like claiming "I have solved 19x19 Go yesterday.".
> It's extremely common in
games research to have papers like this, where results are reported but
completely unverifyable and it drives me nuts.
Right.
--
robert jasiek
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/