Re: [Toss-devel] Imperative Rule Representation

[Lukasz K. <luk...@gm...>]

> I'm sorry for not checking the paper right now, but I think that rules
> based on adding/deleting are already fully implemented (including a
> nice syntax), with *a lot* of flexibility (for example, with compiling
> defined relations into structure relations). Providing rules by
> structures is just an option.

I forgot that we already even had the syntax - that's really nice!
Now I think that we should move most our standard examples to
the imperative rule syntax, what do you think? Except for rules
where explicit grid-like structure syntax is used, this seems to be
much more readable, right? I just commited a change and moved
Tic-Tac-Toe, but I'll wait with the other games to know what you think.

Another small change I made is to replace MATCH with pre. I am
not 100% sure, but it seems to me that there is no real difference if
you write precondition inside MATCH or later after pre, right?
If so, then I prefer to have it all together in the front - exactly as
it is done in YAGI in the paper I cited in the previous mail (maybe
you don't need to read the whole paper, but just the first column of
the "YAGI by example" section on page 3). Do you agree with this
change? I hope that this will really make the rules more readable!

Lukasz