Re: computer-go: Engineering (was: Most simple Go rules)

[Robert Jasiek <jasiek@xxxxxxxx>]

Mark Boon wrote:
> The Chinese rules, [...], the NZ rules [...], 
> all require to play on longer and count territory and stones, which I
> personally don't like because I find it more laborious. That's why I like
> Japanese rules over any other rules that I'm currently aware of.

To understand what either of us is saying, we need to be more
precise what we are talking about:

Most rules with Chinese scoring usually proceed as follows:
1) strategically required alternating moves
2) two successive passes
3) agreement about removed stones
4) not changing the situation any more except possibly for 
   counting

Most rules with Japanese scoring usually proceed as follows:
1) strategically required alternating moves
2) two successive passes
3) filling dame, etc.
4) removal of stones
5) not changing the situation any more except possibly for 
   counting

***

So if you find rules with Chinese scoring more laborious or
longer than rules with Japanese scoring, then the only
difference you could be referring to is whether "filling of
dame, etc." takes place before or after "two successive
passes". The effort of performing "filling dame, etc." is
exactly the same because all dame, etc. are filled. Calling
the Chinese way more laborious refers to doing it alternately
instead of doing it informally. Calling the Chinese way 
longer refers to counting only those stones put on the board
that are put there before "two successive passes".

I understand that this is your preference for human play.
However, do you suggest that programs should also follow
the Japanese practice of performing "filling dame, etc."
INFORMALLY? If so, why have the programs not done this
themselves so far? CLearly because it is not a most simple
procedure for programs since they would have to resolve
conflicts of both programs occupying a same intersection at
the same time.

> those [Japanese] rules allow for quick ending of the game

Inhowfar is an ending by rules with Japanese scoring quicker 
than an ending by rules with Chinese scoring? If you look at
the procedures above, then you notice that not the procedures
as a whole have different performance times but only the moment
of stopping the count of alternately played stones is different.

> and fairly straightforward counting, something
> I'm missing in other rule-sets.

Inhowfar do you consider Japanese counting to be fairly
straightforward in contrast to counting methods used in
other rulesets?

> When very precise rules start to get in the way of normal common
> sense we call it bureaucracy.

Calling it "bureaucracy" is politically motivated; it might as
well be called "clarity":)

> And I try to avoid senseless bureaucracy
> whenever possible.

How many lines of code do you guess you need for implementing
Chinese scoring or for Japanese scoring, respectively? If 
the former were senseless, then the number of code lines would
be roughly the same. It won't be.

> Moreover this kind of bureaucracy will definitely not win
> over any new beginners for the game.

If you tell a beginner "Look, if you want to be BUREAUCRATIC,
[...]", then there is no doubt about the effect of such a 
statement to beginners:(

> If we can have the best of both worlds then I'm all up for it.

Here we will never agree, it seems:)

> And so far I have not seen anyone propose anything I consider
> a workable alternative.

What would you call workable?

> We're
> trying to make a computer play a game that people play, not a game that's
> played by computers only.
> The fact that some of the rules are harder for computers than others is
> irrelevant. We shouldn't be fiddling with the rules just because they're
> difficult to program.

The major purpose of rules is not the degree of their simplicity
but the degree of fairness of their application. It is the
logical rules (which by coincidence also rather often happen to 
be simple) that allow measuring their degree of fairness so that
this can be understood by the players, organizers, referees, 
beginners, the media, and the general public.

--
robert jasiek