Re: [computer-go] Data structures for Go

[Heikki Levanto <heikki@xxxxxx>]

On Wed, Oct 08, 2003 at 01:20:12PM -0500, Richard Brown wrote:
> Here's a vote for linked lists of points.  [...]

Here is a different vote. In Gnu Go we keep things in arrays. Used to be
2-D arrays to reflect the board, but we optimized that into 1-D arrays.
Mostly to save time in passing coordinates as parameters. 

We do not do much of undo in our reading, rather we push the whole
structure into a stack, and instead of undoing, we merely pop it back.

This works quite nicely and effectively. 


> When the Ing prize was around [It's expired now, right?] they said that
> no part of your program could be written by anyone else.  So, even if I
> would "find useful" some code that you produce, I wouldn't have been
> allowed to use it in my program, in competition, would I have been?

Of course we all build on the shulders of the giants who have developed
the basic algorithms and data structures for us. Often coding them all
over again, as often borrowing from a (more or less) standard library.

> But maybe times have changed.
> Is it now all right, in computer go, to "find useful" the work of others?

Gnu Go was never specially aiming at the Ing prize. It has competed
(quite successfully) in some competitions. As you may know, it is
written by a team of volunteers, and is released under the GPL, so
anyone can look at the source and borrow things he "finds useful", as
long as he releases his code under the same GPL. If this would ever
prevent us from participating in competitions, we could happily do
without - our goal is write a good Go program.

Regards

  Heikki Levanto

-- 
Heikki Levanto  LSD - Levanto Software Development   <heikki@xxxxxxxxxxxxxxxxx>

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://computer-go.org/mailman/listinfo/computer-go