Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The Book Problem, Open Book Format proposal, Request for comments.

Author: Dann Corbit

Date: 10:08:01 03/04/06

Go up one level in this thread


On March 04, 2006 at 10:22:47, James Swafford wrote:

>On March 03, 2006 at 18:20:24, Dann Corbit wrote:
>
>>On March 03, 2006 at 17:10:03, Mark Boylan wrote:
>>
>>>On March 03, 2006 at 17:06:01, Steve Maughan wrote:
>>>
>>>>Dann,
>>>>
>>>>I don't like the idea of a SQL based system.  I'm not an expert in SQL (which is
>>>>one reason I don't like the idea of SQL :) ) and I also think it would mean that
>>>>an engine that wants to support the format would be *much* more complex than one
>>>>that supported a well defined record based structure.  I've also concerns that
>>>>SQL may be too slow for very fast blitz games - but as I said I'm not a SQL
>>>>expert so my fears may have no basis.
>>>>
>>>>Regards,
>>>>
>>>>Steve
>>>
>>>
>>>I have to agree with Steve. I think that SQL is a great idea for book-building
>>>tools and such. But if I had to add the psql headers and link to those libs, I'd
>>>probably just say "forget it." I would think there needs to be an export format
>>>for engines.
>>
>>Just use an API.
>>
>>It will be far faster and better than a custom formet.
>>Also, all the bugs would be removed.
>>It would be more flexible and basically better in every conceivable way.
>>
>>Not using SQL is definitely the wrong approach.
>
>
>I've come to the same conclusion.  Let database engines do what they
>do best; no need to reinvent the wheel.
>
>I'm starting a rewrite of Prophet, and I'm going to use a Postgres backend
>to log games played, eval scores, book moves, etc etc etc.
>I use Postgres at work and *love* it.  Postgres is free, and now there are
>even Win32 builds available.

The Win32 port has some code that I wrote in it.  We did the first port to Win32
here with PostgreSQL 7.1 and we donated some of the code to the PG project
(actually, it was probably simultaneous with the Japanese effort which was a
threading model, rather than a process model).

>The downside to this approach is that installing a database to use my
>engine is going to turn most people off, but I'm not writing my engine
>for most people, so I can accept that.

The SQLite database uses a subset of PostgreSQL syntax (including things like
VACUUM).  It is an embedded database that requires no installation.  You can
even do an ASCII pg_dump and then load your SQLite database with that file.
There may be some small issues due to schema support in the recent PostgreSQL
builds, but I guess that a couple simple edits would fix that.

>Have you put much thought into what an opening book schema might look
>like?

I see that you found it.  BTW, the book stuff is absolutely public domain.  Do
absolutely anything you like with it.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.