Re: [Bug-gnubg] Database interface

[Joern Thyssen]

On Tue, May 25, 2004 at 03:26:44PM +0100, Jon Kinsey wrote
> 
> Or we could just remove it, is it helpful - after all I guess everything 
> in the database is under the same schema?

I can't remember why I made the schema... 

> You're right, in that I've hardly changed database.py to use a different 
> database.  I'm not sure how many users are likely to do this though!
> 
> Perhaps this gets to the point I was trying to make - how do we see this 
> being used?  I doubt many users are database experts, this is what I 
> think is more likely:
> 
> Ordinary user:
> * Minimal initial configuration of a database
> * Use gnubg to store matches in database and retrieve basic statistics
> * Possibly use python to run (supplied) scripts to gather other data 
> from database
> 
> Power user:
> * Use python to write scripts to gather more specialized data from database
> * Possibly query database directly
> 
> For windows users (who invariable won't have a database on their 
> machine), I would imagine supplying sqlite and not even telling them 
> that it's there would be the best policy.
> They then wouldn't necessarily have to install python.

I'm not really fond of using the C APIs, neither sqlite nor any other.
You could also argue  that most windows users have Access installed, so
we should use that instead. 

I think we should document the database-specific part of database.py
really careful, and write about it in the manual, e.g., 

Using gnubg and SQLite:

Create the database by executing blah blah yada yada. Locate and modify
database.py as follows: yada yada yada.

BTW, I don't imagine any users to use python to query the database. I
would expect them to use the query tools available for their specific
database. We can supply SQL queries to be pasted into MS Access,
postgresql etc, but I don't think we need to write any python scripts to
do that.

Jørn