[plt-scheme] PLT and Database access

> (David Neu, can I write to you offlist with some questions, or would
> you prefer them here?)

Either is fine with me, if you think that your questions would be of
general interest, let's try it on the list.

On Fri, May 27, 2005 at 05:35:42PM -0700, nishad at ptolemy.tlg.uci.edu wrote:
>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme
> 
> 
> 
> Thanks to everyone for your very helpful replies.  I'm playing with
> scheme-pg, and it's looking like the answer (David Neu, can I write to
> you offlist with some questions, or would you prefer them here?)
> 
> 
> Jacob Matthews wrote:
> > 
> > 
> > I said before and I'll say again that once the web site project is done, 
> > I'm going to try to write up an entry for the Scheme Cookbook 
> > summarizing the current state of affairs in the world of connecting PLT 
> > Scheme to a database. It's a topic that I think is in dire need of 
> > better documentation.
> > 
> 
> My troubles so far have mostly to do with the ODBC factor.  I have,
> from the bottom up, a Postgres release, a unix ODBC driver manager
> (two choices), a Postgres-specific driver (at least three choices),
> the v209/299 differences (albeit temporary), SrPersist (tricky to
> compile), and SchemeQL (buggy, possibly moribund), all of which have
> to come together with compatible versions.  The psqlodbc releases
> alone sometimes compile with one driver manager, sometimes with the
> other.  Based on the fun I had trying to get this to work on Solaris,
> I can't say it's the way to go unless 1) the cross-platform promise of
> ODBC is vital to your work, and 2) hacking a package like
> SrPersist/SchemeQL is as easy to you as using it (the line between
> these two sorts of programmers seems to be much finer in Scheme than
> in other languages, which is in some ways a dangerous thing).
> 
> <Wishful thinking> 
> I'm not sure what a universe solution would be.  A standard (SRFI?)
> interface of database bindings would need an implementation layer per
> DB program per OS.  This could be done either scheme-pg-style (use the
> DB's libraries) or spgsql-style (do it entirely in Scheme).  That
> still seems like less work, and potentially more stable, than the ODBC
> route.  DB libraries and protocols are much more stable than driver and
> driver manager releases, AFAICT.
> 
> At any rate, one longs for the day when all this is as easy and
> brilliant as the parser tools, XML, and web programming collections
> that PLT provides.
> </Wishful>
> 
> Cheers,
> 
> nishad
> -- 
> "Things ain't been the same since the blues walked into town."
> 
>                                                 -- Larry Love