[BioSQL-l] Re: Postgres version support

[Chris Mungall]

On Mon, 21 Oct 2002, Hilmar Lapp wrote:

>
>
> > -----Original Message-----
> > From: Chris Mungall [mailto:cjm@fruitfly.org]
> > Sent: Monday, October 21, 2002 5:16 PM
> > To: Hilmar Lapp
> > Cc: Elia Stupka; Biosql; Thomas Down
> > Subject: RE: [BioSQL-l] Re: Postgres version support
> >
> <snip/>
> >
> > one chaneg i made in bioperl-db was treating db_xrefs in the
> > feature table
> > as both dbxrefs and seqfeature_qualifier_values
>
> Yeah I've seen this in the logs. I was going to introduce this on the main trunk as well.
>
> >
> > i haven't tested the mysql->pg converters on the latest mysql schema
> > Hilmar has added to the repository. ideally i'd like to see
> > us move to a
> > schema neutral format and add all those Type=InnoDB to the converter.
>
> That'd certainly be preferable. Then one could also have a more
> powerful description language for the schema.
>
> What I did is modify the converter's grammar to recognize the Type=XXX
> syntax (and ignore it in Pg output). BTW asking the converter to
> produce Mysql format was (and still is) broken, not sure who ever used
> that.
>
> The MAGEstk folks in fact came up with a XML-schema definition for
> describing a schema, and a corresponding converter to DDL of
> particular flavors (which I believe so far is Pg only, not sure).
> Jason, correct me if this description is too sloppy.

personally I'm in favour of keeping the core schema in SQL (without all
the horrible mysqlisms), but we could easily convert to an XML
representation. In what sense would XML-schema definition allow for a more
powerful description language?

I'm interested in the schema definition schema though - I'm rewriting the
SQL DDL parser along with Ken Clark @ cshl and would like to use this as
another format

> 	-hilmar
>