Here is my opinion:
The instance where a table will have a FK to two tables with the same
name but different schemas is going to be rare. I think analogous to
the views having two foreign keys to different subclass views of the
same table. Probably even rarer.
Make the common case fast and easy. Keep the "getGene()" syntax. When
there is such a clash, then use the schema name "getDotsGene()" &
"getFooGene()"
The ability to have a modular schema is more important in my opinion.
cheers,
Angel
On Feb 17, 2004, at 6:58 PM, Steve Fischer wrote:
> folks-
>
> i think that we should consider a rule about GUS table names:
>
> "table names are unique across schema spaces"
>
> there are pros and cons here.
>
> pros:
> - this way, in our object layer, we can have simpler accessor names,
> such as getGene(). Otherwise, we need to have getDoTSGene() (yuck)
> - this may help us port, if we need to, to systems which don't
> support schema spaces
>
> cons:
> - the obvious one: schema spaces are no longer truly modular. the
> must be aware of other schema spaces that are or *might be* used in
> the same system.
>
> Comments???
>
> Which do people see as the lesser of the two evils?
>
> steve
>
>
>
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
> _______________________________________________
> Gusdev-gusdev mailing list
> Gus...@li...
> https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev
|
