[BioSQL-l] towards schema freeze

[Hilmar Lapp]

I encourage people looking at the present schema definition to see 
whether there's anything that makes you feel uneasy or unsure.

  From my perspective, there are no real problem areas left that I would 
be aware of at this point.

The best test of course is always subjection to the real world. The 
current HEAD revision of bioperl-db should actually work with the 
current HEAD of the schema, if someone wanted to test-drive this. Also, 
there is Aaron's load_ncbi_taxonomy script, which already caught a 
minor bug.

To finish the clean up, there is one remote corner I don't find pretty: 
location_qualifier_value has columns value (a varchar) and int_value 
(an integer). The only advantage of having both is if someone wanted to 
efficiently (indexed) query against an integer value (like all 
locations that are fuzzy and have min_start = 10) in this table. 
Otherwise it is just a waste of space, plain ugly, and potentially 
confusing as to which column you're going to find the value in that 
you're interested in. Do people feel strong about keeping both columns? 
If not, I'd propose to remove int_value.

If there's anything you learned to dislike about the schema, *now* is 
the time to bring it up. I'd like apply the freeze by the end of this 
month at the latest.

	-hilmar
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------