[Biocorba-l] Database versioning

[Alan Robinson]

On Tue, 6 Feb 2001, Ewan Birney wrote:

> On Tue, 6 Feb 2001, Alan Robinson wrote:
> 
> > 
> > On Tue, 6 Feb 2001, Ewan Birney wrote:
> > 
> > > On Tue, 6 Feb 2001, Brad Chapman wrote:
> > > 
> > > > Alan:
> > > > > Btw, the version number of a database - should it be a 'string' or an
> > > > > 'integer'??? Currently, it's the latter - But I'm now not so sure if
> > > > > that's appropriate...
> > > 
> > > I am a big believe that version numbers should be integers such that you
> > > know that if you have version 1 and version 2 you should look at version
> > > 2.
> > 
> > I agree absolutely - Unfortunately, the reality doesn't appear to be like
> > that. The GenBank and EMBL databases are well behaved in this respect.  
> > But, for example, what versioning would you use on Ensembl? I know the
> > current setup for the software/data is described as "Ensembl 0.7.5 with
> > the September 5th dataset". Is there a database version stored somewhere
> > else?
> 
> each *database entry* has a well behaved, integer version. 
> 
> shame on you to think otherwise.
> 

Hi Ewan,

First a sanity check - Are we talking about the same thing??? 

Certainly, to have anything but an integer for a database entry is asking
for trouble. But we're talking about the version "number" of the database
itself, e.g. "GenBank 121.0" [ftp://ncbi.nlm.nih.gov/genbank/gbrel.txt] or
"EMBL Release 65 (December 2000)" (Does GenBank ever have release versions
such as 121.1?)


Alan.


PS There are probably more pressing issues than this in the IDL -
Personally, I'm chewing over the Annotation / Feature / Location /
Fuzziness issues again.