Quoting Anita Richards <amsr-at-jb.man.ac.uk>:
> VOTable seems relatively
> easy to evolve;
I agree with everything you say Anita except this bit! VOTable is not the right thing to *start* with when looking at the next stage. Let's use XML as it was meant to be used, not take XML, apply a schema, then try and squeeze everything into the resulting IVO-specific format.
I think we've got diverted by VOTable because we feel 'we have an agreed format' but this is misleading. Effort spent on agreeing how a SED (for example) will be described in VOTable will be much better spent agreeing on an SED schema - there is no more work required in agreeing such a schema than in how it will be represented in VOTables, and we then have an industry standard way of validating SEDs, describing web services that take or produce SEDs, etc. We lose all that using VOTable as a starting point.
We can then expect SEDs to evolve; indeed we can expect the way that we create the schemas to change. We can expect certain common elements (eg metadata ones) to appear across different schemas that might originally come from VOtable, and eventually be factored out into a metadata schema that is 'included' in these new formats. This is a Good Thing!
It may be that people are frightened of working with new schemas rather than VOTable (which is a kind of 'schema' the astronomers are familiar with), but it's not that hard, really! And using this VO-specific votable as our 'schema' that will hold all astronomical data formats will *not* encourage astronomers to take up our services.
Cheers,
Martin
-- Martin Hill 07901 55 24 66 www.mchill.netReceived on 2004-04-14Z14:44:37