Re: VOTable alternative?

> (1) I agree with Bob that we should 'evolve' the VOTable specification
> and toolset, rather than move to something else.

I'm sorry to day I probably agree with this, while VOTable isn't perfect I think we may be too far down the standards path (with enough invested) that we shouldn't rock the boat too badly.  

> (2) Because we are arguing over syntax, not semantics, our job should
> not be difficult.

Weeelll...

Depends, I think we are (or will be in part) arguing over semantics, at least, we will be when it comes to time series data. For instance (having done both) the X-ray and optical view points are somewhat different. I'm sure people will have opinions!

> (3) Therefore I suggest a "pure XML" skin for the VOTable that I shall
> VOTable2. It is derivable from VOTable (and back?) and allows a
> VOTable to be used for:
> -- validation
> -- automatic code generation
> -- subclassing
>
> (4) In VOTable2 we split VOTable file into (a) a "record schema" is
> made from the metadata of the table, and (b) a conforming document is
> made from the data part, that is a sequence of these records.
>
> Therefore in place of:
> <TR> <TD>red</TD> <TD>2.1</TD> </TR>
> we see
> <colour>red</colour> <mass>2.1</mass>

I like it, makes sense to me, hideously verbose, but thats not necessarily a bad thing for small documents being passed between services. For larger documents (tables) the data element will (obviously?) be a binary table and the meta data document will describe it (as we have now)...

Al.

-- 
Dr. A. Allan, School of Physics, University of Exeter