An STC "when" and "where" example

[Alasdair Allan]

Rob Seaman wrote:
> Note that the real complaints are with the verbosity of XML.

It was nothing it do with the verbosity of the XML. It's XML, it's 
always verbose, it's designed to be verbose...

What I was complaining about was the baroque nature of the STC 
specification. The total lack of any assumptions, of optional things, 
things that can be left out because there is a default and the default 
is known because it's declared as standard by the schema. You don't 
have to specify every last little detail very time you want to create a 
document.

> Here is Arnold's original example as an STC-S string expression:
>
>>> ObsDataLocation
>>> ObservatoryLocation
>>> Position GEO_D TOPOCENTER SPHER3 248.4056 31.9586 2158
>>> ObservationLocation
>>> Time UTC TOPOCENTER 2005-04-15T23:58:55 Error 5
>>> Position ICRS TOPOCENTER 148.88821 69.06529 Error 1.0 1.0
>>> Spectral TOPOCENTER 460 unit nm Resolution 40
>
> This is fewer characters than your example with a much greater 
> informational content.

Yes, but it's not self describing. Which is the entire point of XML and 
a schema or DTD, which is having a specification you can validate your 
document against in real time. You can do this with the above 
serialisation.

> Perhaps we're just viewing the situation too closely.  We aren't 
> really seeking a standard way to convey coordinates and times.

Sorry, I thought that was the entire point of STC?

>  VOEvent is seeking precisely what STC is selling:  a standard way to 
> convey "Space-Time (and relate coordinate axes) metadata".

I think using serialised STC is overkill for something like VOEvent.

I think we need something much simpler, at least to start with, it's 
possible that in time, we may need to add complications into the 
specification. But to start with at least I think we have to, must, 
keep the VOEvent spec as lightweight as possible.

Sorry, if I'm going to have to implement this thing I'm not going to be 
stuck with something that'll take me 3 months or more to write a parser 
for...

>  Arnold and his collaborators view velocity/redshift as a key element 
> of a coordinate specification and therefore spectral information also. 
> Bottom line - what is a VOEvent without a spectral characterization?

Eh? Unless we're trying to design something very different than I think 
we're trying to design, I really don't see spectral characterisation as 
being central to a VOEvent specification.

Al.