VOEvent working draft published: version, param

[Tony Linde]

Hi Rob,

> I'm skeptical about my ability ...
> Aren't extension schemata therefore automatically supported?

I'm probably even less capable on the xml schema front.

So, you may be right about the automatic support - I was arguing that it be
explicitly supported with examples so that people could see that it was just
as easy if not easier and much less prone to error.

Cheers,
Tony. 

> -----Original Message-----
> From: Rob Seaman [mailto:seaman at noao.edu] 
> Sent: 24 June 2005 23:36
> To: Tony Linde
> Cc: Rob Seaman; voevent at ivoa.net
> Subject: Re: VOEvent working draft published: version, param
> 
> Tony suggests:
> 
> > Add the option of extensions into the spec, explain how 
> they make the 
> > data both interpretable and correctly typed in order to secure the 
> > operation of downstream applications, BUT also allow the 
> <param> and 
> > <group> elements as now.
> 
> I'm skeptical about my ability to explain any such thing 
> myself - however, a VOEvent "metadata packet" is simply an 
> XML document.  This is what the specification already says 
> about the subject:
> 
>      "XML structures other than those listed in this document 
> should be used with care within a <VOEvent> element."
> 
> Aren't extension schemata therefore automatically supported?
> 
> On the broader topic of included a schema, note that the full 
> majesty of STC is available for any VOEvent publisher 
> (strictly speaking,
> creator) to craft an arbitrarily rich coordinate expression - 
> an expression of formally correct temporal or spatial 
> coordinates that is likely to be completely unintelligible to 
> an unwary, which is to say, unprepared, subscribers.  It 
> seems to me that the real issue is completely independent of 
> either strong or weak typing or naming or XML schemata in 
> general.  Rather, each VOEvent aware application has to 
> encapsulate some level of comprehension of astronomical 
> "meaning", whether this is called semantics or ontology or 
> "knowledge of the problem domain".  I really doubt that the 
> world's best schema (and I'm a strong supporter of STC) will 
> resolve this issue.
> 
> Bottom line is that VOEvent is only of value if issuers of 
> events adopt the standard and if robotic observatories are 
> willing to automatically commit their expensive assets to the 
> resulting alerts.
> 
> Rob Seaman
>