Testimonial

[Rob Seaman]

> Oh no!  The Counter-Reformation!

Let me expand on this.  VOEvent needs to avoid a schism.

Time series support is an important element for building specialized  
time domain workflows.  Support for tables, on the other hand, is an  
element in a more general VO publication model.  Both are important,  
but workflows strike closer to VOEvent's soul of enabling autonomous  
architectures for transient response.

VOEvent can address both by supporting *both* SimpleTimeSeries and  
references to external VOTables.  As Roy has discussed, the latter can  
reference local resources such as a VOTable entrained over the same  
transport.

Ensuring that SimpleTimeSeries properly captures the essence of an  
astronomical time series is an ambitious enough chore for v2.0.  I've  
asked John to post his latest version of the schema for comment from  
the WG.  He has been working with Arnold to get the coordinate issues  
pinned down.

Perhaps we will later broaden the schema to address more use cases.   
Or perhaps we will add a separate SimpleTable subschema.  Or perhaps  
references to external tables will continue to suffice.  One  
anticipates that LSST requirements will drive the continuing evolution  
of these features.  However, v2.0 is an important milestone to  
capturing the market share needed in the short term to guarantee that  
VOEvent remains at the heart of the effort as LSST and SASIR and other  
major projects develop over the next several years.

Rob