Consensus

[Roy Williams]

On 12/10/2010 4:21 PM, Arnold Rots wrote:

 > This is a totally inadequate format since it carries no metadata
 > whatsoever. This leads us back to pre-VO days where no one bothered to
 > include complete metadata, because, after all, all friends and buddies
 > knew exactly what was meant - or so one pretended.

The Table element can carry the metadata that Doug Tody defined in the 
recent Time Series document -- if we allow the Params and Fields to have 
a Utype. The metadata then goes into the utype/ucd/unit/description 
attributes. Mike wrote "IVOA is now defining a common understanding of 
TimeSeries", and I take that to mean the document that Doug wrote:
http://www.ivoa.net/internal/IVOA/LightCurves/STSP.pdf

So I don't understand -- perhaps you mean that my simple example light 
curve does not carry enough metadata? Then I agree entirely.... it was 
just an illustration!

 > I suppose I can point to stcTab.xsd till I am blue in the face and
 > still have no one take this stuff seriously.

We decided to follow VOTable, another IVOA standard, rather than the STC 
version of table.

 > I also proposed a compromise last spring that I would have been
 > willing to defend as IVOA-compliant, but there did not seem to be any
 > interest in that, either.

Perhaps this would be a good time to start that discussion again?

Roy





> Roy Williams wrote:
> [ Charset ISO-8859-1 unsupported, converting... ]
>>
>>> IVOA is now defining a common understanding of TimeSeries, why should
>>> this differ from the VOEvent version of the same data-type?
>>
>>
>> Mike.... Not sure if my earlier message made it, so I will repeat it. I
>> think the time series below is not very different from the iVOA common
>> understanding ...Roy
>>
>> -----------------------------
>>
>> There is an alternate representation of time series in the VOEvent2
>> standard, through the Table element, which is based loosely on the
>> VOTable structure. But specifically, it could be used for a time series.
>>   From Doug Tody's recent document on representing time series, I have
>> taken some utypes, and come up with the following very simple timeseries
>> that can be used in VOEvent2 (below).
>>
>> Cheers
>> Roy
>>
>>
>> <Param name="zeropoint"
>>          utype="CoordSys.TimeFrame.Zero"
>>          value="51234.0"/>
>> <Param name="band"
>>          utype="CoordSys.Band.GenID"
>>          value="R"/>
>> <Param name="timeunit"
>>          utype="Char.TimeAxis.Unit"
>>          value="day"/>
>> <Table>
>>     <Field name="Time"
>>            utype="Char.TimeAxis.Coverage.Location.Value"
>>            unit="day"
>>            ucd="time.epoch"/>
>>     <Field name="Rmag"
>>            utype="Data.FluxAxis.Value"
>>            unit="mag"
>>            ucd="phot.mag;em.opt.R"/>
>>     <Data>
>>       <TR><TD>1.635</TD><TD>16.78</TD></TR>
>>       <TR><TD>1.934</TD><TD>15.87</TD></TR>
>>       <TR><TD>3.456</TD><TD>14.99</TD></TR>
>>     </Data>
>> </Table>
>>
> --------------------------------------------------------------------------
> Arnold H. Rots                                Chandra X-ray Science Center
> Smithsonian Astrophysical Observatory                tel:  +1 617 496 7701
> 60 Garden Street, MS 67                              fax:  +1 617 495 7356
> Cambridge, MA 02138                             arots at head.cfa.harvard.edu
> USA                                     http://hea-www.harvard.edu/~arots/
> --------------------------------------------------------------------------
>