DALI 1.1 working draft available

[Patrick Dowler]

As has been discussed numerous times (!) in the past, JD and MJD are *not*
types. For MJD values (eg Obscore t_min and t_max) one uses
datatype="double" unit="d" and describes the coordinate system metadata
using STC as Markus has described in a Note. The xtype attribute is to tell
consumers of the votable that the value can be further parsed to turn it
into a structure of some type, not further interpretted to mean something.
I think that is an important distinction that is often ignored

I'm not sure about the high precision "days" + "seconds" double array. You
need that in TAP only if you actually want to expose a single column with
the multiple values, but I don't know off hand if anyone does that. (I will
admit to having some double array vaulued columns in our TAP service; they
are in a view called caom2.SIAv1 to support the coupleof columns in the
SIAv1 output that are arrays... but they don't have xtypes).

Finally, the restriction of timestamps to a specific ISO8601 in UTC and
without timezone comes from upstream of DALI -- VOResource iirc. It has
been th accepted standard for years and DALI is only formalising some
technical bit of it for usage in services like TAP, SIA2, etc

Emphasis: xtype tells the consumer how to further parse the value.


-- 
Patrick Dowler
Canadian Astronomy Data Centre
Victoria, BC, Canada
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dal/attachments/20160219/75a45807/attachment.html>