VOEvent 20110215 notes

[Norman Gray]

Steve and Bob, hello.

On 2011 Mar 17, at 23:23, Steve Allen wrote:

> On Thu 2011-03-17T16:21:08 -0700, Bob Denny hath writ:
>> 
>> I would strongly suggest we explicitly adopt the date/time format as understood
>> by XML (type="xs:dateTime"). This is indeed a sub-set of ISO-8601. Straying from
>> this would preclude machine parsing of dates within an XML code/object-model
>> binding.

I agree with Bob, though there does seem to be some ambiguity in <http://www.w3.org/TR/xmlschema-2/#dateTime> over what year '0000' represents.  Quoth the spec:

> Those using this (1.0) version of this Recommendation to represent negative years should be aware that the interpretation of lexical representations beginning with a '-' is likely to change in subsequent versions.


Back to Steve:

> The XML dateTime denies the existence of leap seconds, so if any
> timestamp for an event happens to coincide with one that will not
> allow it to be legally expressed.

If I'm understanding the spec correctly, then the denial of leap seconds appears only in Appx E, talking about arithmetic with dates and offsets.  In the lexical specification of the dateTime type, it appears that all of the fields are unconstrained, so that as far as I can see, "9999-99-99T99:99:99 is a lexically valid time.  That's slightly unexpected, but appears to be permit a leap second to be expressed, even if the arithmetic on it can't work as that spec's Appx E says.

Perhaps, then, the VOEvent spec needs to simply cut-and-paste the lexical definition in the XSD spec (essentially, anything matching "'-'? yyyy '-' mm '-' dd 'T' hh ':' mm ':' ss ('.' s+)? (zzzzzz)?"), but noting that leap seconds are permitted, noting that this is compatible with XSD, and deciding what negative years mean.

The document notes that "The timescale — for curation purposes only — is assumed to be Coordinated Universal Time (UTC)."  Is that general enough, by the way?  Would that be acceptable to pulsar timing folk, or do we have to permit TDB and all that jazz?  I can't remember.

Best wishes,

Norman


-- 
Norman Gray  :  http://nxg.me.uk