Agreed,
I believe that this issue was raised a while back (before we
wrapped things up) where he suggested that the spec REQUIRE
the use of the canonical represntation of timeInstant (now
dateTime).
I thought we had adopted this, but I see that the spec doesn't
reflect this requirement. I don't recall if there was any opposition.
Barring any research back to the previous drafts of the XMLSchema
specification, I would recommend that we adopt an errata that
both updates the ebXML MS specification to conform with the
now W3C Recommendation for XMLSchema [1] as well as adopts the proposal
that all dateTime elements be required to be expressed using the
canonical representation of dateTime as defined in section 3.2.7.2
of [1].
Further, since the canonical representation may omit the timezone,
its absence SHALL be interpreted as inferring UTC.
Therefore, I would propose that the following change/errata be
adopted.
[802] The REQUIRED Timestamp element is a value representing the
time that the message was created. It SHALL be expressed as a
canonical representation of the dateTime datatype as defined in
section 3.2.7.2 of [XMLSchema-2]. Absence of the "Z" timezone
qualification for Coordinated Universal Time (UTC) SHALL be
inferred as meaning UTC.
[1] http://www.w3.org/TR/xmlschema-2/#dateTime
Cheers,
Chris
"Cherian, Sanjay" wrote:
>
> Hi,
>
> These are the places in the specification where <TimeStamp/> appears:
> 8.4.6.2, 8.5.2.3, 8.6.1, 8.13.2 and 8.14.1.
> One of the descriptions has been included below for reference:
>
> [Line 801] 8.4.6.2 Timestamp element
> [Line 802] The REQUIRED Timestamp is a value representing the time that
> the message header was created
> [Line 803] conforming to an [XMLSchema] timeInstant.
>
> This statement does not mandate that UTC be used when date and time is
> specified with <Timestamp/>. On the other
> hand, there is no facility in the CPP to define the time zone that one party
> operates within and no facility in the CPA
> to capture the timezone that two parties must agree upon.
>
> The use of UTC is suggested through all the examples in the specification
> that use <Timestamp/>. For example:
>
> [Line 953] <eb:Timestamp>2000-07-25T12:19:05Z</eb:Timestamp>
>
> I understand that the Z at the end indicates that the time specified here is
> relative to UTC (the same as GMT).
>
> In my opinion, it is a small matter to require the use of UTC in the ebMS
> specification.
>
> Thanks,
> Sanjay J. Cherian
> Software Architect
> Sterling Commerce
> Irving, TX
>
> ------------------------------------------------------------------
> To unsubscribe from this elist send a message with the single word
> "unsubscribe" in the body to:
