It makes me think that, as much as as we want to have some restrictions so
that the metadata is structured, it may not be wise to use only the W3C
date and time formats here (i.e. xs:date), because then you can't come up
with any conventions that says what you want, because it isn't defined in
that profile of 8601. We had some discussion about this in the initial
draft of the schema, because the data dictionary itself doesn't restrict
dates to this form. It allows for both alternatives defined in ISO 8601,
i.e. YYYY--MM-DD or YYYYMMDD for one thing.
I've been involved in discussions within the Dublin Core community (and
elsewhere) to write a new W3C note to allow for different types of dates
(as well as to allow for the YYYYMMDD form)-- W3CDTF is very limited and
doesn't cover all that we need. It doesn't seem that this work was ever
completed. So no help here.
This is something to be discussed in the upcoming revision of the PREMIS
On Fri, 8 Dec 2006, Charles Blair wrote:
> On Fri, Dec 08, 2006 at 10:26:37AM -0500, Youn Noh wrote:
> > Apparently, 0000 will eventually be allowed for 1 BCE and xs:dateTime may
> > be truncated:
> > http://www.w3.org/TR/xmlschema-2/
> thanks. there i read:
> D.3.2 No Year Zero
> The year "0000" is an illegal year value.
> D.3.3 More Than 9999 Years
> To accommodate year values greater than 9999, more than four digits
> are allowed in the year representations of dateTime, date, gYearMonth,
> and gYear. This follows [ISO 8601:2000 Second Edition].
> this implies that 9999 is considered a real value, not code to mean
> something else.
> i think that the restriction to ISO 8601 for these dates is akin to
> being asked in a court of law "have you stopped beating your wife
> yet?" and only being permitted to answer yes or no.