[dsdl-discuss] Re: Datatypes strawman

Rick

[Issue 1-5: The XML Schema representation of period only allows decimals to
be used to qualify seconds. Should DSDL be similarly constrained, or should
it remain more compatible with ISO 8601?]

The XML Schema WG had a strong disinclination to allow multiple lexical
forms.
But if ISO 8601 allows them, so should we, with a caution that XML Schemas
does not.

OK, so you want to allow the abbreviated form of 20020202T02.2 for 02:12 on
the 2nd of Febuary 2002 as well, but what about all the other options in ISO
8601, such as W21 for week 21, etc? I have no problem with defining dateTime
as any valid ISO 8601 format, but then I would argue that we should not have
separate datatypes for things like Gregorian Recurring Day in Month (gDay),
etc, but would simply allow any permissible subset of ISO 8601 to be
applied. The problem then comes with processing dates. Its much easier to
process the subsets than the whole thing. And what happens if users want to
instruct something to happen on the last day of each month, a situation ISO
8601 fails to cope with?

>There are several different lexical forms for number, date, and boolean,
but they
tried to keep these to a minimum.

>While I agree that a canonical lexical form is useful, I completely don't
buy that the
document should be restricted. Instead, transformations should be able to be
specified to convert to canonical lexical spaces (notionally) and thence to
values spaces.

I have restricted the lexical forms for canonical number to simple decimal
and floating point representations, provided just two options for the
boolean and distinguished periods from specific dates. There are two
"commonly used binary formats" and a general purpose example of notation. (I
wondered whether to subsume the binary formats within the notations one, but
have not yet succumbed to this idea.) From this small set of cannonicals a
wide set of derivatives has been defined. Are there any useful datatypes
that cannot be defined based on these primitives? (I'll come back to the ISO
11404 question in an answer to your other message.)

Martin Bryan

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)