Hi John,
I belive the W3C XML Schema WG will:
- permit use of the "/" separator for ranges,
- separate out:
- the general datatype, eg ISO 8601
- from an optional specific profile, eg ISO 8601 with 4-digit years
- if no profile is specified, then *all* of ISO 8601 will be permitted
If you want to allow the "/", the issues before you include:
- where? everywhere?
- the "/" is not in the WTN and there is currently no prospect of that
Note being reissued
- you might decide to define a DC profile which points to the WTN,
plus allows the "/" [some|every]where
I, personally, would not permit the whole of ISO 8601 and would,
therefore, go for some string other than "ISO8601". A problem with
"WTN8601" is that the meaning of the "8601" is unclear. How about
"ISO8601DC", ie a profile of ISO 8601 based on the WTN, but (possibly)
allowing the "/"?
Misha
> How should we name our canonical Date scheme (the W3C profile of ISO8601)?
>
> So far people have used "ISO8601", but in the "DC in HTML" draft [1] I'm
> effectively proposing the new name "WTN8601" by showing examples of it.
> There's been little comment, but this is too central a change to let
> pass without discussion, and therefore I'm asking for guidance from
> the DC community.
>
> Here's my analysis. The concept that we're dealing with is
>
> "A profile (or subset) of the ISO8601 Date/Time scheme, as defined
> in a W3C Technical Note." [2]
>
> which I summarize as
>
> "A W3C Technical Note profiling ISO 8601."
>
> and then boil down to the one token/word, "WTN8601", suitable for use with
>
> scheme = "WTN8601"
>
> This is what I used in examples appearing in the "DC in HTML" draft.
>
> So What's The Problem?
> ----------------------
>
> Unfortunately for many months now, we have (without discussion I think)
> tacitly consented to the sloppy practice of referring to our particular
> flavor or subset of ISO8601 in a way that is indistinguishable from the
> much more complex mother concept of ISO8601 (which allows ranges, 2-digit
> years, other short cuts, alternate forms, etc). So now abuses such as
>
> scheme = "ISO8601"
>
> pollute existing metadata collections, incorrectly advising others that
> processing a Date string may require the heavy software artillery of a
> fully compliant ISO8601 parser.
>
> The problem lies not with the providers of these collections, but with
> the DC community which failed to invent a distinguished name for its
> particular profile of ISO8601. Now we have a decision point.
>
> Choices:
> --------
>
> 1. Continue using the name ISO8601 to refer to our profile of ISO8601,
> as defined in the W3C Technical Note [2].
>
> CONS:
> -----
> - calls something ISO8601 that isn't ISO8601; agents that only parse
> the profile will incorrectly believe they can't process the element
>
> - to distinguish real ISO8601, a new name needs to be invented, e.g.,
>
> scheme = "REAL_ISO8601"
>
> - co-opting the term "ISO8601" will create real political trouble with
> the ISO directorate; esp. bad if DC wishes to win acceptance there
>
> PROS:
> -----
> + installed base argument -- doesn't disturb existing collections
>
> + software and people who use DC will "know what we mean" anyway
>
> + full ISO8601 may be unimplementably complex and never exist anyway
>
> + if we co-opt "ISO8601" early, our subset meaning should prevail
>
> 2. Discontinue using "ISO8601" to mean our profile; use "WTN8601" instead.
>
> (PROS & CONS: reverse above lists, change names, get totally confused...)
>
> 3. Discontinue using "ISO8601" to mean our profile; use a new name instead.
>
> (Propose a new name other than "WTN8601".)
>
> 4. None of the above.
>
> (PROS & CONS: please give your reasoning)
>
>
> Comments by the end of April
> ----------------------------
>
> Please try to get your comments on this and other aspects of "DC in HTML"
> to the dc-general list by the end of this month.
>
> Thanks,
>
> -John
>
>
> References
>
> [1] Encoding Dublin Core Metadata in HTML
> <ftp://ftp.ietf.org/internet-drafts/draft-kunze-dchtml-00.txt>
>
> [2] W3C Technical Note - Profile of ISO 8601 Date and Time Formats
> http://www.w3.org/TR/NOTE-datetime
>
>
> Relevant excerpt from current "DC in HTML" draft:
> -------------------------------------------------
>
> Date (of creation or availability of resource; [WTN8601] recommended)
> ----
>
> <meta name = "DC.Date"
> content = "1972">
>
> <meta name = "DC.Date"
> content = "1998-05-01">
> <meta name = "DC.Date"
> scheme = "WTN8601"
> content = "1998-05-01">
>
> <meta name = "DC.Date.Created"
> content = "1998-05-01">
> <meta name = "DC.Date.Available"
> content = "1998-05-11">
> <meta name = "DC.Date.Valid"
> content = "1998-05-20">
>
> <meta name = "DC.Date.Created"
> content = "triassic">
> <meta name = "DC.Date.Acquired"
> content = "1957">
>
> <meta name = "DC.Date.Accepted"
> scheme = "WTN8601"
> content = "1998-01-05T08:15">
>
> <meta name = "DC.Date.DataGathered"
> scheme = "ISO8601"
> content = "1998-01-05T08:15/1998-01-05T13:15">
>
> <meta name = "DC.Date.Issued"
> scheme = "ANSI.X3.X30-1985"
> content = "19980501">
------------------------------------------------------------------------
Any views expressed in this message are those of the individual sender,
except where the sender specifically states them to be the views of
Reuters Ltd.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|