At 03:39 PM 23/5/97 +0100, Jon Knight wrote:
>Hope the helps (and I'm awaiting the flood of 'Oh, no it isn't!' shouts
>from the other meta2er's... :-) ).
No disagreement from me Jon - what I need urgently is some (understandably
interim) recommendations I can implement. It seems to me that in many
cases all we need is someone to nominate a scheme name and a URL that
points to a usable short specification. By scheme name I mean an exact
token we can use in a (scheme=xxx) statement. EG does "ANSI X3.30" become
"(scheme=ANSIX330) and who maintains a short but accurate interpretation at
a specified URL that can be referred to by the Reference Description.
The schemes we use mightn't be the ones that 'win' in the end, but as long
as we agree on them and put them in the DC reference spec, we can implement
them, and any later implementations can recognise them and do conversions
as necessary (where practical eg date format).
Can someone (Stuart, Renato, whoever knows) give us an update on likely
timelines for a revised (post DC4) Reference Description, and news on
progress of working groups? PLEASE??
My immediate requirements are:
Generic Schemes
===============
Are the scheme names "URL" and "EMAIL" as I have proposed in EdNA metadata
standard
http://www.edna.edu.au/edna/owa/info.getpage?sp=&pagecode=5212
acceptable?
Can anyone give me a URL for a formal specification of URLs and EMAILs. Eg
I have only ever understood the use of quote marks around names and <>
around email addresses by example. Are there any issues in putting a quoted
name strings in a field which contains an email address.
DC.TYPE
=======
Jon - would you care to nominate a scheme name for
http://www.roads.lut.ac.uk/Metadata/DC-ObjectTypes.html ?
DATE FORMATS
===========
(for dc.date and dc.coverage.temporal)
I prefer a scheme that results in 1997-05-23 rather than 19970523 because
its just a whole lot more readable for a human. Also a format is required
for ranges, eg "1997-05-23 - 1998-05-22"
I've looked at
ftp://ds.internic.net/internet-drafts/draft-newman-datetime-01.txt on
Misha's recommendation but is *much* too complicated to refer anyone who is
trying to just understand how to format a date for inclusion in a DC
record. Also I think we need a scheme where time of day and corrections for
time zones are optional.
>> 3) CREATOR : is the recommended/default form First_name Last_name or
>> Last_name, First_name?
>
>I've always used "First_name Last_name" but as "Last_name, First_Name" is
>unambiguous (the comma differentiates them) I think the two can co-exist
>peacefully as one default.
=========
JG: This raises interesting possible clash with multiple values in a field.
eg: NAME=DC.CREATOR CONTENT="Smith, Adam" is one value but
NAME=DC.LANGUAGE CONTENT="French, English" is two values.
=========
>I notice that Rebecca mention's the "Library
>of Congress Name Authority File" as a SCHEME; I'd also add in "BibTeX"
>name formats as well as they are widely used in the scientific community
>(plus the spec is freely available on the Net). Same goes for the
>CONTRIBUTOR element as well.
========
JG: Can we get a URL for these and would someone like to nominate an agreed
scheme name?
========
>> 4) LANGUAGE : will the recommended/default form be the revised Z39.53
>> three character code or the ISO 639-2 or still "free text"?
>
>The jury is out on that one; the choice seems to sway with the wind, phase
>of moon and who you talk to. Z39.53 keeps DC aligned with the library
>community but ISO-639 keeps it align with IANA and the Internet community
>(more or less - our good old .uk domain being the only hiccup in the
>latter case due to historical accidents). Before DC-4 I left the
>DC-SubElement's draft's default as freetext to neatly sidestep the issue
>(ie: if you feel passionately that it should be Z39.53 or ISO-639 then
>fine, but you have to supply a SCHEME qualifier to tell people that).
>
>> 5) DATE : is the recommended/default form of date the ANSI X3.30 e.g.
>> 19970523 (as written in the Reference Description <URL:
>> http://purl.org/metadata/dublin_core_elements>) or the ISO 8601 e.g.
>> 1997-05-23 (as suggested in Dublin Core Qualifiers/Substructure : a
>> proposal by Rebecca Guenther <URL:http://www.loc.gov/marc/dcqualif.html>)?
>
>Again jury is out; ANSI X3.30 is/was popular, as was IETF.RFC-822 (the one
>I had as a default in the pre-DC-4 draft). According to Rebecca's
>document there is a DATE subgroup from DC-4 looking at this - anyone care
>to step forward and say what's happening with that? In lieu of someone
>doing that before your report needs to go in, I'd put ANSI X3.30 in as the
>default, but with a note that this may change.
JG: to me the issues is not 'what is the default' - to which I thought the
answer was always 'free text if no scheme is specified' but what is the
exact scheme name and what is the URL of the reference information for
whatever scheme(s) we use.
end =============================================================
Jack Gilding ph: (03)9628-4652
Project Manager, VET EdNA Project fax: (03)9628-2472
Communications & Multimedia Unit [log in to unmask]
OTFE, PO Box 266D Melbourne VIC 3001 http://www.edna.edu.au/vetwp/
(level 4 Rialto Sth Tower 525 Collins Street Melbourne Australia)
|