[stdsreg] Message from Makx Dekkers, Dublin Core

[Ketchell John <john.ketchell@cenorm.be>]

John,

> We held our conference call meeting yesterday on the above;  I
> imagine that
> you had another obligation, but I hope you will be able to join a future
> discussion.  The next is on Tuesday 21st May at 18h Brussels time.

Very sorry about this. I did not read your previous e-mail well and
thought the call would be on the 17th rather than on the 7th. I have
now noted the 21st of May in my diary and intend to participate.

> I understand that Bob Hager at ANSI (in copy) has also been in touch with
> some DC people on the US side, so maybe we can close the loop.

Yes, Bob has talked to Stu Weibel and this led to our willingness
to participate in this activity.

>
> Meanwhile maybe you can shed some light on one issue mentioned yesterday?
> Standard ISO 11179-3:1994 contains metadata attributes generally
> drawn from
> the DC reference description version 1.1, but there are also some detailed
> differences.  It was not clear to the participants why these differences
> exist and whether there is any significance to them.

As ISO 11179-3 was written before DC was even started, the
attributes were not 'drawn from DC 1.1', rather the definition of
DC 1.1 tried to base itself on ISO11179. Now we had a question
recently from someone who questioned a statement in the reference
description of DC1.1:

> I have a general question about the "Dublin Core Metadata Element Set,
> Version 1.1, Reference Description:"
>
> The text says that "Each Dublin Core element is defined using a set of ten
> attributes from the ISO/IEC 11179 standard..."
>
> It appears to me that the elements are being defined by ten data attribute
> descriptors (as specified in section 4.1 of 11179-3), rather than
> the actual
> data element attributes themselves (which are specified in
> sections 5 and 6,
> and of which eight are mandatory).  My understanding is that the
> descriptors
> apply to the attributes, and the attributes apply to the data
> elements.  Am
> I misinterpreting ISO/IEC 11179-3?

My answer to this was the following:

> Your question is certainly interesting and gave rise to quite a bit
> of internal communications. The text you refer to was written a
> couple of years ago by someone who is no longer working with us.
>
> I tried to familiarise myself with ISO11179 and find out what the
> idea behind the text in the Reference Description of DC1.1 is.
>
> From my analysis it appears that we used the ten General Descriptors
> of ISO11179 (that are supposed to be the descriptors of data element
> attributes, in our terminology the elements of the Dublin Core) as a
> basis: name, definition, obligation, condition, maximum occurrence,
> datatype, maximum size, character set, language, comment.
>
> We deleted three of those (condition=none, maximum size=unlimited
> and character set=ISO636) that we didn't see use for, and added
> three more from the basic attribute set (Identifier, Version and
> Registration Authority) that we needed for maintenance of the DCMES.
>
> These last three are actually used in ISO11179 as attributes for the
> data elements (in their terminology the things that are described,
> so in our terminology that would apply to the resources).
>
> We use the Identifier to link translations of the Dublin Core
> element set (something that ISO11179 does not seem to allow for),
> and the other two are metadata about the Dublin Core Metadata
> Element Set itself. Again, it is strange that ISO11179 does provide
> a mechanism to control versions and identify Registration
> Authorities for data elements (which for them are things like lists
> of country codes) but not for the attributes/metadata elements.

This of course is based on my understanding of the principles of
ISO11179 and I would be happy to hear from people who are more
familiar with the history and background of ISO11179. I hope this
sheds some light on the discrepancies.

Please forward this to the exploder - I will try to subscribe today.

Makx.







------------------------------------------------------------------------
Best regards
John Ketchell
Director, CEN/ISSS - Information Society Standardization System

URL:http://www.cenorm.be/isss

Rue de Stassart, 36	email (direct) john.ketchell@cenorm.be
B-1050 Brussels	email (secretariat) isss@cenorm.be
Belgium	                        Tel (direct) + 32 2 550 08 46
Fax + 32 2 550 09 66	Tel (secretariat) + 32 2 550 08 13
Tel (GSM) +32 475 594 828