> From: Geoff Mottram [mailto:[log in to unmask]]
> Subject: [MODS] Typeless names
> Since you have done such a wonderful job of simplifying the
> MARC format in
> MODS, I would like everyone to consider removing the vestigial tail of
> distinguishing between types of names (personal, corporate,
> etc...). Is there really a need in this day and age to make such a
I don't think that you want to remove the distinction between names but
you do not need separate tags. Basically, you can think of the problem
you describe in terms of what the XLink standard does with its "role"
attribute. The "role" attribute in the XLink standard allows you to
define a relationship to the object you are describing. The benefits
of thinking in an object/relationship fashion is that you only need to
define one generic object and some standardized URI's for relationships.
Implementers can define additional relationships, by using their own
URI's. What's important here about "role" is that the "object" is still
the object. You are not redefining the "object" but how you are using
that object. As far as the MODS standard it could implement this as
The MODS standard would define the "object" [in MODS context is would
be called aspect], e.g. <mods:name> where any name can be placed. The
mods:role attribute could be optional, in which case the contents of
the object is a name with an undefined relationship. The MODS standard
would then create the following URI's for use in the mods:role attribute:
As an implementer, if I needed to make the distinction between "for profit"
and "not for profit" corporations then I could define my own URI's to be
used as roles. The important point here is that I still would specify the
aspect, e.g. XML element, as <mods:name>. Anyone making use of the metadata
would know that <mods:name> is a name aspect of the metadata object. What
role that name is playing in the relationship to the metadata object may or
may not be important to the person/program using the metadata.
> Over the past 17 years I have developed and sold MARC-based cataloging
> systems to institutions with large non-book holdings. The
> biggest hurdles
> to "selling" the MARC format to archives and museums has been
> the steep
> learning curve and the additional time and cost required when
> creating AACR2
> MARC records. The problem is caused by the excessive
> subfielding of data
> and the need for so much coding and classifying of each
> element ("this is a
> forename, a surname, a family name, a personal name, a
> corporate name, a
> conference" and so on). Not only are these seemingly
> arbitrary distinctions
> hard to explain but they tend to drive users to non-standard
> and proprietary
I disagree with your notion about excessive encoding of the data being
arbitrary distinctions. As a matter of fact AACR2 doesn't do enough
encoding. Markup is an important concept in any metadata standard. The
markup allows you to easily retrieve aspects of the metadata object. You
could define a metadata object as one blob of text. It would not be very
useful for machine processing. Markup allows you to find the tree in
For example in the MODS Map test record, the MARC21 255 element mapped to
the XML <cartographics> element, sub-element <coordinates> and <scale>.
The current MODS standard plops the AACR2 form down into those elements.
The problem is that <coordinates> should have been broken down into the
individual coordinate aspects of longitude and latitude rather than
relying on the AACR2 punctuation. This is what metadata markup is all
about. I should not have to parse punctuation to derive content.
As MARC21 record encoding now stands, LC's own records don't always have
the correct punctuation. This makes it difficult to pull those longitude
and latitude aspects out of the AACR2 data thus requiring all kinds of
special cases. This may be the problem with AACR2 that you eluded to above.
However, as I previously indicated, you do not need five different tags
to describe the same metadata aspect. You should describe one aspect
that has varying roles.