Dobbs, Aaron wrote:
> Staying with the MARC / ISO2709 notation does us few favors with personal names.
As I said, forget ISO2709, it can be entirely gotten rid of without any
changes in MARC substance. It is ONLY still around because of all those
tools and interfaces, and it doesn't sit in most ILS's anyway,
internally.
> The MARC+AACR2 implementation of naming makes it very difficult...
>
> This notation: "<name>Smith, John</name>" from "<bag1>" could be either:
>
> <bag1>
> <surnomen>Smith</surnomen>
> <praenomen>John</praenomen>
> </bag1>
>
> Or
>
> <bag1>
> <personalname>
> <reference1>Cloud-based Personal Names Authority File1 ID</reference1>
> <reference2>Cloud-based Personal Names Authority File2 ID</reference2>
> </personalname>
> </bag1>
>
> (there may be potential problems with these Anglo-centric naming conventions)
>
There *will* be these problems. And don't forget terminology changes and
designators that become politically incorrect and all that.
Numeric tags have been and will be the method of choice; there's no
salvation in verbal tagging. This does not preclude exports in all
manner of copiously verbose ways, I'm talking of exchange standards
only which are for content that must be stable for a relative eternity.
> Continuing the theme, each work in a "local" catalog could be referenced to a Cloud-based Works Authority File,
Sure, a nice idea, but don't drive this too far; it creates new
dependencies on infrastructure and will likely be performance critical.
>
> Discoverability becomes a query against a separate, large, common index
Yes; this could then also be a switchboard between national standards
for extending queries beyond borders.
>
> I'm not saying MARC is bad; however, I am suggesting the current
> Local-ILS implementation of MARC may be an evolutionary dead-end.
>
Maybe, but this doesn't help as long as there is no new strategy
that does more at lower cost.
B.Eversberg
Received on Tue Apr 20 2010 - 09:44:18 EDT