Re: [ARSCLIST] Memorex CDs and more!

[steven c <stevenc@xxxxxxxxxxxxxx>]

----- Original Message ----- 
From: "Jon Noring" <jon@xxxxxxxxxxx>
> What is really needed is not a database format, but a data model for
> discographical information that can be expressed in XML so the data
> may be exchanged between different database systems.
> 
> There's a lot of issues involved with designing such a data model,
> including understanding the scope, what's been done already, other
> data models for related information (such as for contemporary
> recordings, song composition databases etc.) Designing such a data
> model is definitely a "committee" effort requiring a group of people
> who really know their stuff in *several* areas. One needs a couple
> people with general data modeling experience, one or two XML wonks,
> the usual experienced record collectors who understand all the
> "exceptions and oddities" (which is what really makes a 78rpm
> discographical data model complicated), experts of digital metadata
> archives, discographical researchers, etc., etc.
> 
> It's also a good idea to build an *open source* and *cross platform*
> reference implementation database application that will import/export
> the XML data, so it can shake out the data model itself. The key thing
> is that the data model should not be tied to any particular platform or
> proprietary application.
> 
Basically, I think what you refer to as a "data model" is what I was
thinking of when I said "database format," since I WASN'T thinking
of which database application should be used...but, rather, of a
definition of fields (size, field type, name and type of data to
be included) in a discographic database. Here, what should be
defined is a minimum set of fields, to be included in any
databases created to list sound recordings (including catalogs,
both private and institutional).

With this defined, then individual database users could add such
fields as they felt necessary for their own purposes...as long as
they used the core fields as "officially" defined. Particularly
with cataloguing databases, there is a fair amount of data
included that is only relative to the given set of phonorecords
being catalogued (i.e the whole world doesn't need...or want...
to know the date I bought my copy, what I paid for it or
which milk box it's in...but it DOES need to know the matrix
number and take on my copy, which could be different than
other copies...)

Here, I think ARSC could play a role in establishing the
standard set of "core fields" in databases covering
phonorecords...which would at least insure compatability
between database files, since commercial database programs
usually have a selection of export choices including
xBASE (and xBASE files can be used to produce readable
text files by stripping off header data).

The next task would be data entry from hard-copy items
(books, printed lists, written notes, usw.) that contain
data on sets of phonorecords (tracking those down will
be a formidable task, considering how many were privately
published if published at all!)

It's been a lifelong thought of mine...and present-day
computers make it possible (and a lot easier than typing
out 3,000,000 3x5 cards!

Steven C. Barr