see end...
----- Original Message ----- 
From: "D P Ingram" <[log in to unmask]>
> On 23 apr 2007, at 08.14, Steven C. Barr(x) wrote:
> > The "Abrams files" are a set of text-based data files compiled
> > by Steve Abrams, which provide label/number/matrix/credited
> > artist for about 300,000 78rpm phonorecord sides, give/take.
> > They do NOT include classical-series phonorecords...and their
> .... Thank you for this information. I am slowly developing the idea  
> of a music catalogue for audiophiles and collectors that would  
> require tight data entry controls to ensure integrity of materials  
> and either "sight of record" or many sources to corroboate the  
> information.   But each time I pause for breath the project expands  
> and the technology demands seem to rise.   It is growing out of a  
> need to catalogue over 100,000 records as there are certainly many  
> duplicates there and it is getting hard to remember what one has and  
> what one doesn't have.
> As many sources of information to piece a jigsaw together are good.
> > However, they provide a foundation on which to construct the
> > ultimate 78rpm database...a list of EVERY 78rpm disc ever issued...!
> Out of interest, what to you constitutes the ultimate database other  
> than every disk issued? Other than the usual track/matrix/artist  
> information ?
Well...remember that, to me, "phonorecord" means "78rpm record"...if only
because the imposing task of trying to list EVERY sound recording...analog
or digital, regardless of format...that has ever existed is simply too
overwhelming to even THINK about!

Now, a discographic database is NOT the same as a catalog (in spite of
the fact that catalogs could be an important data source for such...!).
A catalog "describes," using a set of data fields, the subset of 
phonorecords held by a SINGLE collector/institution/wotever. Thus,
entries like "Condition," "Price Paid," "When acquired" and such
become applicable (note that these aren't, and CAN'T be in any practical
sense, part of a discographic database!). As well, a catalog should
provide on the storage location of a given phonorecord in the collection;
a discographic database has the option of identifying holders of a given
phonorecord, but nothing more exact is needed.

IMO, we can assume that as many as 3 million 78's were issued here in
North America...and I have no idea of similar estimates for other
countries in the world. That means our "ultimate 78 database" (which
has to be a discographic one, since no such collections exist...even
Lennick's...!) will have to contain data on three million phonorecords...
or, more likely, six million sides. The associated data fields COULD
include all sorts of items (one idea might be images of the sides'
labels, or even the full sides...?)...but will need to include the
usual fields...Label, Country, Title, Credited Artist, Actual Artist,
Matrix (if known), Take (ditto), Type of label, Date Recorded, Date
Issued and a good-sized Remarks data field (a necessity in data tables!).

In other words, the data necessary to define the record...who made it,
what songs, when (as close as we know), and discographic data like
whose matrix, what number, which take, control number (if any)...

Steven C. Barr