Re: Databases for multi-image documents -Reply

Bob Rosenberg (mailto:rarosenb@RCI.RUTGERS.EDU)
Fri, 8 Mar 1996 14:56:30 -0500

Message-Id: <mailto:199603082021.OAA29693@library.wustl.edu>
Date:         Fri, 8 Mar 1996 14:56:30 -0500
From: Bob Rosenberg <mailto:rarosenb@RCI.RUTGERS.EDU>
Subject:      Re: Databases for multi-image documents -Reply
To: Multiple recipients of list IMAGELIB

I would like to thank the list members for all the replies I got, on and off
the list.  Lists are, in my view, one of the premier functions of the net,
and this one performs admirably.

The long and the short of my conclusion is this: We will probably (why stop hedging now? Anything could happen in the next two months) use a relational data structure, with documents records in one table and image records in another, linked by our document identifiers (which are archival designations, arbitrary but unique). The image records will have an ordinal field to note which image they represent in a multi-image document. There's more, but that's the essence of the answer. Since we have a site license for Oracle, that is a good bet as software.

Bob Rosenberg

Robert Rosenberg * "Genius is all bosh. Thomas A. Edison Papers * Clean hard work is Rutgers University * what does the business." mailto:rarosenb@rci.rutgers.edu * --TAE