On Tue, 22 Dec 1998, Susan Hamburger wrote:
> HELLO! I hope some of you are working over these holiday weeks. As I get
> ready to embed the DAOLOC in the EAD finding aids and work with the Access
> database, we need to resolve what we are going to name each image.
Actually, this is already taken care of by the database. The database
automatically assigns a non-metadata encumbered file name to each image.
We (Berkeley) devised this system under Daniel Pitti's tutalage in the
California Heritage Digital Imaging and Access Project. The more we
wrestled with having file names that somehow related to the content or
provenance of a particular item, the more problems we encountered. If you
just enter items into the database, you will have your file names.
> Can we as a group decide on the naming convention so we don't accidentally
> duplicate a name (especially if we use numbers) and to readily identify the
> image as belonging to the specific institution?
This should not be a problem, since each institution will store the images
locally. Also, the database is assigning an institutional "prefix" (which
you can set in the defaults table) which will go at the beginning of all
of the filenames you create.
> When Daniel Pitti was here last week to help me straighten out my EAD to
> XML to HTML display issues, among others, he suggested
>
> <DAOGRP>
> <DAOLOC ROLE="thumbnail" SHOW="embed" ACTUATE="auto"
> ENTITYREF="item5219.t"></DAOLOC>
> <DAOLOC ROLE="reference" SHOW="new" ACTUATE="user"
> ENTITYREF="item5219.r"></DAOLOC>
> </DAOGRP>
>
> The "item5219" is the individual image number in the Pennsylvania Bridges
> Collection that was assigned during processing.
>
> When we played with the beta version of the MOA2 Access database, we used
> H-PABRI1.01, H-PABRI1.02, etc. as our numbering to identify the image as
>
> H = Historical Collections and Labor Archives (unit within Special
> Collections)
> PABRI= Pennsylvania Bridges Collection (name of collection)
> 1 = part of collection being scanned
> .01 = photograph being scanned
>
> For the Order of Railway Conductors and Brakemen we used H-ORDER1.01, etc.
> H = Historical Collections and Labor Archives (unit within Special
> Collections)
> ORDER = Order of Railway Conductors and Brakemen
> 1 = Minute Book A
> .01 = November 16, 1884 (date of minutes)
>
> We need to decide now what we're going to call the images so we can
> properly embed the DAO in the EAD finding aid. Would it make sense to use
> a combination of the two systems to identify the image (H-PABRI1.5219.t) or
> use the simpler naming convention (H-PABRI1.01.t )? Should we be adding
> PSt at the beginning to identify the image as belonging to Penn State
> (PSt-H-PABRI1.01.t)?
>
> What are y'all doing out there?
Again, the database is handling derivative naming, so you don't have to.
We went over this at the NY meeting, and if you look at your databases,
you will see how it works.
> On another note, in the EAD finding aid are you expanding your container
> list to include a listing of the images being scanned (e.g., for Minute
> Book A, listing all the dates), or only listing the broader approach
> (Minute Book A, Minute Book B, etc.) and then pointing to an artificial
> table of contents for the user to navigate through the scanned minutes by
> date?
As a matter of fact, some of us Berkeleyites met last week to try and
decide what we will do here. We will be "exploding" the container list at
the item level, but to what degree, I don't know. Certainly it will be
very cluttered if we start including all of the information we are storing
in the database. But is it enought to show the first page of a multipage
item, and then link from there to the tool for that particular object?
This is what I am leaning towards.
> It seems that listing all the dates in the EAD is way too
> cluttering. But then where will the "table of contents" (the EBIND "page
> turner" reside)?
The "page turner" (not EBIND, but a Java tools of Rick's devising) lives
on the Berkeley Sunsite.
I hope this gives you enough of an answer.
Merrilee