Watermark Descriptions (IPH 3.1.3-3.1.5)

[Virginia Tech]

Tomas Stohr made some comments in his last letter which we did not respond
to in our first reply, but which seem to us important points for us all to
consider. We refer to his comments B and C. First, regarding Comment B, on
coded watermark descriptions.  (We will send a second note regarding
Comment C, on digital camera facsimiles, a little later.)

At  9:46 AM 11/25/97 -0800, Ing. Tomas Stohr wrote (COMMENT B):

> B) Difficulties with IPH code: Recently I went through the "Catalogue of
>Watermarks in Italian Printed Maps" by David Woodward, University Chicago
>Press 1996. Even though a great effort was made in this book to show the
>IPH code for each watermark fascimile, unfortunately there are several
>inconsistencies among these assignments. Many of the watermarks shown,
>are surrounded by a circle under a six pointed star but in the
>identification one can find a large variety of codes:
>   in # 16 to 22              U1[J5/4]
>   in # 35, 36                [U1-J5/4]
>   in # 91/92                 [U1]-J5/4
>   in # 107 to 110            [U1]-J5/3
>   in # 121to123,and 126to130 <f:U1{b:J5/4}>
>   in # 150,158to176,214/215
>        221/222 and 238to242  <f:U1{t:J5/4}>
>   in # 146                   <U1>-{t:J5/4}
> and I would probabily have written:         <f:U1>-{t:J5/4}
>    I am not criticizing the publication on italian maps, but I want
>enfasize how difficult it is to find an agreement even on a simple
>pattern. I think there is a need of elaborating or reconsidering the
>coding system, or at least produce an extense users manual.

You will see in the model pages which we set up that we have not yet
addressed the question of how to describe the watermarks, and whether the
IPH system will work for a WWW-based watermark archive.  We simply
reproduced the IPH fields as they presently exist, and entered a simple set
of codes in fields 3.1.3 (Motif) and 3.1.4 (Class). (You can see this on
the sample description at the following URL:

http://www.bates.edu/Faculty/wmarchive/wm-initiative/Watermark/athphigen003,
1a-w.html

We, personally, find the IPH codes very difficult to get used to, ambiguous
when applied to more complex watermarks and likely to introduce errors as
different readers of descriptions interpret in various ways the codes
others have written (as Tomas' example demonstrates very clearly).  It
seems to us that we might do well to explore other alternatives and how
they might be implemented in a WWW-based archive system. In the Commentary
on the IPH Standard which we compiled as we worked on the model for a
WWW-based system of distributed archives, we therefore wrote the following
in the commentary on sections 3.1.3 and 3.1.4:

     In our model, we would like to use
     straightforward, standardized language
     and eliminate the codes. In our model
     watermark description, however, we have
     for the present retained, for discussion
     purposes, the IPH code.

We do have some initial thoughts on the methodological issue.

When the IPH project was started, there was no possibility of ready
availability on line of visual images, so the codes were devised to
describe *everything* in such a way that it could be searchable.  We now
have the advantage of having images as well as descriptions on line.  So we
ask you:  is there any continuing need for the watermark description to
describe *everything* -- every element and aspect of the watermark?
Wouldn't the presence of online images associated with the database
eliminate the need for such detailed codes as the IPH system constructed?
In light of the WWW's expanded graphical capacities, should the purpose of
the watermark description simply be to enable us to do searches that narrow
the field to a reasonable degree preparatory to visual study, so that we
don't have to download too many images over the wires?

So we have before us several related questions about how to handle
watermark descriptions, and you, the participants on this discussion group,
will hopefully think of still others:

Should we abandon codes?

How much and what kinds of information should we include in the watermark
description (IPH 3.1.3 - 3.1.5)?

What should be in the glossary of standardized descriptive terms (whether
coded or not)?

What can/should we do in the design of this database to assure the
effective use of standardized language?


Bob Allison <rallison@bates.edu>      Jim Hart <jhart@bates.edu>


Robert W. Allison
Dept. of Philosophy & Religion and
Chair, Classical & Medieval Studies
Bates College,
Lewiston, Maine, USA 04240

E-MAIL:         rallison@bates.edu
TEL:            (207) 786-6307
FAX:            (207) 786-6123