Commentary on
Vol. XXI, No. 2
Posted 22 September 2008
|
Commentary on |
|
Posted 22 September 2008 |
This comment was made by Harrison Eiteljorg, II, and posted at the same time as the Newsletter article, September 22, 2008.
Any database structure can be critiqued by an outsider. Therefore, I think it is necessary to be careful when commenting on the Dunhuang database structure as shown in Ms. Jones' review (an image of the downloadable PDF showing the structure of the data tables).
However, Susan Jones and I looked at the structure -- and its expression in the data searches on the web site -- and came to rather different views during the course of Ms. Jones' work on the review. I was particularly concerned when I realized that the semi-prepared searches offered on the site were designed so that a user could seek items that were considered to be textiles under the heading of "Type of Artifact" and under the heading "Form."
When I saw that, I worried about the apparent confusion in terminology; so I looked more carefully at the organization and content of the data tables. The organizational chart is very impressive, and I was especially pleased to see a concordance for terminology in the various languages at issue. I believe that shows a level of fore-thought that is both unusual and commendable. What I could not find, however, seemed to me to be as important as what I could: no sense of what an archaeological artifact is. To be more precise, I would expect to find a data category for materials (ceramics, chipped stone, worked stone, hide, shell, etc.), a separate one for more specific material (glazed ware, obsidian, horse, cowrie), something about function (tool, vessel, ornament), along with something more detailed about function (awl, cooking pot, bracelet), chemical analyses if applicable, other tests if applicable, and additional tables as appropriate for objects that fit into corpora that have been well studied so as to be parts of larger typologies.
Looking at the Dunhuang tables, I found evidence for none of the above. The best I could find was a single entry labeled materials value but with no link to any other table where that might be made more meaningful. Dimensions are supplied in one table, but they seem not to be the kinds of dimensions one would expect for archaeological objects (diameter for a pot, for instance, does not appear to be there) but a kind of generic list. The entry for color is apparently an alphanumeric one, but there are no links to another table; so I assume this is just a description by someone unnamed (a process my own partial color-blindness makes me see as improper). I believe this is the natural result of the database designer using a single table to contain virtually all the basic information about a given object, as found here. For instance, in the Items Table one finds Paper Texture, Paper Layers, Paper Thickness, largest panel height, and largest panel width. However, only the single entry for materials value can be found. I can only hope that I have missed something or that the organization chart is not complete or . . . . Otherwise, it seems to me that the database is woefully inadequate for archaeological materials. (There is a Catalogue Table, but its link to the Catalogue Entry Table leads ultimately to the Texts Table; so this is apparently not the place for objects.
Susan Jones has examined the data far more carefully than I, and her opinion is very different from my own. On the other hand, my own examination of the structure of the data tables has been as complete as the database organizational diagram makes possible, and I am concerned that the data tables as currently conceived are not up to the task at hand.
LINK TO PRIMARY ARTICLE, Web Site Review: The International Dunhuang Project, by Susan C. Jones
Table of Contents for all CSA Newsletter issues on the
Web