I examined Jim Zwick’s Anti-Imperialism in the United States, 1898-1935. Overall, I thought the online archive was well-done for a private individual (and--I assume from the really annoying advertising--under funded effort). Mr. Zwick captured a significant amount of information regarding anti-imperialism movements and other associated efforts. However, he could incorporate a couple of simple tools to his site that would dramatically improve its usability.
In this blog, I will first discuss how I would improve the site to exploit the digital medium. Then I will discuss a project based on this site that uses the unique capabilities of New Media.
Mr. Zwick added each new source as a separate HTML page instead of using XML or some other database technology. I don’t think this is a critical fault, however. Mr. Zwick obviously did most of the work on the site, himself. What his site lacks in sophistication, it makes up for in its “brute force” approach--the mass of data collected. I believe that his in-site search engine significantly aids the user in searching through an otherwise unmanageable mass of data. However, I think that a couple of minor changes would dramatically ease searchability for his site.
First, I would incorporate an advanced search feature that uses the text within the document and metadata. Those who use the search engine within the site are likely to be disappointed if they are used to advanced search features in many archives and online journals. Though the search engine handles Boolean strings of data, it doesn’t handle them well. For example, I typed “McKinley and Philippines”—a search that will guarantee information overload—and I received hits that were not as relevant as I had expected. If Mr. Zwick incorporated some very basic metadata on his individual pages, then he could dramatically improve the accuracy of his search engine. I would recommend metadata that discusses date, provenance, subject, speaker, genre of argument, party affiliations of the authors, and other related documents.
Second, I would also incorporate metadata and ALT tags for the images. As the site is presently configured, the user must download the complete image before he knows what it is. Also, the in-site search engine does a poor job of capturing these images if the user does not know the date or cartoonist for each image. From the search page, the user can navigate to an index of all images, but this page has no thumbnail control or descriptions of the illustrations—again, searching becomes a “hit or miss” affair.
And third, this site screams for some innovative navigational features. For example, an interactive timeline, a few pages that serve as intellectual histories for selected arguments, or trees of correspondence would also provide excellent interfaces for managing all of the texts. At the very least, Mr. Zwick should incorporate a site map that shows visited locations. This site is so extensive that a reader can get lost very quickly.
Also, because Mr. Zwick spent the effort to make all of his text machine readable, he could take advantage of some of the other capabilities that this type of text provides. First, he does use hypertext links within the text to discuss specific people, places, or events. He could very easily add hover or rollover notes to provide a brief synopsis of these points without forcing the user to go to another page. Also, he could link specific documents with other related media of the time. For example, one of the editorials on his site was posted on the same page as one of his cartoons. Why not link them both? Also, he could incorporate some GIS to explain the locations within his texts. In many of the narratives I read, the author discussed ambush locations or other sites. However, I could not easily cross-list those sites in the text with map locations. A little GIS would go a long way to alleviate this problem.
Finally, I would improve his site by placing the actual scanned images of documents where the machine-readable conversion has forced a loss of information. I was particularly disappointed with his petitions. While the specific cases made by the petitions are important, the people who signed them can tell the historian much more. Unfortunately, OCR software can’t read handwriting, so Mr. Zwick was forced to just tell us how many people signed.
For my research project, I would build an interactive timeline and GIS support system that linked Mr. Zwick’s documents in time and space. From there, I would analyze the documents in his sites against the actual events on the ground to see if there was a correlation between specific actions and public outrage. From here, I could probably find how sensitive public opinion (and anti-imperialism) was to reported events. In other words, did success on the ground quell anti-imperialism opinion, and did failures inflame it? This project could also spawn other interesting studies. For example, how did the subject matter of political cartoons shift with public opinion? Did regional politics influence the illustrations. Where were the anti-imperialists located? What else was influencing them?
By linking documents in this site through GIS and other tools that use metadata, I believe that the researcher could find trends that discourse analysis about various items would not. While this project could be done without digital technologies, Mr. Zwick would have spend another lifetime correlating this data and determining trends. With the path I’ve discussed, he could dramatically improve the usability of the site while providing tools that could significantly aid historians of this period.
Posted by Stephen B. Sledge at November 7, 2004 07:55 PM