24: Representation and Analysis

Wed 6 Jun, 11:00-12:30
Session 24: Representation and Analysis

Collex: facets, folksonomy, and fashioning the remixable web
Bethany Nowviskie

Expressing Complex Associations in Medieval Historical Documents: The Henry III Fine Rolls Project
Arianna Ciula, Paul Spence, José Miguel Vieira, Gautier Poupeau

Reading Tools, or Text Analysis Tools as Objects of Interpretation
Stéfan Sinclair, Geoffrey Rockwell


B. Nowviskie gave an engaging overview of NINES (est’d 2003): they seek to provide tools that many scholars can use and benefit from, in the wake of the current academic publishing crisis. [More on that in many places, e.g., here.] One question for NINES was whether the energy already being put into C19 scholarly work could be channeled into collaboration that looks and feels like scholarly communication to people used to print-based processes. Thus, NINES added a peer review structure; helped to apply standards-based tools, such as Collex (blog); encouraged scholars to tap local IT resources more fully; and linked scholars together into a supportive as well as critically analytical community.

Collex is also a convergence: people can upload RDF describing their resources and test objects in a sandbox before submitting them for peer review by members of the NINES board; a Solr-based facet system with Lucene search and “more like this” contextual links; an “integrated collections mechanism” that allows annotation; and an in-progress exhibit builder.

JSTOR, CUP, UVa Press, and Project Muse are partnering with Collex to allow “discovery” of relevant resources within the system. Next stop: OCLC.

Tools are needed that foreground informed subjectivity and reflect the expectations of more traditional colleagues yet also push them to become more familiar with tools enabling digital manipulation.

(D. Sewell commented that Collex may be intimidating to people not already intimate with Web 2.0 tech. Nowviskie responded that feedback so far is anecdotal; a formal usability study is coming. Another person asked about the distinction between formal markup [I think the part enabling facets, in context] and folksonomy [user-applied tags]—they look rather similar to a user; at the same time, the people who apply approved markup are often users as well, so informal tags may sometimes be applied in well-informed ways, so why are the two buckets split at all? [Missed Nowviskie’s reply.])

(Here is a review of Collex.)


P. Spence and A. Ciula spoke.

The Fine Rolls project (said Spence) is a collaboration of the Nat’l Archives (UK) and King’s College London. The first version of the website was launched in Apr 2007, with content from 1216 to 1248; an index is included for people, places, and [something—Spence is going rather quickly]. Each roll –> TEI P4 doc includes physical structure, structure of the calendar and its content, and semantic content (names of individuals, locations, etc.) One result of the project’s socio-historical approach is a set of revelations: 3439 men / 401 women, ~4062 locations, 1157 topic-subjects. The project uses authority lists and associations (isDaughterOf, e.g.) to track personal names yet privileges the orthographic forms used in the texts; they use RDF/OWL rather than MADS.

(Ciula; v. quiet) The XML markup was used to populate the ontology (via XSLT from the TEI markup); till Henry III was mentioned in a text, he did not “exist” within the project, and docs participate in the set of associations. Thus, a doc with a certain ID “refersTo” a person. The project built their framework so that both content and associations could be used by other undertakings. The encoded associations enable clarification or emendation for the indices without making the intervention visible in the text. Protégé-OWL = Java-based ontology editor enabling continual refinement. Ciula showed visualization shots demonstrating how the associations work for persons.

The TEI markup can benefit from the ontological associations as well, but should they be kept synchronized? It’s difficult from a maintenance perspective, and the ontological work is effectively annotative.

(Some of the talk’s content is available here. S. Bauman asked about P5 use, which the project plans to adopt once P5 is more stable, perhaps during the anticipated second half of funding that’d complete Henry’s reign (d. 1272). O. Eide cautioned against feeding too much back into the TEI, partly because it’d bloat things and partly because it might limit other kinds of ontological research.)


“A dialogue, with no slides.” 😛

They began with an allegory of (basically) Prometheus, who brought his new tech to his dean instead of his people; his dean told him to get the work peer-reviewed. What would one lose from adopting new tools? What if we read tools as text—studied their instructions without trying to implement them? Or, what if we played instead of trying to read (first | during)?

It’s not tools’ fault, but with personal computers (instead of mainframes) we gained access yet lost the communal strengths of lab work, such as ongoing discussions of method.

[Much harder to take notes on than a formal talk! This omits much.]