Sunday, December 8, 2013

Composite text search and score generation in CDLI

Composite text search and score generation in CDLI [via email]
I am pleased to announce, after the recent implementation of a new search and aggregation tool keyed to cylinder seal occurrences in CDLI text transliterations, a second, similar utility for searching and displaying composite and witness entries that document the occasionally quite rich Mesopotamian tradition of text copying and transmission. Supported by the programming skills of UCLA CS graduate students Eunice Yuh-Jie Chen and Henry Li, we have begun entering to CDLI catalogue all “Q” identifiers of such texts, following the numbering system of Oracc’s Qcat, and, in line with the system implemented for seals management, have made these Q numbers searchable at our slightly revamped search page. Thus now a search for the Gudea text (RIME 3/ =) Q000911 results in, currently but ever growing, 399 hits, led by Dan Foxvog’s transliteration and translation of the composite version and followed by entries documenting 398 physical artifacts. While the Q numbers will be less familiar to users, this system, like its cousin seals ID field, allows a casual viewing of some random witness text, for instance a recently tagged cone fragment in the University of Durham collection, to expand, by clicking on its corresponding catalogue composite ID Q000915, to a full view of the 96 CDLI entries that are associated with that text, led again by Foxvog’s treatment of an idealized composite version. Indeed, we limit all translations of such texts to these composite entries.

The corresponding tagging of composite text transliterations in CDLI for the purpose of generating score versions of each text, that we delighted in producing as ambitious students of Assyriology, will follow along at a slower pace. <> contains the first several dozen such completed score (“Partitur”) versions, primarily now of the easier to tag RIME inscriptions, but in time the much more involved scores of scholarly texts will be added. Given the great numbers of exemplars of royal inscriptions catalogued by the RIM project, we have decided to limit Q-tagging of transliterations to those that are better documented than by their simple entry in RIM witness catalogues, usually with existing image files, with the exception of texts with no such documentation. In these cases, the first otherwise minimally registered transliteration will serve as a single witness until further documentation is available to us. Thus while the score of Gudea 37  seems imposing, it comprises just 250 of the 1365 registered witnesses to this inscription. These seals-and-composites tools are, for us, primarily technical means to improve the standardization and reliability of our existing transliterations, but should also be of general interest to colleagues who seek to check the fidelity of the various published or circulated forms of Partitur apparatus that represent some of our most treasured research tools.

Bob Englund

No comments:

Post a Comment