Text Selection Taskforce

Membership
The EEBO-Text Creation Partnership Text Selection Task Force will meet at the CLIR offices in Washington DC 30-31 March 2000. Members of the task force include:

• Gay Dannelly (co-chair), Ohio State University
• Thomas Hill, Vassar College
• Laura Janover, ProQuest Information and Learning
• Austin McLean, ProQuest Information and Learning
• Bill McPheron, Stanford University
• Maggie Powell, Yale University
• Mark Sandler (co-chair), University of Michigan
• Daniel Traister, University of Pennsylvania
• John Tuck, University of Oxford

Charge
Review relevant background information on selection of a subset of the image corpus for full text conversion, including:

1. Statements and discussion documents.
2. Input of faculty queried in advance of the meeting.
3. Focus group information gathered by ProQuest Information and Learning.

Having completed this review, and any suggestions and arguments have been clarified, the Task Force will draft recommendations to the Board on the following:

1. The primary underlying approach to selection of the text corpus: chronological, item-by-item, random sampling, stratified sampling, or a combination of the above. If item-by-item selection is to be employed, specify the criteria. If sampling is to be employed, specify a method.
2. Proposed exclusions to the converted corpus: non-English language materials, materials other than English and Latin, dialects of English, multiple editions, serials, almanacs, dictionaries, government publications, sermons, and heavily illustrated works.
3. After reviewing the work of the DTD Task Force, implications for selection based on physical attributes of works, such as fonts, scripts, tight bindings, and poor press strikes.
4. Subject areas, if any, to be emphasized or excluded.
5. The proposal to set aside 10% of partner contributions to convert specific titles recommended by individual partners.
6. The proposal to avoid converting titles available in widely-distributed proprietary collections.
7. Offer the Board any further advice regarding the ultimate shape of the text corpus.