Agenda for November 6, 2015 ALTO board face-to-face meeting @ iPRES in Chapel Hill NC.

  1. Review action items [Frederick leads discussion]
  2. Review and adopt board membership terms, membership criteria, decision making and voting procedures, and recruiting new members (issue 17 and issue 20). [Jukka and Frederick lead discussion]
  3. Discuss and vote recently nominated on board members [Frederick leads discussion]
  4. Discuss reading order and text block writing mode change requests issues 12) and 18). [??? leads discussion.]
  5. Discuss OCR correction attributes change request issue 21). [Jean Pilippe leads discussion.]
  6. Discuss fragment identifier API for ALTO (issue 33). [Jean Philippe leads discussion.]
  7. Other business to be added ...

View all open action items here.

Attending members

Minutes

The ALTO board met face-to-face after the 2015 iPRES conference in Chapel Hill NC, USA, in Manning Hall on the University of North Carolina campus.

The board discussed issues pertaining to board membership, board member terms, and voting. The results of the discussion are summarized in the board membership criteria. Changes and revisions from the discussion at the meeting are in italics and prefaced with [PROPOSED]. We will discuss and finalize the proposed changes at the next board teleconference.

Discussion and changes to issue 22 have finished. Frederick promised to pester board members who haven't yet voted on it. Once voting is finished, the issue will be open for public comment (draft schema). The board hopes to publish the schema changes necessary for this issue along with any schema changes needed for another (simple) issue at the beginning of 2016.

Recently Jean Philippe added issue 33, an API for ALTO fragment identifiers. See issue comments for a summary of our discussion about the API.

In order not to publish a new schema with changes only for a single issue, the board considered all open issues and choose a "simple" issue which could be quickly resolved so that any necessary schema changes necessary for it could be combined with the changes made for issue 22. We choose issue 23, confidence value calculation. We decided that it is not possible to require a uniform way of calculating confidence values. At best we could ask ALTO producers to document their method for computing confidence values, perhaps in OCRProcessing. This would not require a schema change. We also decided to merge issue 23 with issue issue 13 since that issue is similar. We also recommended that OCR confidence calculations documentation be improved. We will discuss issue 23 further during our next teleconference.