Forgot your password?
Views
XML Hebrew Syntax
Document Actions
February 26th
Click on the "subscribe" tab above to be notified of all changes to this page and its children by email!
What Andi has done:
Action Points
- Kirk creates schema and sample trees, using the Tree Editor.
- Andi will generate grammar from schema and trees
- Andi will create a grammar for converting Morph (while preserving all data) into a Book/Chapter/Verse/Morpheme hierarchy, with the Morpheme elements displayed as flat leaves in TreeEditor. Do only Genesis. Andi will create the input file this afternoon (Feb 26). Completed.
The information in the Morpheme element/leaf would be:
- surface form in Michigan encoding
- parse string
- lemma
- English gloss of the lemma: use the glosses from the
csv9:Analysistable.
- Kirk will write grammar rules with conditions and restrictions in terms of morphology, POS, context and cantillation.
February 27, 2005
Today Kirk is using the test data file generated yesterday by Andi and using TreeEditor to create trees. Two kinds of feeback are being given: (1) GUI and functionality and (2) what kind of analysis.
- ten or fifteen trees were created with various alternative analyses given
- a little work done on the schema (and a lot of reading on XML schema!)
- comments on both the usabilility of TreeEditor and on the experimental data model
February 28, 2005
Snow is predicted to begin at 6 a.m. with a total accumulation of up to 10 inches...But that didn't deter Andi and Kirk! They worked straight through until 5 p.m. and slipped and slided home...
The discussion today focused around the nature of the grammar needed to be submitted to the parser. A data model is becoming clearer. Finally, a Roadmap was created for future collaboration.
March 1, 2005
More snow flurries today...
Our discussions included how George Yeager's fine Hebrew parser (written in prolog) could be utilized to leverage our grammar. Andi generated a new version of the lexer to use the disjunctive feature of the cantillation system to group morphemes into phrases. It has become clear that we will be writing, in effect, an entirely new kind of Hebrew syntax. For all practical purposes, the documentation of the parser grammar will be a book on Hebrew Syntax. And we had better do so carefully, otherwise we will forget what we did and why!
Concluding Action Points
- Kirk will work on analysing trees, detailing the syntax data model (syntax schema), xslt Morph2xml filter (including schema)
- Andi improve the
lexerfor analysing trees (only gn1 has been completed), modify parser for handling incomplete trees, tree editor feature requests, finish the pure cantillation tree grammar
