Construction Arabic WordNet in Parallel with an Ontology

1
Database Design
Item Offset Item name (u) Item type Source
name Gloss (u) Author Date
Link name Item 1 Item 2 Author Date
(hypernym, subclass etc.)
(synset, term etc.)
Item
(SUMO, PWN1.5, etc.)
Item
Word (u) Frequency Corpus name
Form type Form (u)
(plural, root etc.)
word
v 1.00
2
Example
283454739 car synset PWN A device Adam 9/2/05
equiv 283454739 485747383 Adam 9/2/05
485747383 Automobile term SUMO A
device Adam 9/2/05
Item
Item
automobile 345 Brown
Plural, inf. autos
(plural, root etc.)
word
3
Process (proposal)

• One XML file owned by each group
• Additions ok, uploaded each week
• Need tools to help with merge and conflict
  resolution
• Day 15 every month everyone sends XML export to
  Horacio
• Horacio runs conflict tool
• Flurry of email to resolve them
• Horacio exports XML and sends to everyone who
  then update their local copies
• Process has to happen quickly so export, resolve,
  merge is atomic

4
Tools Needed

• XML import and export to standard mySQL database
  format
• Interface from everyones tools to common mySQL
  DB (optional, long term)
• But lexicography (MEANING) tools will be modified
  and used in common between Manchester and
  Barcelona

5
Tasks

• Adam defines XML schema
• Adam implements import/export from XML to Sigma
• Bill implements import/export from XML to MEANING
• Christiane implements import/export from XML to
  Princeton tools
• Bill and Horacio install Sigma locally
• Horacio modifies MEANING to call Sigma when a
  SUMO term is clicked on
• Horacio implements conflict checker/resolution
  tool
• For example, XML exports have a link from AWN to
  a SUMO term that is deleted or renamed between
  XML file exports