JavaNLP Meeting Notes 10/21/03

1) Updates:

Dan:
- checked in new parser stuff, but parser not working on regular text
input 

Galen:
- Checked in NormalizedCounter for normalized/smoothed views
  - immutable
  - some extra fields (numberOfKeys, reservedMass)

Huy:
- eliminated dependencies in util package 

Jenny:
- in classify.old deleted LinkedList.java and Node.java, and have other
code just use util.Collections LinkedList class.  (Note: current one mixes
iterator functionality into LinkedList class.) 

Roger: 
- implemented GeneralizedCounter
- fixed DummyTokenizer documentation 
- adapting JFlex/StreamTokenizer to new Tokenizer architecture
  - implementing JFlexTokenizer class that takes Lexer parameter

2) Dev branches in CVS:

Why we might want development branches:
- allow CVS to bridge the transfer of file between your local machines
(for IntelliJ) and the nlp/db machines
- allow people to check-in code that doesn't compile so you have backups,
revision history, sharing, etc.

Action:
- Create dev branch in cvs with personal folders.
- Can check out personal folders using "cvs co dev/myfolder -d target_dir"

3) Document/DocumentReader:

- Need DocumentReader that iterates over files in a directory (in addition to
  specifying delimiters) needed.
- Jenny will research TreeBank, give comparative evaluation, and then
  we'll decide what to do about Tree/Document stuff.