Several items discussed in today's code walkthrough

[tomerm]

• runtimeHome - explain what is it used for
• sub - remove this property if it is not used
• tokenization - rename to "preprocessing"
• typeToks - remove, assume "tagger" value is the only one which is supported
• remove all server properties (servStop, servPort, servSource ..)
• taggerPath vs. rtTaggerPath - why do we need both ? Semion to follow up on that.
• w2vDim - must be consistent with actual w2v model (in case latter was created off line) . Clarify that in the comment.
• log file - write error messages into log.err and informative messages into log.inf
• runFor - fix values in comment since code works only with lower case

[tomerm]

@matanzuckerman FYI.

[semion1956]

• options runtimeHome and sub removed;
• Process (and section) tokenization renamed to preprocess. Now P() should be used in requests instead of T() ;
• typeToks and all options, belonging to server, removed;
• taggerPath refer to the jar, used by Preprocess (it performs all preprocessing tasks), just as rtTaggerPath used by DataLoader and in runtime (it performs tokenization only);
• the same w2vDim option used by WordEmbedding process, which creates W2V model and by Model prosess, which loads this W2V model. If W2V model was created "off line", w2vDim should be updated manually;
• no need in log.err, because any error causes the end of pipe's working.
• value of runFor can be typed in lower or in upper case.