msokolov commented on issue #862: LUCENE-8971: Enable constructing 
JapaneseTokenizer with custom dictio…
URL: https://github.com/apache/lucene-solr/pull/862#issuecomment-530494478
 
 
   > Should it be marked experimental then ? The fact that we ship a single 
dictionary within the jar also ensures that it is built from the same version 
but this change breaks this assumption. What kind of compatibility are we 
expecting here ? Should we require users to rebuild binary dictionary on each 
minor version ?
   
   Yes, these are good questions. I think experimental makes sense for this 
given that we are not providing detailed documentation and really only experts 
with knowledge of NLP will ever use this. With expert features there is no 
compatibility guarantee, so I think that rebuilding with each version would be 
the recommended policy. I would think users would be well-advised to rebuild 
whenever they build their software, treating the Kuromoji dictionary as a 
binary artifact produced from (textual dictionary) source code. Does that make 
sense?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to