RE: [edict-jmdict] database schema

[Jim Breen <Jim.Breen@**********************>]

[Stuart McGraw (RE: [edict-jmdict] database schema) writes:]
>> Also, as a general observation, not a comment on what you said
>> about the indexes....
>> while a loader program will be used frequently while evaluating
>> and developing the database, after the database is in operation, 
>> the loader script will be only of historical interest.  (There 
>> are better ways of backing up the database than relying on re-
>> loading from a assorted text and xml files.)  So advanced load 
>> optimizations will probably have a short lifetime.

I think this is correct.

[..]

>> - I don't see any distinction between "entities" (which you have 
>>   all in one table) and other keywords like the language and 
>>   dialect (which you have in separate tables).   The difference
>>   between them seems to be only in a choice Jim made for their
>>   representation in xml.  Had he defined entities for the language
>>   and dialect keywords, would they belong in the entities table?
>>   What would have changed about the information they convey?

None of this is set in stone. The original JMdict DTD was me
teaching myself XML many years ago. Now I'd probably put
many things which are currently entities in as attributes, as is
the case with the kanjidic2 structure.

I'm not sure how the entity vs attribute decision impacts on the
database design. Perhaps not at all.

Jim

-- 
Jim Breen                                http://www.csse.monash.edu.au/~jwb/
Clayton School of Information Technology,               Tel: +61 3 9905 9554
Monash University, VIC 3800, Australia                  Fax: +61 3 9905 5146
(Monash Provider No. 00008C)                ジム・ブリーン@モナシュ大学