>If I remember correctly I think the purpose of a twp phase database was
>to avoid having to do a full rescan when you have changed some setting
>that usually requires it. There was a lot bigger need for this with the
>new schema approach as you may want to re-configure the schema without
>having to do a full rescan. I'm pretty sure it wouldn't make
>incremental rescans any faster. 
>
Yes, I agree that there would be benefits in a two phase DB.  Storing the raw 
file tag data (acting like a cache), would be useful for other situations, and 
breaks the problem up into two nice phases:

1. Read content from file system and store in a DB.
2. Read tags from that metadata DB and process into an optimised SBS DB.

It would probably make fault diagnosis easier.  eg. at the moment, due to 
diacritics being subsumed into artists with no diacritics, it isn't easy to 
find files with incorrect tags.  But if there was an interim DB, it would be 
easier to find bad tags (rather than searching tags in files).
_______________________________________________
beta mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/beta

Reply via email to