Re: The competition (CDDB2)

Nick Lamb (njl98r nospam at ecs.soton.ac.uk)
Fri, 3 Dec 1999 00:51:08 +0000

On Thu, Dec 02, 1999 at 09:08:26AM +0100, Michael Kaiser wrote:
> I am getting close to the point where I convert from a flat-file-database to a
> rdbms and I'd like to join forces as much as possible. Are you (still)
> interested in a cooperation and how do you think we could work together more
> closely?

The major barrier (I have said this before, but I will say it again) to
close co-operation is that FreeDB has all the weaknesses of CDDB on the
data side. This isn't your fault (it's not even the fault of the current
owners of CDDB) but it's a tremendous technical weakness.

You cannot generate a valid CDindex identifier from the database info
in the FreeDB at the moment, and the CDDB-style identifiers are hopeless
(even FreeDB's relatively small collection contains about 1% collisions
already) for uniquely identifying CDs, especially singles.

Is it possible for you to get LEAD OUT data into FreeDB records as they
are updated? This requires support in (at least) submitting clients,
and probably in the server code too. With LEAD OUT we can transform
FreeDB records into CDindex ones.

It would be nice to have a proxy which accepts CDindex queries and
returns matches from CDindex (preference) or FreeDB otherwise, with
the user request and FreeDB match tied up to create new CDindex data,
I don't think the quality hit from this would be too awful, if the
duplicate merging code is working.

Do you have a list of stuff FreeDB has which would be "nice to have"
or essential for lossless conversion of FreeDB to CDindex?

Nick.