Attached is an email which I received which I think makes some good
points.
--
___________________________________________________________________________
Ian Clarke
http://www.dcs.ed.ac.uk/~iic
"A subversive is anyone who can out-argue their government"
--------------FCD385ED7DCA7BC6C5430E9D
Content-Type: message/rfc822
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Return-Path: <robin nospam at acm.org>
Received: from postbox.dai.ed.ac.uk (postbox.dai.ed.ac.uk [129.215.41.196])
by muck.dcs.ed.ac.uk with ESMTP id MAA01131
for <iic nospam at dcs.ed.ac.uk>; Tue, 9 Mar 1999 12:55:30 GMT
From: robin nospam at acm.org
Received: from ro.nu (robin nospam at the-clock.bc.nu [194.168.151.226])
by postbox.dai.ed.ac.uk (8.8.7/8.8.7) with ESMTP id MAA28698
for <ianc nospam at dai.ed.ac.uk>; Tue, 9 Mar 1999 12:55:27 GMT
Received: (from robin nospam at localhost)
by ro.nu (8.8.5/8.8.5) id MAA09699
for ianc nospam at dai.ed.ac.uk; Tue, 9 Mar 1999 12:55:21 GMT
Date: Tue, 9 Mar 1999 12:55:21 GMT
Message-Id: <199903091255.MAA09699 nospam at ro.nu>
To: ianc nospam at dai.ed.ac.uk
Subject: CDDB
You wrote:
> This CDDB thing really makes me angry. The database that is used
> by these people was built up by ordinary users of programs like
> QCD who wanted to help out other users of the system by entering CD
> details. The CDDB people are now trying to sell the product of that
> good will. It is cheeky at best, and downright immoral at worst.
I quite agree. The problem is having a central database which is
susceptible to single point of failure and to administrative revolutions
like this.
> I second the call for a free CDDB-a-like which is protected from
> potential future exploitation by a GPL-like licence.
The issue is the database rather then the protocol or software, but
I agree that the whole cddb system needs a successor which is better
protected. There is plenty of scope for improving the existing
cddb protocol anyway; I always thought it was terrible.
Basically there are three things to sort out: what can you use as
the unique signature of a CD, what fields to store in the database,
what protocol to use to get at the database. IMHO cddb gets all three
wrong: the unique signature includes the arbitrary classifaction
of the music, which can't be computed, the fields stored are
very minimal, with vague and widely ignore instructions on how to
format the free text values in some of the fields, and the
protocol is overly-heavy TCP.
The replacement system should have a totally self-contained algorithm for
computing the unique disk ID---something simple and easy like taking the
MD5 checksum of the table of contents. It would be nice if the database
fields were totally relational, so links between composers, performers,
tracks and so on could be maintained, but I'm not sure it would be
workable to implement this in a small and simple client. But at least
there should be specific field labels for the most common information.
Finally, the basic protocol should be UDP for speed. Having email and
web methods is also a good idea.
A network of servers can exchange update information and make requests of
each other, so no one server has to shoulder the whole load. Something
vaguely like the DNS system would be appropriate.
> I am happy to come up with an interface spec (email ianc nospam at dai.ed.ac.uk)
> if anyone has a nice server and some programming time!
I'd love to help too. I don't think that I could cope with the world
descending to query a 50Mbyte database, but I don't suppose I have to
worry about that just yet! I can certainly come up with the processing,
disk space and connectivity for development. If we can come up with a
way to share the load, we make it easier to distribute widely (since
no one site has to cope with everything) and more robust (since any
connected part ofthe network will continue to function.
> Alternatively, would
> it be possible to use some kind of clean-room method to "re-discover"
> the CDDB interface protocol, and circumvent these idiots.
This doesn't help recover the data in their database. Still, I should
think that a good many writers of cd playing software would be
willing to switch to a new format, particularly if it offered some
facilty advantages over cddb (in addition to being guaranteed free).
I wonder if the existing database is worth preserving anyway. There
are many mistakes in it, and it would be hard to split the free-text
information fields in to something more computer-friendly. If we change
the CD unique ID algorithm, it is also hard to match up entries.
Anyway, there is obviously lots to talk about. I can easily set up a
Majordomo mailing list to conduct further discussions if you have had
enough replies to warrant it. In fact, I'd have set it up already but
I can't think of a catchy name:-)
Robin.
--
R.M.O'Leary <robin nospam at acm.org> +44 7010 7070 44, PO Box 20, Swansea SA2 8YB, UK
--------------FCD385ED7DCA7BC6C5430E9D--