[mb-users] "The need for free and open music metadata", but MB not good enough

[Brian Schweitzer]

I have to agree with Mudcrow.  I admit I'm not familiar with David, but I do
find it rather bizarre, too, that he first decides that MusicBrainz is
"**Complex
model, incomplete and inaccurate content, difficult to maintain and update",
so his solution is the make an entirely new music database, also user
contributed, using a means quite similar to MusicBrainz to support itself -
then as to how he wants to actually pull it off?

"However, should people take to our implementation, we are going to need
some help. Maybe Jimmy Wales <http://blog.jimmywales.com/> can take it over
as part of the Wikimedia Foundation<http://wikimediafoundation.org/wiki/Home>.
It is using data from his baby after all. Maybe
Mozilla<http://www.mozilla.org/foundation/>or Music
Brainz <http://metabrainz.org/> can help and show how this can be managed.
We are open to and are actively seeking suggestions."

Umm, though I don't see MusicBrainz users as a group doing anything to block
them trying to do whatever they want, why would MusicBrainz actively help to
develop a new database that does exactly what MusicBrainz does and supports
itself almost exactly the same way MusicBrainz does?

David seems to have fallen into the same trap that seems common these days;
namely, the blind belief that "open for anyone to edit" immediately leads to
"best and most accurate data".  The one *can* lead to the other - as I
think, at least in some areas, we prove.  But how you get from the one to
the other...  that's where the magic stuff is.  (And much of it the issues
we debate frequently.)

"**Complex model" - well, yes.  David's model is based only on
    1. An artist, or
    2. A group, or
    3. A song, or
    4. An album or music collection
This sounds like MusicBrainz, circa v. 0.1.  No labels, no Advanced
Relationships, not much else - you might as well use Google Music, for the
amount of interesting/useful data that's lacking in his "non-complex"
structure.

"incomplete ... content" - Is this not true of any database, Wikipedia
included?  I know of several recordings that are not contained in *any*
music databases.  (Ones I've yet to add to MusicBrainz, so even we don't
have them.)  Is the better solution to a) create an entirely new database,
or b) add any missing releases you know of to an existing database?  (I
choose b, he seems to choose a...)

"...and inaccurate content" - Again, though he only says this about
MusicBrainz, it is also quite true across the board.  I don't like that this
is true, but again, how to balance inaccuracy entering the database against
complexity of editing (and thus disincentivizing editing at all) is one of
the main issues we frequently debate.  There is no magic solution.  But a
path to a solution would include more people working on fixing errors, not
attempting to fragment the userbase by starting yet another separate
database.

"difficult to maintain and update" - I'm not sure what he means by the
maintenance or updating of existing data - if it's entered correctly, it
shouldn't need updating.  Adding events, relationships, etc does imply some
opf this, but again, his structure doesn't even include such data, so I'm
not sure what he's talking about.

Don't get me wrong.  I do understand that any database - MusicBrainz
included - has a somewhat steep early learning curve.  But his comments
appear to be based on a very shallow look at each.  He claims AMG is highly
accurate, while we all know AMG is frequently incorrect, and very
US-biased.  According to him, Audioscrobbler data is private and cannot be
updated by third parties, while the reality is that they do have an API and
license for its use, and have added at least basic 3rd party input, if not
direct editing, with the voting on what the correct name for an artist is.
Wikipedia he sees as "highly accurate" - need I even discuss how wrong this
is?

One of my favorite quotes, just to show how the Wikipedia model breaks on
non-subjective data, such as music, from the Mozart discussion page:

"Also, for the benefit of the general users of Wikipedia, who do not really
understand these compositional technicalities, simplifying the List of
Violin Sonatas by Mozart by counting K.6 as Violin Sonata No. 1 may be a
good idea."

This idea was then followed by another editor, for all works (not just
violin) - attempts at reverting this clearly bork logic have been reverted
each time.  So Wikipedia now has their own work numbering for compositions,
without regard to any historical accuracy or original reason behind the
real/official work numbering.  :P

Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.musicbrainz.org/pipermail/musicbrainz-users/attachments/20080426/d89be8a2/attachment.htm