MusicBrainz Non-Profit White Paper

MusicBrainz Non-Profit White Paper
by Robert Kaye
February 2003

1. Introduction

MusicBrainz aims to create a music information commons where the community
creates and maintains a public database of information about music. This
music metadata will enable non-ambiguous communication about music, and will
allow the Internet community to discover new music without any of the bias
introduced by marketing departments of the recording industry.

The MusicBrainz project has been around since the fall of 1998 (previously the
CD Index), and is now gathering more support from the community and
partnering companies. In order to give MusicBrainz some legal muscle and to
ensure the future availability of the dataset, it is proposed that MB be
incorporated in California as a non-profit corporation.

However, creating and running a non-profit corporation costs money, and with
limited resources, MusicBrainz will depend on donations from the community
and industry sponsors to elevate it to the next step.

2. MusicBrainz Today

The first version of MusicBrainz, which nears completion during the first
quarter of 2003, focuses on creating an open database of basic music metadata
which can be used for identifying audio CDs and digital audio tracks (MP3,
Ogg/Vorbis, WAV, etc.). MusicBrainz is comprised of three separate components
which all work together to enable users to semi-automatically identify music
and apply clean metadata tags to their music collection:

MB Web site: The MusicBrainz web site allows anyone on the net to search,
browse, and maintain the community metadatabase. The web site users
(moderators) can add new metadata to the site, edit or correct existing
metadata, and delete incorrect metadata via the web-based moderation system.

MB web service/client library: All of the MusicBrainz data is available to the
public via the RDF-based web service. A web service client can search for and
request information about any artist, album or track in the database. A
client library released under the LGPL is available for developers who would
like to support MusicBrainz in their application. This client library
abstracts out the details of interacting with the MusicBrainz web service,
and allows a client developer to add metadata lookup to their applications in
a short period of time.

MB Tagger: This 32 bit Windows application (similar applications with support
for other platforms are also in development) takes an end-user's collection
of MP3, WAV and Ogg/Vorbis files, generates an acoustic fingerprint (TRM Id)
for each track and, using the fingerprint, looks up the track metadata. If
the main server does not have the metadata available, the application guides
the user through the process of entering the missing information into
MusicBrainz so that future users may benefit from the new metadata. After the
proper metadata has been downloaded/entered, new metadata tags are written to
the user's audio files.


The basic metadata includes a list of artists and artist aliases (e.g.
alter-ego names, alternate band names and common abbreviations) and for each
artist a list of albums and the tracks for each album. MusicBrainz assigns
each artist, album and track a unique identifier, which can be used to refer
to a particular artist/album/track without having to deal with the semantics
of correct spelling and conflicting names in the database.

These identifiers provide the Internet community with a means to establish a
meaningful computer-based dialog about music. This unambiguous dialog is
enabled by an RDF based web service interface and presents the first baby
steps towards the "Semantic Web", where computers can carry on a meaningful
discussion without involving human beings. The RDF used in the web service
uses portions of the Dublin Core and is documented on the MusicBrainz site.
MusicBrainz encourages others to use the RDF in other future music
applications to enable a host of new applications and features that are not
possible today.

For instance, it is not possible today to exchange a playlist with a friend,
since your friend may not have the same files that you do; even if your
friend does, the files may be located in a different location on the hard
drive. Using MusicBrainz, a user can create a playlist that consists solely
of MusicBrainz track identifiers, and then send that playlist to their
friend. Their friend will be able to feed the playlist to their
MusicBrainz-enabled audio player and then have the player match up the
available tracks. If some of the tracks are not available in the collection,
the player could go out to music sites such as EMusic.com, MusicNet or
Pressplay to download the missing tracks. The MusicBrainz identifiers allow
future audio applications to carry on unambiguous conversations about music
and to enable a whole new set of features for music enjoyment and music
discovery.

The MusicBrainz dataset has been created and maintained by its user base of
over 2000 volunteers. Since its inception as the CD Index in the fall of
1998, and the consequent renaming to MusicBrainz in the fall of 2000, the
database has seen more than 160,000 additions and changes (moderations) to
the database. Even without any promotion of the site, and all of the software
just now emerging from a beta state, the dataset is growing and improving in
quality. To see the latest statistics on MusicBrainz, please visit:
http://musicbrainz.org/stats.html.

MusicBrainz's human moderation approach encourages participation in the data
maintenance process and thus yields higher quality data, since many eyes will
spot even the smallest mistakes. Active moderation, concise technology for
identifying music and a carefully designed database allows MusicBrainz to
collect data with greater accuracy than services like GraceNote. The
GraceNote service suffers from an overwhelming number of errors and duplicate
entries in their database, and without a focus to reduce duplicates and to
correct errors in the database, they cannot compete with MusicBrainz in the
long run.

Furthermore, GraceNote charges serious amounts of money for severely
restricted access to its data. FreeDB, the free alternative to GraceNote, has
not created any new technology to advance the state of the project. FreeDB's
goal is to provide a service that is free and backward compatible to the old
GraceNote/CDDB service. This gives MusicBrainz the advantage to create the
first well-edited, highly structured and comprehensive music encyclopedia on
the net.

Once the TRM (acoustic fingerprint) and audio CD based music identification
portions of MusicBrainz have been completed, the service is poised for a
significant increase in the number of users contributing to and using
MusicBrainz. This will provide a powerful alternative resource for
non-commercial music developers and a very low cost alternative for
commercial music services and channels.

3. MusicBrainz Tomorrow

The basic metadata framework that the first generation of MusicBrainz puts
into place will enable more comprehensive and subjective metadata to be added
to the community metadatabase. A few possible additions include:

Reviews/biographies/ratings: Unlike the rest of the existing MusicBrainz
dataset, artist/album reviews/ratings and artist biographies and are not
factual metadata, and thus they will require a different approach in
collecting and maintaining. However, this subjective metadata may present the
most significant revenue source for MusicBrainz. (see below for details)

Music Discovery: The advanced music classification from above will allow
MusicBrainz users to browse the available genres and discover new music as
they find genres that describe their own musical tastes. Combining the music
classification with user-contributed information about their own musical
collections will enable MusicBrainz to offer collaborative filtering services
to its users.
Advanced Music Classification: Today's music classification systems leave a
lot to be desired, since music classification is a highly subjective task,
and few subjective systems have been developed to date. However, MusicBrainz
can harness the power of many users to create a representative classification
system that will evolve over time as musical genres evolve. Using data
collected from thousands of users will enable MusicBrainz to statistically
infer Genre Curves for artists and albums.

Detailed Music Information: MusicBrainz will expand to cover more information
about music such as artist web pages, official fan web pages, detailed
support for classical music (e.g. composer, opus number, orchestra,
conductor, etc.), and any other relevant pieces of information that will make
MusicBrainz into a comprehensive music encyclopedia.

Music Genealogy: MusicBrainz may keep track of which
artists/performers/engineers contributed to a piece of music, and when these
contributions took place. Combining this contribution data with data on how
artists influenced each other will create a genealogy of modern music.
Imagine being able to track Britney Spears back to Beethoven!

These are just a few of the possible future directions of MusicBrainz. The
actual directions will be heavily influenced by the MusicBrainz
partners/sponsors to create a mutually beneficial relationship between
MusicBrainz and its partners and sponsors.

4. MusicBrainz Licenses

MusicBrainz is devoted to using the right licenses for the right job and thus
the GPL (GNU's General Public License) is used for the server software and
the LGPL (GNU's Lesser General Public License) for the client library. The
use of the LGPL allows even closed source applications to use the client
library to access the MusicBrainz server.

The overall goal is to remove as many obstacles to accessing the MusicBrainz
dataset as possible and to foster the inclusion of MusicBrainz technology in
third party applications. To support this goal, MusicBrainz makes the dataset
available to the public by placing portions of the dataset into the Public
Domain and releasing other portions under Creative Commons'
Attribution-NonCommercial-ShareAlike License 1.0:

Core data: The core data is comprised of the artist, artist alias, album, and
track information, as well as the CD Index identifiers, and TRM identifiers.
All of this data is released into the Public Domain.

Derived data: The derived data consists of artist, album and track text
indexes, as well as moderation and voting information, which is released
under the Attribution-NonCommercial-ShareAlike License from the Creative
Commons.


Subjective data: In the future MusicBrainz will collect artist biographies,
album reviews, music ratings, and other non-factual data and also release
them under the Attribution-NonCommercial-ShareAlike License.


To some people the use of the Public Domain for the core data may come as a
surprise. However, the United States Supreme Court decided that facts are not
copyrightable and all of our core data is essentially comprised of facts.
This limitation, combined with the desire to have commercial enterprises use
the MusicBrainz core data to extend the reach of this data, makes the Public
Domain a perfect choice.

5. MusicBrainz and Commercial Enterprises
Even though MusicBrainz is an open source and open data project, MusicBrainz
actively encourages companies to participate in the MusicBrainz community.
The availability of the core dataset in the Public Domain encourages
companies to work with and link to the MusicBrainz dataset without having to
navigate a complex maze of license requirements.

MusicBrainz is not hostile towards commercial (for-profit) corporations! On
the contrary -- MusicBrainz will only reach its full potential if commercial
corporations use the dataset and encourage their customers to participate in
the MusicBrainz community. Any and all corporations around the globe are
encouraged to use the MusicBrainz core dataset to establish meaningful and
non-ambiguous conversations about music.

The derived and subjective data components in MusicBrainz are licensed under
the Creative Commons Attribution-NonCommercial-ShareAlike License, which
prohibits the use of the data in a commercial setting. However, MusicBrainz
will make commercial licenses to the data available to companies that wish to
use the data in a commercial setting. The income from these license
agreements will provide MusicBrainz with the needed revenue to ensure that
the dataset continues to evolve and remains available to the public.

However, many companies are skeptical about using open source software because
there is no one to call (or hold responsible) should the software fail. Open
data projects like MusicBrainz are in a similar position -- what if the data
is wrong? Or not in the database at all? The answer to this lies in the
MusicBrainz community -- the community is comprised of individual
contributors who work hard to enter and correct the data in the system. The
MusicBrainz server software also enforces a peer review system, under which
users must review and approve changes made by other users. The peer review
system combined with the motivation, expertise and pride of its contributors
will ensure that the data in MusicBrainz will be comprehensive and reasonably
correct.

Only reasonably correct? No one can guarantee that all the data in a database
is correct. Not even the commercial companies that provide metadata services
can give this assurance. The MusicBrainz community will respond to problems
found in the database and fix mistakes faster than any commercial company
with paid contributors can, since the MusicBrainz community is global and is
never closed for business. Furthermore, the community is more supportive of
MusicBrainz than of other commercial services due to its open nature.

Another area corporations are skeptical about is the issue of service
reliability. The MusicBrainz servers have always lived in professional
colocation facilities with excellent connections to the Internet, and even
though there has not been a legal corporation watching over the servers for
the first four years of its life, MusicBrainz has had only a handful of minor
service interruptions.

In the future, MusicBrainz plans to create a network of mirror servers that
will mirror the dataset across the globe. Any corporations that would like to
work with MusicBrainz, but would prefer to handle their own servers for
reliability and added load balancing, will be welcome to operate their own
MusicBrainz mirror server. This option leaves all the service reliability
concerns in the hands of the corporation.


6. MusicBrainz Non-Profit Corporation

In order to ensure that the MusicBrainz dataset will continue to exist and
continue to be available to the public, a tax-exempt non-profit corporation
(503.c.3) should be created. This non-profit should adopt a set of bylaws
which will state that MusicBrainz will make all metadata created by the
MusicBrainz community available to anyone who wishes to download the data.
The MusicBrainz corporation should consider itself the guardian of the
MusicBrainz dataset and its community, and should take the necessary actions
to ensure that MusicBrainz can continue its mission.

The MusicBrainz non-profit should strive to become self sufficient over the
course of 2-3 years. To achieve this independence, it should pursue the
following possible revenue streams:

Contributions from the community: Users of the MusicBrainz Tagger will greatly
benefit from the project by having the tagger automatically clean up the
metadata present in a user's collection. For-profit companies charge for this
service, and MusicBrainz should ask users for a $10 contribution for the
service of cleaning up the metadata.

Google style ad-words program: As MusicBrainz gains more users, it will be
possible to offer an ad-words program similar to the one pioneered by Google.
Third parties will be able to purchase small and unobtrusive advertisements
that will be shown on artist/album pages.

License artist/album reviews and biographies: When MusicBrainz provides the
infrastructure to collect and manage album/artist reviews and biographies, it
will ask the authors of these works to assign the copyright to MusicBrainz.
These reviews and biographies will then be made available to the public under
the Creative Commons Attribution-NonCommercial-ShareAlike License.
Furthermore, as this collection of reviews and biographies becomes
comprehensive, MusicBrainz will offer a commercial license to this content
for use in commercial applications and web sites.

Provide MusicBrainz dataset services: As corporations switch away from
proprietary music metadata services, MusicBrainz will gain a larger user
base. However, since MusicBrainz is community funded it will be unable to
provide the bandwidth for millions of users to access the dataset. Large
commercial customers will be encouraged to setup their own MusicBrainz mirror
servers to handle the load of their own customers. However, some commercial
customers will not want to deal with this in-house and would rather contract
out these hosting and integration services. MusicBrainz will be available for
hire to carry out the hosting and integration of the dataset on behalf of
corporations. In the same spirit, if commercial customers would like to have
a dedicated support staff for addressing problems with the service or data,
MusicBrainz will also be able to provide these services.

The above revenue streams will take some time to develop, but over time
MusicBrainz will strive to grow its revenue and become self sufficient.
Should MusicBrainz find itself in a position of having excess revenue (where
a for-profit company would pay a dividend), it will offer grants or awards to
open source/open data/music projects and their developers.

MusicBrainz has never had anything to hide, and all of its business has been
visible to the public. The finances are transparent and all discussions are
carried out in a public forum. With this approach, MusicBrainz will attempt
to create a new kind of non-profit corporation that can continue to hold the
trust of its community.

Conclusion

Community feedback about the MusicBrainz project has been overwhelmingly
positive; now is the right time to take MusicBrainz to the next level and
create a non-profit corporation. If you believe that MusicBrainz has the
power to make a difference, please consider contributing money to
MusicBrainz. While we are looking for sponsors to contribute larger
donations, we welcome any donations. Anything helps to move the project
forward and keep it alive.

--
`The moroccans with the carpets
seem like saints
but they're salesman'