Re: CWR - Common Works Registration

Hi;

On Mon, May 05, 2014 at 03:29:32PM -0500, John DeSoi wrote:
> Hello,
>
> I'm new to MusicBrainz and have been evaluating it for possible use in a project.
>
> I notice that the works data is very thin in comparison to the number of recordings (over 13 million recordings, but less than 500 thousand works). It seems this could be improved a great deal by implementing an import for CWR - Common Works Registration. CWR is being used by publishers and societies around the world to register and exchange information about musical works.
>
> Here is the specification document:
>
> http://www.commonworksregistration.com/CisacPortal/downloadMyRessource.do?fileName=/home/ftpdata/FileRepository/IS-CRF/EN/IS-CRF08-0027R5_CWR_Functional_Specifications_2005_08_04_EN.doc&name=IS-CRF08-0027R5_CWR_Functional_Specifications_2005_08_04_EN.doc
>
> It would not be necessary to implement the entire specification, only the parts relevant to the MuicBrainz database.
>
> Would this be useful for the MusicBrainz project? If so, would it be acceptable to implement in a language other than Perl? My first choice would be PHP and I assume that could be easily called from Perl to process CWR files.
>

There's a couple places something like this could be hooked in, at
various timeframes. The first thing to note is that the community is
generally not fond of fully automated imports, at least not without very
comprehensive checking. There are, however, several routes to using this
data as a tool for editors to use in bringing in work data more
efficiently -- that is, not as a fully automated process, but ensuring
that editors adding the data don't need to type or manually copy-paste
the information.

In the longest term/most correct, we'd probably want to write tools to
import this style of data into geordi [0], which is intended to be our
general place for putting data from external sources for processing and
eventual import to MusicBrainz itself. It's not currently entirely
complete, however (and is in fact in the process of a relatively large
reworking, albeit on slight hold as we finish up our semiannual schema
change release).

More directly useful for you might be the interface geordi itself will
use for providing the data to MusicBrainz for editors, that being
so-called "seeding" of our various editing pages ([1] for
release-oriented fields -- somewhat less documented, and generally done
via GET rather than POST, for other forms). These interfaces are used by
most of the userscripts that ease the process of importing data from
other sites (e.g. discogs, bandcamp, jamendo); most of them are for
releases, but http://userscripts.org/scripts/show/94676 is at least one
that provides importing tools for works (in this case, from JASRAC).

In this situation, you, presumably, would create a small tool that
creates an appropriate url and/or form for submission of the works with
included data. One thing to note is that a patch coming next release
will also add tools to the work add/edit page directly for editing
relationships, which are especially important for works (where they're
used to denote composers, lyricists, etc., as well as to link to
recordings); since the next release is a schema change that code isn't
on our usual beta server, but it is on http://ianmcorvidae.mbsandbox.org
(which doesn't touch the production database, of course, either). Then
the remaining piece would be making the tool available and encouraging
people to use it.

More nuts-and-bolts talk, you'd presumably want/need to ensure accurate
lookups of linked properties like the artists and recordings, and some
tools for avoiding duplicates.

And of course, since it'd be a separate project entirely, you could do
it in whatever language you desire :)

For the longer-term possibility of putting it in a tool like geordi, I'm
curious where we'd source data in this format; is it possible to
download this data directly (and for fairly unrestricted use) somewhere,
would it need to come through some sort of negotiated official channel,
or something in-between?

Hopefully my description is helpful/informative, and of course feel free
to ask further questions if you have them!

[0]: https://github.com/metabrainz/geordi/tree/big-refactor
[1]: http://musicbrainz.org/doc/Development/Release_Editor_Seeding

> Thanks,
>
> John DeSoi, Ph.D.
>
>
> _______________________________________________
> MusicBrainz-devel mailing list
> [email protected]
> http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-devel

--
Ian

_______________________________________________
MusicBrainz-devel mailing list
[email protected]
http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-devel

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)

iEYEARECAAYFAlNoKisACgkQq+0avXExp2+hCQCfaZ2dk5uiXQ3uDy/EZFoK9d1x
th8AoM612E4NbQQUa3WbUKjbY+/Vte68
=wpjQ
-----END PGP SIGNATURE-----