[pdb-discuss] Some things to think about at tonight's meeting

Mon Jan 15 17:28:16 UTC 2007

Here are some suggestions for things to think about in relation to 
tonight's meeting.

1. What we plan to build. In particular do we want to build our own db 
of work metadata or contribute to an existing one? Originally I  had 
thought we would need to construct our own as nothing appropriate 
(structured + open licensed) already existed. However I spent quite a 
bit of this afternoon looking again at musicbrainz:

   http://musicbrainz.org/

When I last looked at this in detail (as it must have been a year ago) 
they appeared to be using CC by-nc-sa which rendered their data 
non-open. However since then they seem to have changed to making all the 
core data public domain and only keeping the CC by-nc-sa for a 
restricted set of add-on data. Thus, is seems to me, it would make a lot 
of sense for us to avoid reinventing the wheel by creating our own 
metadata database and instead focus on:

   1. Contributing data on *old* works to musicbrainz (old data is what 
we are interested from a public domain point of view)
   2. Developing software and algorithms to determine which 
works/performances are in the public domain

There are some drawbacks to using musicbrainz of course. For example, 
(AFAICT) they don't always draw a clean distinction between 'authors' 
and 'performers' (they do have a composer category in their Advanced 
Relationships section though). However stuff like this seems very minor 
compared to the benefits.

~~ If we go down this route then things we can work on: ~~

2. Contributing data. This breaks down into several parts:

   1. Entering data by hand (e.g. by looking up dates in wikipedia)
   2. Getting hold of full datasets either 'by hand' (i.e. finding them) 
or robotically (e.g. from library of congress)
   3. Extracting the data we want from the datasets we acquire (for 
example from the composer list we were given or from the BBC data)
   4. Once we have structured data uploading up to whatever storage 
system we are using (musicbrainz or our own)

3. Developing a front-end to show current list of public domain works, 
questionable works (i.e. status unclear etc etc).

4. Developing a project website (based at 
http://www.publicdomainworks.net/ or whatever other url we choose). I 
suggest we use wordpress for this and move the current demo wiki-based 
site to alpha.publicdomainworks.net.

Regards,

Rufus