Re: question for the Arista/Transmageddon groups

On Tue, 2010-09-28 at 21:01 -0700, Jeff Cobb wrote:

Hi Jeff,

> (..) dvdmetabase (..) sends a request to my back-end server (..)

Out of curiosity - is the database public? If so, what is the license of
the database data on the server? Is it possible to download a snapshot
of the database for experimentation on a local machine, for example?
Would the data license allow people to take the data and set up their
own server elsewhere (for non-commercial purposes) if you lost interest,
turned evil or somesuch?

 Cheers
  -Tim

Hi Tim; I am so glad you asked that. I have been turning that question
over and over in my mind and would like your (or anyones) thoughts on
it.

First to get the 'turn evil' question out of the way, I have either
GPLed or LGPLed all of the source, as much as any of it were original.
As for the metadata (what is stored on the server), my desire is to
license it in such a way that no one can ever lock-up the data and
ideas. I have worked for huge companies that tended to lock up every new
idea or invention, even if it only peripherally (if at all) related to
the day job. Since Sony has/had so many fingers in so many pies, you got
to hang on to almost anything. While other reasons are equally valid,
this is one reason why I quit and moved to a more narrowly-defined
company. If I came up with a good original idea I wanted the freedom to
keep it free so that others might gain use from it and overall, the
actual item itself being free also meant that I could continue to work
on and extend my own work. It has happened so many times in the past
where I came up with something I found interesting and mostly because it
encouraged more and different improvements. Thus if anyone on this list
(or otherwise) suggested a free license that does what I want as far as
being free goes, I will more than likely choose that one. So, anyone can
use it, others are free to extend it without fear of reprisal.

That said, something the missus and I have discussed at length over the
years is the licensing of the data itself. Can anyone suggest that one
company or the other owns any kind of license for this data? I mean,
CDDB et all do something similar and they are still OK (but got sold a
number of times). The information in the database from the US side, I or
the missus tracked down the proper track names, etc and it was our
sweat-equity that got it this far. Not surprisingly, the same disc in
England would invariably get a unique hash, even if the tracks were all
the same thing due to PAL making for a different dataset.

So what license is the data on the back-end? I don't know but putting a
restrictive license on this information would not be a cool thing and
the irony of putting a restrictive license on this could end up leaving
me with an invention of my own that I could not even use. Bull. I guess
at the end of the day, it should be free as in freedom and free as in
beer because I see so many real-world uses for this that I would rather
an open-source company keep it rather than keeping up with it myself
only to have it taken away. Thus if the information is free enough to
use, polish, extend, etc that is what I want.

Does this make sense? I hope I have answered your question; if not, let
me know and we can work it out. This was a very good question though.
One last thing: since the missus and I have entered every disc in the
database (with the exception of any PAL titles which my friend in
England has entered/worked with) it is up to us/me what the fate of the
data is.

Finally, in answer to your snapshot question it was always intended to
encourage experimentation. I do not have a clean/automated way to make a
snapshot of the database but have done it for others by hand. Say you
wanted to do what you say, machine A is your front-end (in my exact case
where you run EZRip from) and machine B (at 1.2.3.4) is the local
database, separated by cat-5. You could do the following:
1. Get a snapshot of the DB (its really quite small)
2. Install sqlite, python, etc on machine B.
3. On machine B, edit the config file to reflect machine B and finally 
4. Start the SQLite daemon (if needed)
5. On the front-end machine, edit the config file to point to a local
machine/point and then roughly, run the ezrip program to rip or to test,
insert a DVD and run chkdisc.py which should do the hash on the disc,
submit it to the back-end for a lookup. Not terribly sexy but if you are
running it locally, either use a database address of machine B (1.2.3.4)
or 'localhost' if on the same machine. This whole thing has been quite
effective for testing front-end, back-end integration so by default it
is aimed at the experimenter and therefore free.  Part of the value that
I want to leave in the hands of the FOSS world is the actual content
itself so while you *could* set up your own private server, there isn't
a large reason to do so since the primary value here is the usefulness
of the data. More, if the public in general were allowed to add to the
database, it would belong to everyone.

Thoughts?