Re: [gnomoradio-devel] An ed2k key/hash database?

[Alexandre Van de Sande]

anyway, i think using their program to music seems a little like
glueing a flashlight to a head cap with silver tape. It's easier but
we want to develop a real hemlet with a light on, not make a kludge
from ready pieces.

Again, there is a whole field of study called collaborative filtering.
It wasn't gnomoradio nor irate who invented it. I am currently very
interested in studying it to develop a open all players colaborative
filter for music.

I see that has been done already, can anyone give me a clue about how
irate or gnomoradios algorithm currently work?

Alexandre

p.s. don't try to understand their logic. copyrighters have thei own
reason. i disagree mostly, but they just think it's the easiest way to
get more money.

On Wed, 15 Sep 2004 11:34:25 +0200, daniele <address@hidden> wrote:
> Uff... u r right! They "accept" very happily new code... but do not
> release their!
> Bad. Worst.
> And stupid: they base their earnings upon how many people join the
> community. More people join, more people may choose to upgrade, more
> webmasters will pay (and pay more) to sponsor their web site... don't
> understand! If anyone could do its own non-harmful script or
> implementation of SU in this or that utility, the traffic and the number
> of stumblers would only increase in number and quality!
> Going to ask them something about their reasons.
> daniele
> 
> Alexandre Van de Sande ha scritto:
> 
> 
> 
> >very interesting daniele.
> >
> >It is a interesting idea, but too simple. what bothers me is something
> >on stumble upon homepage:
> >
> >"Founded April 2002, StumbleUpon has developed a patent-pending
> >community-based websurfing system"
> >
> >Stumble upon is patented, wicth means they are not wanting to anyone
> >to freely use their code. Best would be to develop a independent open
> >coded, copyleft program specifically for our needs, that could be
> >adopted in many softwares, like
> >
> >1-colaborative filtering for music (like gradio and irate)
> >2-colaborative filtering for wepages (like stumble upon)
> >3- colaborative filtering for anything somebody thinks about it (news,
> >videos, friends, messages, blogs, anything)
> >
> >But thanks for showing me stumble upon, i didn't know it.
> >
> >Alexandre
> >
> >Don't care for poor language grammars, care for communication
> >
> >
> 
> 

 daniele <address@hidden> wrote

Thinking about how could it be built a common database of CC art for all
possible applications, I had that idea.
The database contains only ed2k (for example) links to the files, and
descriptions about them, providing facilities to search and know more
about licensing, artists, donations, comments made from others etc.
A web server generates a page for each entry in the database, displaying
all info, containing all necessary links etc. From the hash and the ed2k
code of a file, it is possible for the user (or a client he uses) to
build the correct address for the web page on the server corresponding
to the shared file.
And the rating mechanism?
Nothing simpler! The rating is provided by an external tool, like
stumble upon. Infact, each song is represented by a web page and any web
page can be rated in the stumble upon community, and it can be opened a
discussion on the page. Than, "stumbling" from the browser in the
preferred category, will show the description page of the songs with the
ed2k links. Integrating the stumble engine into an external app (like
gnomoradio) will generate a radio playlists /throw that mechanism)
according to the selected genres.

Comments/idea?
daniele
ps: sorry for my bad english!!!