[vorbis] Obtaining tag-independent track uniqueness?

[Greg Wooledge]

Tom Wadzinski (orca_twadzins at yahoo.com) wrote:

> Open Nap, and perhaps Napster, use a "hash the non-tag portion" strategy
> (see http://opennap.sourceforge.net/napster.txt search in the text for
> hash)

Interesting reading.  Note the disclaimer at the top of it (reformatted):

  Disclaimer: The following information was gathered by analyzing the
  protocol between the linux nap client and may not resemble the official
  Windows client protocol.

This is a reverse-engineering document.  Section "3." (MD5) describes
the utter uselessness of Napster's md5 hash -- it's not even consistent
between different versions of the official Napster products.

The last time I used OpenNap was a couple years ago.  At that time,
there was no use for the md5 hash except *possibly* file integrity
validation (and I doubt it even did that much).  There was definitely
no multiple-source download.

> and in Bitzi (see
> http://bitzi.com/bboard/message?message_id=5119&forum_id=1667 search in
> text for "audio part hash" ). 

OK.

> I'm guessing that this is exactly how some of the current p2p programs
> with swarming work that it is okay if different users have different
> tags, because the songs get pieced together by swarming the non-tagging
> content and a new tag is applied to resulting file.  Audiogalaxy appears
> to work this way.

Hmm... whose tag wins?  Is it chosen randomly?  Does id3v2 trump id3v1,
or do you get one of each?


-- 
Greg Wooledge                  |   "Truth belongs to everybody."
greg at wooledge.org              |    - The Red Hot Chili Peppers
http://wooledge.org/~greg/     |


-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/pgp-signature
Size: 241 bytes
Desc: not available
Url : http://lists.xiph.org/pipermail/vorbis/attachments/20020313/d2724b13/part.pgp