[X Newbies] Itunes...

[Alex]

On Friday, Mar 26, 2004, at 22:30 Canada/Eastern, Vincent Cayenne wrote:

> [...] The CDDB is a vast storehouse of information about many audio 
> discs. The information was freely contributed by many, then somehow 
> commercialised (not a complaint, I just don't know the details) and is 
> licensed by/as Gracenote. [...]

A succinct account of the story can be found at 
<http://www.ibiblio.org/tkan/software/cddb_wsj_12.31.01.pdf>, and it's 
endorsed by Ti Kan, who started it all. What the article doesn't cover 
is the fact that a highly significant fraction of the value of CDDB was 
due to volunteer contributors. In other words, those who reaped the 
profits of the sale of CDDB reaped the profits of unpaid labour.

On Friday, Mar 26, 2004, at 23:08 Canada/Eastern, Vincent Cayenne wrote:

> [...] Unless your tastes runs to a genre where many have contributed 
> info that is pooched <g>. But seriously, I get about one entry with 
> slightly screwy info about every 10 retrievals or so (artist/track 
> switched, wrong case, etc.) and a "not found" every 20 attempts or 
> thereabouts [...]

First, the database was originally designed for pop music; hence, it 
couldn't accommodate the kind of information required, say, for 
classical music. In the absence of a standard, each contributor used 
his own method of entering such data.

Second, the database was created chiefly by volunteers. Consequently, 
it has all the warts of such projects. There are the typos. Then, there 
are all those Anglophones (English-speakers) who don't know the 
(relatively simple) rules of title capitalization of their own 
language, and who, not content to mangle English track names, happily 
apply the same ignorance to non-English titles.

Third, there's the popularity problem. If you're a fan of Elvis Presley 
or Pink Floyd, you're one of a great many, and very likely someone has 
already entered the info in the database. But if your tastes run to, 
say, Tassos Chalkias or the octets of Georges Enesco, then you might 
very well be out of luck and have to type the info yourself.

Finally, there's the "magic number" problem. As Vincent pointed out, in 
most instances, Audio CDs do not contain any identifying info 
(incidentally, does anyone know if iTunes recognizes CD-Text?). 
Therefore, they are identified by a "magic number", computed on the 
basis of, among others, length and track number. By and large, that's 
adequate, but in some cases different CDs may have the same "magic 
number" and get confused. For instance, I'm a fan of classical 
Hindustani music; quite often, these CDs will have only two or even a 
single track, and confusion happens fairly frequently.

f