tag:blogger.com,1999:blog-6555947.post114642031228508320..comments2014-01-12T10:46:48.153-07:00Comments on The Geomblog: Inadvertent Hashing and CDDBSuresh Venkatasubramanianhttps://plus.google.com/112165457714968997350noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-6555947.post-1146631045947530892006-05-02T22:37:00.000-06:002006-05-02T22:37:00.000-06:00"so once there are 10^10 CDs or so, a collision be..."so once there are 10^10 CDs or so, a collision becomes likely. "<BR/><BR/>NOT EVEN THAT. I have several hundred CDs and I run into collisions about 1 in 100 times. <BR/><BR/><A></A><A></A>Posted by<A><B> </B></A>AnonymousAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1146519509094549302006-05-01T15:38:00.000-06:002006-05-01T15:38:00.000-06:00CDDB (or Briklin) might not need a mathematician, ...CDDB (or Briklin) might not need a mathematician, but a calculator would help. 100^10 = 10^20, not 10^11. However, even if you start with this assumption (which is dubious, since track times come from a lower entropy distribution), the "birthday paradox" means that the probability of a collision is roughly 10^20/(#cds^2), so once there are 10^10 CDs or so, a collision becomes likely.  <BR/><BR/><A></A><A></A>Posted by<A><B> </B></A>David ApplegateAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1146512729025249812006-05-01T13:45:00.000-06:002006-05-01T13:45:00.000-06:00well you have to remember that CDDB is not being h...well you have to remember that CDDB is not being hacked. if your only goal is to get a decent hash function, it's a lot easier than trying to get a hash function that can't be hacked. <BR/><BR/>Moreover, mathematicians don't own the copyright on good ideas :) <BR/><BR/><A></A><A></A>Posted by<A><B> </B></A>SureshSureshhttp://www.blogger.com/profile/15898357513326041822noreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1146495800320491972006-05-01T09:03:00.000-06:002006-05-01T09:03:00.000-06:00Interesting how the simplest, silliest solutions e...Interesting how the simplest, silliest solutions end up being as good as what 12 mathematicians locked up in a room for a month would come up with.<BR/><BR/>I'm just feeling cynical today. <BR/><BR/><A></A><A></A>Posted by<A><B> </B></A><A HREF="http://www.daniel-lemire.com/" REL="nofollow" TITLE="lemire at acm dot org">Daniel Lemire</A>Daniel Lemirehttp://www.blogger.com/profile/17350105204859937215noreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1146436741219345742006-04-30T16:39:00.000-06:002006-04-30T16:39:00.000-06:00Is it CDDB or FreeDB that they use? Original CDDB ...Is it CDDB or FreeDB that they use? Original CDDB screwed a lot of people over when they became commercial. More information in the Wikipedia entry: <A HREF="http://en.wikipedia.org/wiki/Cddb" REL="nofollow">CDDB Wikipedia entry</A> <BR/><BR/>MusicBrainz is a very cool alternative with a different approach: <BR/><A HREF="http://en.wikipedia.org/wiki/MusicBrainz" REL="nofollow">MusicBrainz Wikipedia entry</A><BR/><A HREF="http://musicbrainz.org/" REL="nofollow">http://musicbrainz.org/</A> <BR/><BR/><A></A><A></A>Posted by<A><B> </B></A>DAnonymousnoreply@blogger.com