I've been playing around with MusicIP Mixer based on Chris' recommendation. After getting through the "Oh my word, this is sucking up all the processing power of this monstrous computer" while it indexes and "validates" your music, it does a really interesting job, finding similar songs across all sorts of genres, albums, artists. In trying to share this wonderful news with the world, I exported one of the playlists, but iTunes just dumps it out as a tab-delimited text file. There's not much I can do with a file like that on this machine; because Google Docs has gotten so good, I haven't bothered to dig out a copy of MS Office. So I threw it into Google Spreadsheets to see what would happen. Bastards auto-magically imported it into a perfectly formatted spreadsheet without that interim Excel step where you do the machine's job and say what it's delineated with, if there are headings, etc.
Unlike Google, MusicIP isn't perfect yet. It brings in more of the exact same artist/ album than I'd like. It has a slight case of feature-itis: I don't need to play the songs in it, I don't need another MP3 tagging solution (and I'm guessing most of the target audience already has an incredibly anal tagging plan* in place). And I'd like to be able to pass it multiple songs from different artists, but that's just me thinking as a person. How the hell does the machine know what x songs have it common? That's what it's supposed to be doing for me. Watch it take over my sonic life.
* Apologies to anyone with a non-music incredible anal tagging plan.