| >As a result, my count of known ABC sites is now 239, an increase of | >45 from a month ago. My tune index now contains 115097 | tunes, | | this is a great tool you have programmed ! | | Is it possible to have a top ten list ? :) | More seriously, have you a report of those website (with the | number of tunes) on your own site ?
Top ten what? Big sites? Titles searched for most often? I've occasionally posted statistics from the search bot. I usually also include a disclaimer about the actual significance of the numbers. Thus, I have one of the larger sites, but I occasionally mention that this is somewhat an accident (due in part to collecting from a number of mailing lists and mirroring other people's stuff). And I don't consider it all that important. I think the smaller sites are probably more important in the long run. Especially with a web search tool, where things are on the web isn't all that important. | >Part of the argument is the ongoing fear of ABC being Napstered. | I hope it will never be the case. Since folk-music is no longer | 'pop' (=popular) music, Abc will certainly remains 'clean'. Yeah; most of the online abc is long out of copyright. Of course, that won't necessarily stop the corporations. Intimidation has a long, dishonorable history, especially in what are supposedly "competetive" markets. It's pretty obvious that the recording industry is running scared, and rather than improving their product or lowering prices, they are attempting to outlaw the competition. The publishing industry understands the situation somewhat better, but it still hoping that the net can be turned into a way to restrict what used to be "fair use". And in music publishing, there's a long history of making vague copyright claims on public-domain material. One of the really useful things about abc is the ease of including historical notes. I've already read of a couple cases of a publisher challenging some online sheet music (abc or gif or whatever), and when told "That tune was published by so-and-so in 1763", the publisher silently slunk away and was never heard from again. This sort of thing is a blatant attempt to claim copyright by intimitation. Knowing exactly when something was first published is a good defence. So if at all possible, you should include historical notes in your tunes. A good place to get information is: http://www.ceolas.org/tunes/fc/ | Can we make something in order to help your bot find easier the | tunes we've transcribed ? Yeah. Send me the URL. (First look up your own tunes to see if I already know about them. ;-) Actually, there are simple ways of helping all sorts of such efforts. There are a lot of people now working on the problems of locating things on the Net. A lot of the work is in specialized searchers that know the terminology of a few subject areas. If you have any sort of music files online, chances are that several music searchers have found them. But they might not be doing a good job of analyzing them. If you're interested, you could try to spot them and offer to help. The main thing is to separate the material so that search programs don't have to dig through too much. Thus, if you have abc and postscript and gif and midi and mp3 music files, you might put them into four separate directories. Then my abc search bot won't have to grovel through the midi and mp3 files, a midi program won't have to examine the abc and gif files, etc. This isn't always convenient for whatever else you may be doing with them, of course, and I wouldn't worry about it that much. But this is a useful approach when it's consistent with the other uses. And it is useful for other purposes, too. If you really want to help such online efforts, you should probably also look around in your server's logs (if you can), and see if there are any searches of your non-html files. If so, you might want to try to identify the searchers and send them some email, asking if you can help them. This is particularly true if you have something like music files, which the big search sites can't handle well. Try locating abc tunes through google or yahoo if you want to learn about the problems. Even better, try finding postscript or pdf music files. You'll be disappointed. One fact about my tune finder that's not necessarily true of other such efforts is that it only returns single tunes. So the most efficient way for it to be used is to have one tune per file. With a big file, extracting a single tune means downloading the whole file and searching it for the tune. Again, this is just one program, and some people have perfectly good reasons to have one big abc file, so I don't push it much. My own collection is mostly single tunes or "medleys" of 3 or 4 tunes that fit on one page. But I do violate my own advice, and I have a few large abc files. One is in my mirror of the O'Neill's Project: http://trillian.mit.edu/~jc/music/book/oneills/ In this case, I decided that 100-tune files were appropriate. But the tunes are also there in single-tune files, to speed up access. (And this artificially increases my own tune count. ;-) There is one thing that I do that subtly encourages small files: If you ask for a title and there are a bunch of matches on tunes with identical titles, my index sorts them by file size, smallest one first. I did a bit of a study of requests, and found that people to tend to work their way down the list. So I lightened the load on this machine somewhat by sorting this way. OTOH, I'm seriously considering adding the ability to unpack zipped collections of tunes. There are several sites with a lot of tunes packed up this way. I haven't done it so far, mostly out of laziness. and I probably should. It would add a bit of cpu time, but not all that much, and it would make more abc tunes available. I would definitely put them at the bottom of the list, though. To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
