aaight... I get yer point there, BUT you see, what do you do when an artists changes it name... forget it, that was a bad choice... anyway... you see, in one of my fields of interests, you got dogs... see, dogs can change name, not just the calling name, but I mean completely change it all. second, they change apperance with growth. So to keep track on "stuff" you need to make a system work for you, not the other way around... Hence, all in the db... but no way someone would be * enought to put ... you know what.. NOone would EVER put a single mp3 file in a db ;) lol. There we agree. My problem originated with the planning... or rather lack of. Now? I think it works flawlessly so why fix it On Monday 06 November 2006 23:00, Richard Lynch wrote: > On Sat, November 4, 2006 5:38 pm, Børge Holen wrote: > > either you end up with a had as method of grouping them together, > > moreover you can have thousands of small files inside one dir with an > > id name > > to it, and yes the last one, thousands of directories with one file > > inside... > > Speaking as a guy who has 65,000+ mp3s "on-line" (though only a > fraction of them available to the general public) I'd go crazy if they > were all in the DB, not to mention that my webhost would kill me... > > But I'm not dumb enough to put them all in one directory either. > > "foobar.mp3" goes in "/f/o/foobar.mp3" > > Actually, which *drive* it is on is determined by the date the audio > was recorded, as I have about a Terabyte available spread across 4 > cheap IDE drives, and just change a simple include file when one of > the drives is nearly full to start using up the next one. > > So it really turnes into one of these: > "drive1/f/o/foobar.mp3" > "drive2/f/o/foobar.mp3" > "drive3/f/o/foobar.mp3" > . > . > . > based on what date the live performance occurred, which is in my > meta-data, which I need to get for the on-the-fly ID3 tags anyway. [*] > > It's all very crude and shoestring budget, but it works. > > Well, except until they turned off the A/C in the office last summer, > and then spilled coffee grounds all over the box... :-( > > But I found an old computer in the closet that I had found in a > dumpster and threw the hard drives in that, and it works. 495 MHz > seems enough for my audio server. :-) > > I'm finishing up a process of copying the files to a "real" host so > I'll have a crude 2-tier fail-over. > > Anyway, you have to plan this out with some idea of what scale and > scope you are dealing with to avoid insanity, but cramming binary data > into the DB seems like the least attractive choice to me personally. > > [*] It's a shame the MP3 players all seem to ignore my nifty JPEG ID3 > data I'm pre-pending to the audio streams. You have to download the > files just to see the artist photo. Sigh. > > PS > Feel free to give a listen if you like accoustic music: > http://uncommonground.com/radio_hifi.m3u > http://uncommonground.com/radio_lofi.m3u > (hifi, lofi, respectively, obviously) > > iTunes Podcast version is in "beta" if you're interested in being a > beta-tester... > Same audio, just in RSS/XML format. -- --- Børge Kennel Arivene http://www.arivene.net --- -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
