[02:22] Join: Caelian_ joined #corewars [02:26] MSG: Ping timeout: 240 seconds [02:56] MSG: Read error: Operation timed out [02:56] MSG: Read error: Operation timed out [02:57] MSG: Ping timeout: 240 seconds [02:58] Join: triple joined #corewars [02:59] Join: Caelian_ joined #corewars [03:09] Join: Mizcu joined #corewars [03:14] MSG: Ping timeout: 240 seconds [03:19] Join: Mizcu joined #corewars [04:07] MSG: Ping timeout: 240 seconds [04:07] Join: Caelian|w joined #corewars [05:03] MSG: Ping timeout: 240 seconds [05:03] Join: Caelian|w joined #corewars [05:37] MSG: Quit: Ex-Chat [08:00] Join: willvarfa joined #corewars [08:06] MSG: Quit: Ex-Chat [08:50] Join: willvarfa joined #corewars [08:52] MSG: Client Quit [08:57] Join: Caelian joined #corewars [09:29] :) [11:34] MSG: Quit: Ex-Chat [11:37] MSG: Ping timeout: 240 seconds [11:38] Join: Caelian|w joined #corewars [12:50] Join: Caelian joined #corewars [14:27] http://groups.google.com/group/comp.compression/browse_thread/thread/a87a6dd960ed1b68/c30f67bfdffbe48e#c30f67bfdffbe48e <- found ya ! [15:18] Join: Roy joined #corewars [15:18] Good find, I read that newsgroup too regularly ;) [15:21] MSG: Client Quit [15:23] damn, escaped me [17:05] Join: Caelian_ joined #corewars [17:08] MSG: Ping timeout: 240 seconds [17:48] Join: Roy joined #corewars [17:48] Sorry, I just come and go :) [17:49] cant blame [17:50] Join: fiveop joined #corewars [17:52] Getting cosy here, hi 5op [17:52] So you are still in the compression-algorithms Miz? [17:52] ive been wondering today about turning a piece of data into a radix tree, but that ends up a a problem how to limit the string at a end of an branch [17:52] good enough answer? [17:53] Hmm.. [17:53] No idea on how to visualize that :) [17:53] you know radix tree? [17:54] Err, I know the radix-sort [17:54] (quick google probably answers that faster than i..) [17:54] yeah :) [17:54] radix-sort is an n-tree where every node with only one child is unified with its child [17:54] http://en.wikipedia.org/wiki/Radix_tree [17:55] so one would end up with say, a -> b -> c -> defgh [17:55] (well, a => b => c -> defgh) [17:56] idea being that one reads couple characters in input, and then takes a string in library [17:57] You could store words in less space, so probably also data... IC [17:57] library being built on pre-existing data [17:59] question is deciding how for an string at end-of branch should be [17:59] for -> long [18:00] or even worse, if huge amount of data is reach, there will be so much data that there are no string-ends at first 4-5 levels of a tree [18:02] so in short, i am just thinking about PPM [18:03] single pre-seen characters can be encoded in under a byte (or size of literal) with Ari or even Huff coding [18:10] I still love the thought of making a range-coder with some (or multible) prediction algorithms underneath [18:11] afk [18:11] range encoder can be used like ari-coder [18:26] Yeah its basicly the same except a range-coder is faster and doesn't have a patent attached to it [18:29] i thing i have been thinking before is using say, lzw together to pack stuff that is not predicted [18:30] but then there is the problem that there are tree options for data; literal, literal packed, predicted [18:31] and using more than one bit per byte for describing seems bad [18:33] technically one could predict small strings from LZW library, which might be bad for hit-rating [18:37] lzw -> predictor wont work, predictor -> lzw.. only if predictor is made for speed rather than accuracy [19:45] MSG: [22:28] MSG: Quit: Lost terminal