00:00:00 --- log: started retro/13.01.05 04:43:03 --- join: obobo_ (~chatzilla@dyn-dsl-pt-76-75-103-192.nexicom.net) joined #retro 04:44:34 --- quit: obobo (Ping timeout: 245 seconds) 04:44:40 --- nick: obobo_ -> obobo 04:57:22 --- join: ncv (~quassel@89.35.216.197) joined #retro 04:57:22 --- quit: ncv (Changing host) 04:57:22 --- join: ncv (~quassel@unaffiliated/neceve) joined #retro 05:35:40 lp:~crc-x/retro-language/hashed-dictionary is the branch where I'm now experimenting with adding a hash field to the dictionary headers 05:39:56 http://bazaar.launchpad.net/~crc-x/retro-language/hashed-dictionary/revision/424 contains code to generate hashes and add them to the dictionary (no conflicts in global dictionary) 06:44:49 possible strategy to deal with collisions: check hashes; if they match, then compare names to confirm 06:45:27 makes sense to me 06:45:36 it's still a huge time savings over checking each word 06:45:51 also... 06:45:55 well... hrm 06:46:22 I was going to say you could re-order the dictionary to put the most frequently used words first. 06:47:22 but the dictionary is laid down in a particular order based on dependencies. So you'd have to re-arrange the links after the fact to do it that way. 06:47:57 also you'd have to know which ones are used the most :) 06:48:29 nm. that's completely over-optimizing anyway. 08:03:26 http://rx-core.org/retroImage_withHashes and http://sprunge.us/DJeF for some code that modifies the search functionality to use the hashes. Still slow though :( 08:16:25 this also seems to break a few things with strings (probably due to the rotating buffer being used), so I need to investigate that further 08:18:14 in parable, I now have it loading a precompiled image/session, so startup is much faster now 12:51:01 --- join: obobo_ (~chatzilla@dyn-dsl-mb-216-168-124-99.nexicom.net) joined #retro 12:53:10 --- quit: obobo (Ping timeout: 264 seconds) 12:53:22 --- nick: obobo_ -> obobo 13:25:15 --- join: Mat2 (~claude@91-65-144-133-dynip.superkabel.de) joined #retro 13:25:19 hello 13:28:53 crc: I think, hashed disctionary access is slow because the interpretation overhead for hashing is huge (because of retro's string handling, use of slow multiplication and the dispatch overhead) 13:35:02 ^dictionary 13:37:27 hey Mat2 :) 13:40:53 hi tanentstorm 13:43:01 just read a (for me) bizarre thread on comp.lang.forth about race stereotypes and gun legislation ?!??? 13:44:37 ... that doesn't sound like it was very on topic :) 13:45:07 not really 14:05:50 I've found ace128 on the net (a unix clone for commodore 8 bit computers like the c128), nice :) 14:19:54 --- join: ncv_ (~quassel@89.35.216.197) joined #retro 14:59:15 --- quit: ncv_ (Remote host closed the connection) 15:40:02 --- join: nal (~nal@173.215.251.14) joined #retro 16:08:17 --- quit: harrison (Ping timeout: 276 seconds) 16:11:58 ciao 16:12:07 --- quit: Mat2 (Quit: Verlassend) 16:16:22 --- join: harrison (~quassel@li89-226.members.linode.com) joined #retro 16:57:12 --- quit: nal (Ping timeout: 252 seconds) 17:08:55 --- join: nal (~nal@173.215.251.14) joined #retro 18:05:29 --- quit: ncv (Remote host closed the connection) 20:01:16 --- quit: nal (Quit: Leaving) 23:59:59 --- log: ended retro/13.01.05