00:00:00 --- log: started retro/12.09.29 01:51:34 --- quit: tangentstorm (Ping timeout: 260 seconds) 01:57:58 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro 06:29:22 tangentstorm: I read reddit from time to time (username there is crc-x) 08:56:27 hi all 09:04:56 --- quit: tangentstorm (Ping timeout: 248 seconds) 10:32:20 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro 10:33:39 hey all 10:54:21 --- join: kumul (~kumul@adsl-72-50-90-224.prtc.net) joined #retro 10:56:56 hiya 11:13:29 hi 11:16:19 good afternoon 11:18:50 the gang's all here :) what are you guys up to today? 11:22:38 looking into using sqlite to store the parable session data instead of flat files 11:28:30 inside the browser? 11:30:06 sqlite is cool 11:31:18 * erider just got finish setup of some symbolic links for easy of uses of a tool 11:34:06 tangentstorm: on the server 11:34:58 I'm shifting away from client side storage in parable. server-side gives me more options and space to do what I want/need 11:36:37 nice 11:37:48 you ever see the rainbowforth multi-user forth app on appengine? 11:37:50 http://rainbowforth.sourceforge.net/ 11:38:10 it's sort of a cross between colorforth and a wiki 11:38:37 like you can go in and claim a block on the shared server.. kind of cool 11:39:25 --- quit: erider (Ping timeout: 264 seconds) 11:39:49 rainbow forth doesn't work with my ipad 11:39:50 i agree though. server-side gives you a lot more interesting options :) 11:40:21 ah yeah.. kinda keyboard-centric 11:41:33 keyboard events don't work well on mobile browsers :( 11:56:20 sqlite would make it easier for me to maintain a history of projects 12:37:56 12:52:14 tangentstorm: the contributing guide with links is now online; thanks! 12:52:48 * crc needs to get the update script to generate the html before synching things 12:58:47 np 12:59:56 i was kind of thinking about running the docs through sphinx... it's what the python docs use.. it's not the best, but it generates a little table of contents and some rudimentary site navigation 13:06:30 re sqlite ... i've been thinking for a while that a forth dictionary already has most of the work of a tablular database implemented.... all that's missing is the relational part, and to make it a little more generic. 13:07:12 sphinx != speech software? 13:07:47 i have a tendency to reinvent the wheel, i know... especially on this project... :) but the relational algebra operators are pretty straightforward 13:08:11 reinventing the wheel is half the fun :) 13:08:28 that would be interesting to see. 13:08:31 docl : haha.. no... i went to install it last night and there are like 900 projects named sphinx.. http://sphinx.pocoo.org/ 13:08:53 ahh gotcha 13:09:35 I like the concept of retro being a platform on top of which other languages can be added. 13:10:04 i was gonna call my db "minrel" .. it's not meant to compete with sqlite... it's more for programming or very small decision-making things... almost like a little prolog... 13:11:14 * docl has played with minimal tables and relational stuff in the past with retro. 13:11:24 not that I actually know what I'm doing or anything ;) 13:11:44 all i was doing was looking at e.f. codd's old papers. 13:12:39 when you think about it, it all boils down to linked lists and offsets where you can put pointers. 13:12:53 it's the same concept as the dictionary in retro. 13:13:06 i think you make a little tiny relational engine and stick it in the browser, for example, and then you don't need a json parser... you just ship tables. 13:13:09 yes, exactly 13:13:51 big serious databases use a b+tree or something to index the data, but if you're only dealing with a few hundred records... 13:14:13 going to json and back again might be nice for human readability and stuff though. 13:15:25 the way i was doing it, i was pumping the tables into some generic visualizations with d3.js so it looked pretty decent 13:15:28 one sec. 13:15:37 hmm. as long as you can get good reversibility, it might be nice to support lots of formats in the long run. but yeah for playtesting we can do it with simple stuff. 13:17:04 (trouble with supporting lots of stuff, as crc said regarding alt languages, is time costs.) 13:17:59 tell me about it :) 13:18:13 we can write all this stuff for fun, but if it isn't maintainable it ends up getting scrapped or obsoleted. 13:19:51 i'm not really doing this for fun... i mean i guess i'm having fun, but there's kind of a goal behind it 13:20:55 what kind of goal? 13:21:17 * docl has been planning to write a mud in retro for several years now 13:22:08 sorry, not trying to be all mysterious or anything... i trailed off thinking about how to explain it. i haven't had much success in the past. :) 13:23:47 no problem, I have hard to describe ideas too. :) 13:24:39 i learned a while back i "have adhd"... whatever that means... and about a year ago, i finally got on a decent medication for it... and ever since then i've been trying to figure out what causes it 13:24:48 planning to write web apps? 13:25:51 wouldn't surprise me if I have that too. the hyperfocusing kind? 13:25:52 because if you look in the DSM-IV it barely means anything.... anyway, i started with the question "how can i get better at getting all these ideas out of my head and into the world?" instead of .. .well, like you basically said... abandoning my projects 13:26:20 "the primarily inattentive type" is what they call it... but yeah, hyperfocus too :) 13:27:38 anyway, i'm willing to allow that there is some physiological issue underneath, but i mostly think it's sort of an issue with memory 13:28:36 so you're working on a self-help tool for that? 13:28:38 well, many tangents later i wound up trying to design a programming language and user interface that fits with my crazy ideas about how the mind works 13:28:49 ahh 13:29:30 yeah, and also kind of a self-help tool i guess... 13:30:29 the database thing is part of that. i think the fact that we're stuck with linear text is sort of an un-necessary constraint that causes a lot of problems for us 13:30:30 well I'd be happy to test it for you. lemme see. I will precommit to 20 minutes at minimum. 13:30:52 have you seen this? http://worrydream.com/LearnableProgramming/ 13:31:10 be careful :D if it works the way i expect it to, saying that will be like saying "oh, i'll try world of warcraft for 20 minutes..." :) 13:31:27 the domain looks familiar... checking 13:32:25 heh, well if I could get the kind of focus wow and so forth give, on a useful project, that could turn awesome. 13:32:49 sorry if I just gave you a lengthy distraction. :) 13:32:54 oh, processing. :) the other aspect of this is that i want to make a programming course... a year ago i was making a game programming course with processing 13:34:41 OH! went to the main page of worrydream.com ... that guy's awesome :) 13:36:19 and that would be javascript i'm looking at, not processing. :) 13:39:12 or processing-js.. wow, good stuff.. i'm going to have to read this again. thanks. 13:39:27 sure thing 13:45:02 hmm. I wonder if a retro word called "tutorials" which pulls up a list of things people can do step by step to learn how it works would be good. 13:59:49 hmm. examples/editor.rx is broken. tab does nothing and the only thing that does anything seems to be enter. I will look into this. 14:03:02 it seems to be trying to process }doc as a word. 14:03:19 it must be losing the dictionary at a bad spot 14:05:23 ah, at the }} is where it starts not understanding things 14:07:36 crc: did the behavior of {{ ---reveal--- }} change recently? I'm using the snapshot from the website. 14:12:17 hmm. still seems to work in a simpler example. 14:14:49 commenting out "128 setBlocks" seems to cause it to compile, although it segfaults if I type edit. 14:16:15 ok, fixed it. move the line "128 setBlocks" to just after the }} and it works. 14:17:00 docl : https://github.com/sabren/b4/blob/master/ref/crc-retro-wisdom.org 14:21:07 it shouldn't have changed 14:21:43 you said earlier it wouldn't surprise you if you had adhd too... i don't know you, so it's entirely possible... but i think the modern GUI computer interface causes very adhd-like symptoms, especially for programmers. 14:22:53 crc: somehow the setBlocks command is causing a problem in }}, or so it seems. 14:23:02 notice sometime how often you have to hold a few bits of information in your head while you switch back and forth between windows sometime ( like remembering which tab you were on in your browser while you follow a link, or worse: have to go google something to understand what you're reading ) 14:23:07 bbl 14:24:13 docl: yup. I moved the 128 setBlock like you did. 14:24:18 I'll have to dig into this 14:25:05 crc: I amend that. setBlocks seems to cause the dictionary to disappear when run before }}. 14:25:07 I think it has to do with header removals (auto removal of hidden headers) introduced a while back 14:25:30 the block buffer is being allocated overtop of it 14:26:57 tangentstorm: good point. 14:27:46 crc: that makes sense. I will see if I can find anything else broken in the examples. 14:27:47 docl: fixed in rev 362 (now pushed to remote repo) 14:33:45 whois docl 14:34:01 * crc has no idea where the whois came from 14:39:48 docl: the only other example that I can see that may have this particular issue is the corpse implementation, and that's running without problems thus far 14:40:38 docl: I'll write up something to help prevent others from running into this in the future 14:51:37 if I load editor.rx, edit a block and save it, then load up autopsy.rx, it gives me a segmentation fault. 14:59:34 apparently the editor is messing with something important for autopsy when it edits blocks. 15:21:21 odd 15:30:44 I'm not sure what is wrong :( 15:47:36 try rev 363 16:04:45 seems to work 16:05:07 good 16:05:25 the editor remaps the key handler, and didn't restore the default one when exiting 16:07:14 ahh, that makes sense 16:09:41 odd that it never showed up as a problem before. just goes to show how little use it actually got 16:24:52 I wonder if there could be a list kept somewhere of every vector that gets set, so you can easily check for bad vectors. 16:26:21 it'd be possible to write a tool to scan the visible dictionaries for revectored functions 16:26:29 I may need to do that 16:35:44 : vectored last [ dup @d->xt @ #8 == [ d->name puts space ] [ drop ] if ] ^types'LIST each@ ; 16:35:52 will work for the global dictionary 17:02:59 I'll add this to autopsy as words:vectored 17:03:08 cool :) 17:05:50 I'll probably add a few others (words:byClass and such) as well 17:13:47 :) 17:14:11 i'm actually trying to do the same thing right now, from the outside, in pascal 17:14:39 maybe i'll start by trying to translate autopsy and casket 17:25:02 hmm. 'explore' only shows the xt of the function being called. can it be made to show the name of the function as well? 17:26:05 docl : i asked crc that in the log i showed you... i think he said the disassembler has the code to do that 17:26:28 docl: I'll see if I can add that 17:26:48 autopsy 17:26:51 that will be awesome 17:27:37 tangentstorm: I remember when we added support for xt->d a while back, which looks up dictionary headers based on an xt. 17:27:38 crc : maybe i can't do what i was thinking of doing... how would i find the address of the end of the linked list for the dictionary without actually running the interpreter? 17:28:18 * crc will be commiting to lp:~crc-x/retro-language/examples for autopsy and other bits until these things are done, then merging into the main repo 17:29:01 last, the variable holding the pointer, is always located at memory location 2 17:29:17 aha! genius. 17:29:39 this, and a few other bits at the start are kept static to allow for external tools to operate on an image more easily :) 17:31:39 crc : i don't suppose you're an emacs guy? 17:31:57 er.. 17:32:15 i should have known that from the commentary, but it drives me nuts switching back and forth 17:32:26 I haven't used emacs in several years 17:32:56 * docl has been using vim more heavily recently. 17:32:58 what do you use? 17:33:31 hrm 17:34:18 vim and nano 17:34:24 sorry, i'm not getting to the point... :) i'm trying to think of a good way to merge the commentary and the actual code for the bootstrap process 17:35:01 tangentstorm: I'd like to be able to do that 17:36:19 there's noweb... or really it could just be commented out... but i've been spoiled by this thing in emacs called org-mode 17:37:21 the only problem with it is that once you put your code into org-mode it's pain in the butt for everyone else and you have to extract it each time you want to compile 17:38:33 nm.. going back to work here. :) 17:39:37 'll look into that 18:30:31 tangentstorm: I just found your reddit post on object oriented programming. great explanation :) 18:33:29 --- join: arescorpio (~arescorpi@222-206-17-190.fibertel.com.ar) joined #retro 18:35:47 * docl learned about python's 'yield' syntax and generators fairly recently. but already seeing parallels in studying haskell IO stuff. 21:52:55 --- quit: arescorpio (Quit: Leaving.) 23:21:28 --- quit: kumul (Quit: WeeChat 0.3.8) 23:59:59 --- log: ended retro/12.09.29