| 00:17.06 | file[lr] | hi Xentac man |
| 00:17.20 | Xentac | hi file[lr] man |
| 00:17.25 | file[lr] | what'cha doin! |
| 00:17.32 | Xentac | being annoyed |
| 00:17.44 | file[lr] | clarify please |
| 00:18.13 | Xentac | the one co-op job I wanted to apply for, and get, they aren't hiring anyone from the west |
| 00:18.28 | file[lr] | ic |
| 00:19.26 | Xentac | so now I have to apply for other co-op jobs in the east... |
| 00:20.14 | file[lr] | uh huh |
| 00:20.30 | Equality72521 | anyone seen Liberty 5-3000? |
| 00:20.49 | Xentac | otherwise I might just get a job here and then take 3 or 4 weeks off in the summer... |
| 00:21.00 | Xentac | KeyserSoze: ummm... no? |
| 00:21.01 | KeyserSoze | anyone seen Sophie? |
| 00:21.19 | Xentac | what the hell? |
| 00:21.43 | Xentac | geez... I bet spammers would get more customers if they spelled the subject headings righ |
| 00:21.45 | Xentac | s /righ/right |
| 00:21.54 | Xentac | what the hell is "aptillery dertruction"? |
| 00:26.21 | Xentac | hehehe... part of the Squid Test: Have you ever picked up traffic cones on the fly and set them in the middle |
| 00:26.21 | Xentac | <PROTECTED> |
| 00:36.30 | Xentac | KeyserSoze: still there? |
| 00:36.35 | KeyserSoze | yup |
| 00:36.40 | KeyserSoze | Sophie came home, too. |
| 00:36.46 | Xentac | did you like the squid test question? |
| 00:36.49 | Xentac | I found another good one... |
| 00:37.02 | KeyserSoze | tell it to me, then! |
| 00:37.07 | KeyserSoze | you big tease |
| 00:37.13 | Xentac | When carrying a passenger, do you accelerate and decelerate more vigorously if the passenger is of the preferred sex (for intimate companionship)? |
| 00:37.31 | KeyserSoze | that's obvious, isn't it? |
| 00:37.43 | Xentac | this is to see how much of a squid you are though... |
| 00:37.45 | Xentac | so it's yes or not |
| 00:37.49 | Xentac | s /not/no |
| 00:38.08 | Xentac | Have you ever taken a friend's key out of their ignition at a stop light and then ridden off? |
| 00:40.51 | Xentac | Do you regularly ride armed with paint-ball gun? |
| 00:42.29 | Xentac | Have you ever stopped at a stoplight and had the people in the car next to you lock their doors? |
| 00:42.36 | Xentac | ^---- what the hell's up with that? |
| 00:42.45 | file[lr] | my mobos and CPU get sent out tomorrow |
| 00:42.47 | file[lr] | :) |
| 00:42.51 | file[lr] | the echeck finally cleared |
| 00:43.50 | file[lr] | 2 day to Illinois |
| 00:43.53 | file[lr] | then standard to me |
| 00:44.26 | Xentac | Have you ever had to strip by the side of the road because a bee got in your clothes? |
| 00:44.33 | file[lr] | no |
| 00:44.42 | Xentac | Did traffic slow to watch? |
| 00:44.46 | Xentac | Did you get any tip money? |
| 00:45.04 | file[lr] | unfortunately not... |
| 00:49.10 | Xentac | Have you ever altered or defeated your bike's rev-limiter? <---- hey my bike has one of those! |
| 00:49.18 | file[lr] | lol |
| 00:49.45 | Xentac | but my bike isn't exactly the sort of bike you'd race with... |
| 00:51.17 | KeyserSoze | corner harder, too, and then they have to hug you tight |
| 00:51.17 | KeyserSoze | stupid internet stopped working again. |
| 00:51.35 | Xentac | KeyserSoze: this is true... it's not like I didn't check that box off ;o) |
| 00:55.42 | Xentac | Have you ever added oil to a British bike by pouring it on the outside and hoping it would leak in? |
| 01:09.25 | *** join/#picogui gonkulator (~brandon@h-66-167-168-91.DNVTCO56.covad.net) |
| 01:10.59 | file | wb gonkulator |
| 01:11.04 | gonkulator | thanks file |
| 01:21.25 | captain_proton | scanline: poke |
| 01:22.16 | gonkulator | *scanline ows |
| 01:24.29 | Xentac | dude... |
| 01:24.36 | captain_proton | DUDE |
| 01:24.38 | Xentac | I'm only 2.179 percent squid |
| 01:24.54 | Xentac | I suppose that is a good thing... |
| 01:24.54 | file | Xentac: isn't... that... nice... ... right |
| 01:25.17 | scanline | hey |
| 01:25.38 | Xentac | after filling out a 1000 question survey.. I find that I'm 2.179% squid... |
| 01:25.55 | scanline | now that I've been discussing my ideas in #fresco, what are the odds that fresco and picogui end up exactly the same other than picogui having better IPC, better drivers, and better apps? :) |
| 01:26.00 | captain_proton | scanline: shall we go do this thingy? |
| 01:26.36 | scanline | would rather do this tomorrow, when it's less cold and I'm not trying to get this equation to work |
| 01:26.41 | captain_proton | scanline: i can't do it tomorrow |
| 01:26.46 | captain_proton | considering its due tomorrow |
| 01:26.46 | captain_proton | :P |
| 01:26.55 | scanline | I'll follow you if ya want |
| 01:27.00 | captain_proton | k |
| 01:27.02 | captain_proton | lessago |
| 01:40.21 | *** join/#picogui Cambo (mmmeeee@cm216036165101.westman.wave.ca) |
| 01:40.45 | *** join/#picogui kidder (~s@h24-65-171-188.gv.shawcable.net) |
| 01:52.56 | njs | "<scanline> If you dot a function's gradient with a unit vector, shouldn't that give you the directional derivative along that vector?" <-- in case you still care, the answer is yes |
| 02:34.18 | Xentac | KeyserSoze: still around? |
| 02:34.28 | KeyserSoze | yup |
| 02:34.38 | Xentac | what were those pvr progs you listed? |
| 02:34.58 | KeyserSoze | mythtv, freevo, freeplaytv |
| 02:35.21 | KeyserSoze | mythtv is imo the best, but freevo and freeplaytv have neat names |
| 02:35.23 | Xentac | I was checking out mythtv and freevo last night |
| 02:35.36 | Xentac | mythtv (IIRC) is written in perl |
| 02:35.39 | KeyserSoze | freevo = free + tv, freeplaytv = free + replaytv |
| 02:35.50 | kergoth | yo |
| 02:35.53 | Xentac | freevo in python |
| 02:35.55 | Xentac | hey kergoth |
| 02:36.52 | Xentac | mythtv is almost entirely tv oriented |
| 02:37.00 | Xentac | freevo handles movies, music, ripping cds, etc etc |
| 02:38.44 | KeyserSoze | http://mythtv.sourceforge.net/mc/bluemusicmenu.png |
| 02:39.02 | Xentac | yes, mythtv does music as well |
| 02:39.11 | Xentac | but not nearly as nicely (what what I could gather) |
| 02:39.30 | Xentac | mythtv can't really handle movies either... |
| 02:40.49 | KeyserSoze | it's got a weather module |
| 02:40.53 | Xentac | this is true... |
| 02:41.09 | Xentac | but I bet it wouldn't be that difficult to write a weather module for freevo... after all, it is python... and plugin based ;o) |
| 02:41.31 | Xentac | each one has it's benefits |
| 02:41.49 | Xentac | I think mythtv looks good right now, but freevo would be my bet (and where I'd probably contribute) in the long run |
| 02:41.51 | KeyserSoze | last i checked, freevo couldn't pause live tv |
| 02:42.07 | Xentac | no, it can't |
| 02:42.11 | Xentac | it just plays it |
| 02:42.26 | KeyserSoze | yes, mythtv can pause it. |
| 02:43.18 | KeyserSoze | mythtv has a php web interface |
| 02:43.35 | KeyserSoze | which comment? |
| 02:43.45 | Xentac | I think mythtv looks good right now, but freevo would be my bet (and where I'd probably contribute) in the long run |
| 02:43.55 | KeyserSoze | why? |
| 02:44.27 | Xentac | because freevo has plans to add the features of mythtv that are nice... but mythtv doesn't have the plans to add the features of freevo that are nice? |
| 02:44.34 | Xentac | and freevo is written in python... |
| 02:45.25 | KeyserSoze | what's wrong with my mouse? |
| 02:45.35 | Xentac | it's not written in python? |
| 02:46.12 | njs | mixing a mouse and python is not recommended, if you actually want a mouse |
| 03:50.05 | *** join/#picogui gonkulator (~brandon@h-66-167-168-91.DNVTCO56.covad.net) |
| 03:50.13 | gonkulator | hmm... |
| 03:50.18 | gonkulator | ibot: seen PicoBot |
| 03:50.22 | | picobot <~PicoBot@aden2-241-dhcp.Colorado.EDU> was last seen on IRC in channel #picogui, 5d 7h 18m 34s ago, saying: 'shut was last seen on #picogui 2 seconds ago, saying: up file [Thu Jan 23 13:31:44 2003]'. |
| 03:57.21 | gonkulator | w00t |
| 03:58.18 | KeyserSoze | how do a character which has a leading 1 to an unsigned int, and printf it as an unsigned int? |
| 03:59.27 | gonkulator | KeyserSoze: beats the hell out of me |
| 03:59.33 | gonkulator | KeyserSoze: I have no idea |
| 03:59.35 | KeyserSoze | i declare a variable as "unsigned int", and I set it equal to the result of a typecast to "unsigned int" on a character that was assigned to 0xFF, and printf it with %ui, and I get 4294967295 |
| 03:59.43 | KeyserSoze | gonkulator: you really don't know? |
| 03:59.50 | KeyserSoze | it's driving me nuts. |
| 03:59.53 | gonkulator | KeyserSoze: I am pretty much clueless |
| 04:00.05 | KeyserSoze | i can get it to print -1, or 4294967295, but not 255. |
| 04:00.29 | gonkulator | KeyserSoze: I don't have any useful input on that one |
| 04:00.54 | KeyserSoze | i'm skeptical. don't you write C? |
| 04:00.59 | gonkulator | I write some C |
| 04:01.01 | KeyserSoze | are you really picobot? |
| 04:01.08 | gonkulator | but I don't know much about bit order stuff |
| 04:01.21 | gonkulator | why are you using %ui? |
| 04:01.38 | gonkulator | why not just use %d? |
| 04:01.58 | gonkulator | anyway, I am wanting to play some games |
| 04:02.07 | KeyserSoze | i tried %d first, and it didn't work |
| 04:02.19 | kergoth | is that an unsigned char or a signed char? :) |
| 04:03.17 | KeyserSoze | there are signed and unsigned chars? |
| 04:03.21 | KeyserSoze | i didn't know that. |
| 04:03.33 | KeyserSoze | i just used "char" |
| 04:03.50 | KeyserSoze | kergoth: thank you!!! |
| 04:03.55 | kergoth | char isnt necessarily unsigned. its implementation dependent. if your'e casting a char, pay attention to that |
| 04:03.59 | kergoth | ;) |
| 04:03.59 | kergoth | no problem |
| 04:04.03 | KeyserSoze | man, i spent half an hour messing around with that. |
| 04:04.09 | kergoth | pure 'char' could be either one, depending on what OS you're on |
| 04:04.11 | kergoth | or what have you |
| 04:04.50 | KeyserSoze | i don't understand that. i can't imagine why char would be signed or unsigned. but thank you. |
| 04:16.03 | *** part/#picogui testuser (~testuser@Toronto-HSE-ppp3659577.sympatico.ca) |
| 05:24.43 | Xentac | tonight I was taped pretending to beat someone up... |
| 05:26.06 | Xentac | KeyserSoze: you wanna know why char can be signed or unsigned? |
| 05:27.05 | Xentac | I think it might have to do with 7 bit and 8 bit chars... and extended ascii and regular ascii.. |
| 05:46.32 | *** join/#picogui gonkey{iBook} (~brandon@h-69-3-49-169.DNVTCO56.covad.net) |
| 06:41.00 | *** join/#picogui agonk{nov} (~brandon@h-69-3-49-169.DNVTCO56.covad.net) |
| 07:28.17 | captain_proton | ~particle du jour |
| 07:28.18 | | excited pseudoelectron |
| 08:52.04 | scanline | ~particle du jour |
| 08:52.05 | | charged antiproton |
| 09:12.24 | *** join/#picogui lalo[out] (~lalo@200.203.12.232) |
| 09:15.27 | *** join/#picogui gonked_out (~brandon@h-66-167-168-166.DNVTCO56.covad.net) |
| 09:51.03 | captain_proton | so this guy works at an aquarium |
| 09:51.14 | captain_proton | and his boss comes to him and tells him that the dolphins are feeling very amorous |
| 09:51.28 | captain_proton | which is a problem, because they have a bunch of third graders coming in an hour |
| 09:51.49 | captain_proton | so she tells him to go get some baby seagulls, because they're the only anti-aphrodesiac for dolphins |
| 09:52.20 | captain_proton | "but be careful!" she says, "a lion escaped from the zoo this morning, and while heavily tranquilized, is still very dangerous" |
| 09:52.53 | captain_proton | so this guy goes out to the beach and picks up some baby seagulls |
| 09:53.05 | captain_proton | but when coming back, he finds the lion, lying across the sidewalk |
| 09:53.13 | captain_proton | as he gingerly steps over it, he gets arrested |
| 09:53.26 | captain_proton | for carrying young gulls across a sedate lion for immoral porpoises |
| 09:54.01 | captain_proton | :) |
| 09:54.43 | captain_proton | okay, off to bed |
| 09:54.50 | scanline | g'night |
| 09:55.48 | *** join/#picogui hyu_ (hyu@aqima.com) |
| 09:59.04 | *** join/#picogui fraggle (~fraggle@tuneless.plus.com) |
| 11:03.23 | *** join/#picogui agonk{nov} (~brandon@h-66-167-168-166.DNVTCO56.covad.net) |
| 13:45.26 | *** join/#picogui hyu_ (hyu@aqima.com) [NETSPLIT VICTIM] |
| 13:45.26 | *** join/#picogui file (jwired@mctn1-1540.nb.aliant.net) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui fraggle (~fraggle@tuneless.plus.com) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui darth_angel (Admin@aden2-92-dhcp.resnet.Colorado.EDU) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui scanline (~micah@aden2-83-dhcp.resnet.Colorado.EDU) |
| 13:49.16 | *** join/#picogui KeyserSoze (~ksoze@12-245-37-229.client.attbi.com) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui captain_proton (~jupiter@aden2-23-dhcp.resnet.Colorado.EDU) |
| 13:49.16 | *** join/#picogui jojo (~jojo@infidyne.com) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui prot-screen (~jupiter@aden2-23-dhcp.resnet.Colorado.EDU) |
| 13:49.16 | *** join/#picogui njs (njs@12-232-144-227.client.attbi.com) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui kergoth`zzz (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 13:49.16 | *** join/#picogui hikke (piste@drinking.olvi.cx) [NETSPLIT VICTIM] |
| 13:49.16 | *** join/#picogui nemo (nemo@61.95.53.30) [NETSPLIT VICTIM] |
| 13:49.29 | *** join/#picogui lalo (~lalo@200.203.12.232) |
| 13:50.36 | *** join/#picogui nemo (nemo@61.95.53.30) [NETSPLIT VICTIM] |
| 13:50.36 | *** join/#picogui hikke (piste@drinking.olvi.cx) [NETSPLIT VICTIM] |
| 13:50.36 | *** join/#picogui kergoth`zzz (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 13:50.36 | *** join/#picogui njs (njs@12-232-144-227.client.attbi.com) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui prot-screen (~jupiter@aden2-23-dhcp.resnet.Colorado.EDU) |
| 13:50.37 | *** join/#picogui jojo (~jojo@infidyne.com) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui captain_proton (~jupiter@aden2-23-dhcp.resnet.Colorado.EDU) |
| 13:50.37 | *** join/#picogui KeyserSoze (~ksoze@12-245-37-229.client.attbi.com) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui scanline (~micah@aden2-83-dhcp.resnet.Colorado.EDU) |
| 13:50.37 | *** join/#picogui darth_angel (Admin@aden2-92-dhcp.resnet.Colorado.EDU) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui fraggle (~fraggle@tuneless.plus.com) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui file (jwired@mctn1-1540.nb.aliant.net) |
| 13:50.37 | *** join/#picogui hyu_ (hyu@aqima.com) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui agonk{nov} (~brandon@h-66-167-168-166.DNVTCO56.covad.net) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui gonked_out (~brandon@h-66-167-168-166.DNVTCO56.covad.net) [NETSPLIT VICTIM] |
| 13:50.37 | *** join/#picogui Xentac (~jchu@amga45cyy30x9.bc.hsia.telus.net) [NETSPLIT VICTIM] |
| 14:20.09 | *** part/#picogui Piete (~abri@61.98.19.77) |
| 14:29.29 | gonked_out | and off I go to school and work |
| 15:40.35 | lalo[out] | cool, Guido also has development ideas in the shower. |
| 15:40.47 | lalo | what's with showers and software design? |
| 15:41.04 | scanline | they're both wet? |
| 15:41.42 | lalo | I suppose if you program using lots of streams... |
| 15:42.00 | scanline | hey, try googleing for "LineaLX" |
| 15:43.14 | lalo | "miniature tabletpc" sounds like a fud-enchanced term for pda |
| 15:43.30 | scanline | yeah, pretty much |
| 15:43.41 | scanline | the part that I thought was interesting was the mention of Royal's PDA near the bottom :) |
| 15:45.30 | lalo | should we tell them? |
| 15:45.56 | scanline | I'm interested in seeing exactly how the folks at Valourtech managed to mangle picogui first :) |
| 15:47.13 | scanline | Terence Hui, who used to work with Agenda, is now doing the PicoGUI stuff at Valourtech.. he says they have almost completed a PIM, media player, email app, and AIM client... and they want my help with the web browser |
| 15:47.58 | lalo | cool :-) |
| 15:48.09 | scanline | I'm still skeptical until I see the code |
| 15:48.10 | lalo | will this stuff be released? |
| 15:48.20 | scanline | dunno |
| 15:50.24 | scanline | In other news, the ideas for the "object kernel" in the new picogui are starting to form something tangible |
| 15:50.37 | scanline | might actually have some code soon :) |
| 15:51.41 | lalo | well, I have a nagging feeling. |
| 15:52.21 | Xentac | lalo: what's nagging you...? |
| 15:53.54 | lalo | duplication |
| 15:54.15 | Xentac | oh what? code in the new picogui architecture? |
| 15:54.57 | lalo | scanline seems to be writing 70% of what will be in om :-) |
| 15:55.06 | scanline | hehe |
| 15:55.15 | scanline | well, it could be that what I'm writing could be used in om... |
| 15:55.27 | scanline | the object model will be a separate library, with nothing specific to picogui |
| 15:55.34 | lalo | when you asked me about using om I thought it wasn't so hot an idea, but now I'm having second thoughts |
| 15:56.05 | lalo | we could write whatever we need for picogui and that would be milestone 1, then I proceed from there |
| 15:57.01 | *** join/#picogui darth_iBook (~evanshee@duan96-177-dhcp.Colorado.EDU) |
| 15:57.19 | lalo | and then when I proceed we magically get fun abilities like remote implementation of widget classes and arbitrary file formats... |
| 15:57.48 | scanline | well, I was expecting to allow both of those in picogui's object model |
| 15:58.07 | scanline | widgets would store all of their data inside the kernel, so you could disconnect a widget, recompile, and reconnect a widget for example |
| 15:58.22 | scanline | and arbitrary objects/formats would just be a matter of defining new object types |
| 15:58.26 | lalo | the first stage of anything is the data types, which is where I stopped in om (when work called and I had to dedicate my free hours to opental) |
| 15:59.05 | lalo | aha, but remember in the om design the file format (persistence) can be completely disconnected with the object class. |
| 15:59.19 | scanline | ahh |
| 15:59.20 | lalo | you could store live widgets in xml files if it suits you |
| 16:00.00 | lalo | no that I can see how this would be useful today :-) |
| 16:00.20 | scanline | indeed |
| 16:00.20 | lalo | but IMO good design is the art of giving people tools to do stuff you didn't foresee. |
| 16:00.40 | lalo | (ooo, that was deep. Let me save it somewhere.) |
| 16:05.55 | scanline | =) |
| 16:05.55 | Xentac | lalo: you just made it into my signature file ;o) |
| 16:05.55 | scanline | well, I just realized I'm really hungry... should go grab a bagel or something, brb |
| 16:05.55 | lalo | but if I was to apply current picoguiNG design to current om design, the widget server (we now call it kernel, right?) essentially stores the objects in a non-persistent storage |
| 16:05.55 | scanline | err, bbiab |
| 16:05.55 | lalo | damn, out of batteries again, brb |
| 16:05.55 | scanline | lalo: I was thinking the kernel could be optionally persistent |
| 16:05.56 | lalo | scanline: om gives you that gratis. ;-) |
| 16:05.56 | scanline | lalo: if apps stored their important data in the kernel too, the apps would inherit persistence or synchronization that's implemented in the kernel |
| 16:05.56 | scanline | :) |
| 16:05.56 | scanline | well, bagel calls |
| 16:06.00 | kergoth | morning guys |
| 16:06.37 | darth_iBook | yay! cos of a potato! |
| 16:07.24 | lalo | heh |
| 16:07.32 | lalo | funniest thing I heard this year: |
| 16:08.00 | darth_iBook | ? |
| 16:08.42 | lalo | <some guy at #hurd> well, this problem made me find the potatos I thought I lost wednesday, but I'd still like to see it solved. |
| 16:09.29 | darth_iBook | lol |
| 16:14.49 | darth_iBook | this is making my head hurt |
| 16:23.10 | *** join/#picogui ahn_class (~Gnu@aden2-135-dhcp.resnet.Colorado.EDU) |
| 16:23.10 | darth_iBook | hey ahn_class |
| 16:23.10 | Xentac | ahn_class: you have a laptop too? |
| 16:23.10 | darth_iBook | ahn_class: did you steal angels? |
| 16:23.10 | ahn_class | no, i haven't changed it back yet |
| 16:23.11 | Xentac | ah |
| 16:23.11 | darth_iBook | oh |
| 16:23.11 | ahn_class | i am supposed to be in class, so i am not going to bother |
| 16:23.11 | darth_iBook | lol |
| 16:23.12 | ahn_class | but my teacher thought i was to sick to be in class |
| 16:23.15 | ahn_class | i wish |
| 16:24.43 | ahn_class | darth_angel: the hell name is this....i am glad you two are oh so happy, but you can be two people you know |
| 16:24.43 | ahn_class | that's just sick |
| 16:31.23 | darth_iBook | she never changed it back |
| 16:31.23 | darth_iBook | that's the nick I use when I'm on her comp |
| 16:31.23 | ahn_class | heh, i use ahkana. |
| 16:31.23 | ahn_class | no wait |
| 16:31.23 | ahn_class | i use ahnkana |
| 16:31.23 | darth_iBook | not to indicate we are the same person (that would be weird ;) but to show where I am |
| 16:31.23 | ahn_class | i can spell my own name |
| 16:31.23 | ahn_class | aha |
| 16:31.23 | ahn_class | still...bleh |
| 16:31.23 | darth_iBook | lol |
| 16:31.23 | ahn_class | are you in your ibook? |
| 16:31.24 | ahn_class | (but then, am i in class?) |
| 16:31.24 | darth_iBook | in? |
| 16:31.24 | ahn_class | when you have darth_angel, you are in angel's room |
| 16:31.24 | ahn_class | so darth_iBook .... are you in your ibook? |
| 16:31.24 | darth_iBook | rather, it is to indicate I am using her comp |
| 16:31.24 | ahn_class | i am just doing what i do best |
| 16:31.24 | ahn_class | causing trouble |
| 16:31.24 | darth_iBook | ;) |
| 16:31.24 | ahn_class | liar....you aren't really there |
| 16:31.24 | scanline | lalo: thought much about security and network transparency in Om yet? |
| 16:31.24 | ahn_class | ~particle of the day |
| 16:31.25 | | charged pseudophoton |
| 16:31.25 | darth_iBook | inversy antielectron |
| 16:31.25 | darth_iBook | ^^ |
| 16:31.25 | ahn_class | thanks darth_iBook |
| 16:31.25 | scanline | hmm |
| 16:31.26 | scanline | guess ibot is lagged... |
| 16:31.26 | darth_iBook | lol |
| 16:31.26 | ahn_class | erm, darn |
| 16:31.26 | scanline | ~particle du jour |
| 16:31.27 | | charged antineutron |
| 16:31.27 | darth_iBook | scanline: quarky antiquasar |
| 16:31.27 | darth_iBook | there |
| 16:31.27 | darth_iBook | =P |
| 16:31.27 | darth_iBook | s/sikc/sick/ |
| 16:31.27 | ahnkana | uh, i wish |
| 16:31.28 | ahnkana | sadly that ivolves sleep, sleep involves time....time involves less scrabble, but scrabble is fun |
| 16:31.28 | lalo | scanline: network transparency comes with the design |
| 16:31.28 | lalo | scanline: security I postponed on purpose, but I was thinking along the lines of capability too. |
| 16:32.23 | ahnkana | wow....that's slow |
| 16:32.23 | ahnkana | lol |
| 16:32.23 | scanline | lalo: hm.. can't seem to find the docs on Om you posted a while ago |
| 16:32.34 | lalo | not really docs, just the motivation manifesto |
| 16:32.39 | scanline | ah |
| 16:32.40 | lalo | http://nongnu.org/om |
| 16:34.19 | lalo | the docs are in cvs, in the todo database |
| 16:34.19 | scanline | 404 |
| 16:34.31 | ahnkana | darth_iBook: por que dos? |
| 16:36.04 | darth_iBook | physics |
| 16:36.04 | ahnkana | oh, you changed it back |
| 16:36.04 | ahnkana | not where are you.....why two? |
| 16:36.04 | darth_iBook | that was supposed to be darth_iBook's not ibooks |
| 16:36.04 | darth_iBook | but it didn't like the ' |
| 16:36.40 | ahnkana | WORD OF THE DAY: sublunary \suhb-LOO-huh-ree\, adjective: situated beneath the moon; hence, of or pretaining to this world; terrestrial; earthly. |
| 16:36.48 | ahnkana | opps |
| 16:36.51 | scanline | lalo: I still can't find it.. that link just gives me a 404 page, if I search for "om" I don't get anything relevent |
| 16:38.14 | ahnkana | s/\suhb-LOO-huh-ree\/\suhb-LOO-nuh-ree\ |
| 16:38.14 | darth_iBook | lalo: how's your kid? |
| 16:38.14 | ahnkana | heh, that looks cool |
| 16:38.14 | ahnkana | too bad it's a spelling correction |
| 16:38.52 | ahnkana | lalo: how's your wife? |
| 16:38.52 | scanline | ahnkana: you'd have to use \\ instead of \ in the pattern, since \ is an escape character |
| 16:38.52 | ahnkana | tell her i miss her, and she should come visit |
| 16:39.25 | darth_iBook | ditto |
| 16:39.25 | ahnkana | scanline: eh, well i am not going to type it again, you get the gist of it... |
| 16:39.25 | ahnkana | scanline: does gist count as a word? |
| 16:39.34 | scanline | you could do: |
| 16:39.35 | lalo | 404? this is bad |
| 16:40.32 | lalo | http://www.nongnu.org/om/ |
| 16:40.32 | scanline | s/s\/\\suhb-LOO-huh-ree\\\/\\suhb-LOO-nuh-ree\\/s\/\\\\suhb-LOO-huh-ree\\\\\/\\\\suhb-LOO-nuh-ree\\/ |
| 16:40.32 | lalo | kid and wife are both in Rio seeing her family and both are great :-) |
| 16:43.09 | ahnkana | oh damn....well alright then.... |
| 16:43.09 | ahnkana | lalo: that |
| 16:43.09 | scanline | :) |
| 16:43.09 | ahnkana | 'opps |
| 16:43.09 | ahnkana | lalo: that is wonderful, will you relay our 'hellos" to her? and to the baby? |
| 16:43.09 | lalo | ahnkana: sure. :-) |
| 16:43.09 | scanline | me too :) |
| 16:43.09 | ahnkana | lalo: thanks |
| 16:43.09 | ahnkana | scanline: did you post our wonderful game? |
| 16:43.09 | ahnkana | eooonichs...:-D |
| 16:43.10 | ahnkana | lol....picogui is eooonichs based isn't it? |
| 16:43.54 | ahnkana | you should put that in the widget....and see how long it takes captain_proton to notice |
| 16:45.32 | ahnkana | ooppps againm |
| 16:45.32 | ahnkana | try this one |
| 16:45.32 | scanline | ahnkana: haven't transferred it off my camera yet |
| 16:46.06 | ahnkana | scanline: it's all good....see you all later |
| 16:46.10 | scanline | have fun |
| 16:49.00 | scanline | <PROTECTED> |
| 16:49.01 | scanline | <PROTECTED> |
| 16:49.08 | scanline | ^ pythonesque :) |
| 16:49.41 | scanline | well, classtime for me too. bbl |
| 16:52.36 | lalo | it's supposed to be pythonesque :-) |
| 17:04.32 | *** join/#picogui darth_iBook (~evanshee@tcom99-216-dhcp.Colorado.EDU) |
| 17:04.41 | darth_iBook | who missed me? |
| 17:12.43 | lalo | I didn't shoot. |
| 17:13.04 | darth_iBook | lol |
| 17:14.35 | darth_iBook | ugh for fizix |
| 17:36.27 | *** join/#picogui gonkulator (~brandon@Metro0276.client.mscd.edu) |
| 17:40.31 | darth_iBook | hi gonkulator |
| 17:43.23 | lalo | hi gonkulator |
| 17:52.29 | gonkulator | hi darth_iBook |
| 17:52.33 | gonkulator | hi lalo |
| 17:52.40 | gonkulator | captain_proton: how much did your cello cost? |
| 17:52.50 | gonkulator | hmm... darth_angel |
| 17:53.08 | captain_proton | gonkulator: a lot |
| 17:53.58 | gonkulator | is it the same "a lot"? |
| 17:54.02 | captain_proton | gonkulator: no |
| 17:54.29 | gonkulator | is it that same a lot? |
| 17:54.59 | captain_proton | perhaps |
| 17:55.05 | gonkulator | 17k? |
| 17:55.46 | captain_proton | ++ |
| 17:56.25 | captain_proton | then again, it qualified as a gradauation gift, two birthday gifts, plus i paid a good amount towards it |
| 17:56.25 | captain_proton | :) |
| 17:57.33 | gonkulator | holy sheite |
| 17:58.02 | gonkulator | hi ahnkana |
| 18:12.11 | lurgyman | gonkulator: http://bzflag.net:3000/music/Mahler_Gustav/ |
| 18:14.27 | lurgyman | :) |
| 18:17.12 | gonkulator | hmm... |
| 18:17.16 | gonkulator | IT might not like this... |
| 18:17.45 | gonkulator | eh, the wurst thing they will do is fire me |
| 18:19.00 | gonkulator | damn... this is cool |
| 18:22.19 | gonkulator | but I should do it when I am home |
| 18:32.06 | lurgyman | it's between universities... things tend to be good between them |
| 18:34.43 | lurgyman | thanks to... big bandwidth! |
| 18:36.06 | lurgyman | i don't want big bandwidth in -my- medicine cabinet |
| 18:40.08 | gonkulator | I've heard that |
| 18:45.03 | *** join/#picogui ahnkana_ (~Gnu@aden2-135-dhcp.resnet.Colorado.EDU) |
| 18:45.03 | gonkulator | hi ahnkana_ |
| 18:56.39 | darth_angel | rawr |
| 19:34.31 | *** join/#picogui ibot (ibot@rikers.org) |
| 19:34.31 | *** topic/#picogui is http://picogui.org || If nobody here answers your question, try the pgui-devel mailing list || PicoGUI v0.44 "Polygons of the Apocalypse" released || New architecture under design: http://picogui.org/doc/neopicogui.png |
| 19:52.45 | *** join/#picogui ahnkana (~Gnu@aden2-135-dhcp.resnet.Colorado.EDU) |
| 20:03.31 | *** join/#picogui darth_iBook (~evanshee@ec237-24-dhcp.Colorado.EDU) |
| 20:03.54 | darth_iBook | yay, made it through poli. sci |
| 20:09.48 | *** join/#picogui merlin{work} (~amccall@crtntx1-ar1-4-60-243-225.crtntx1.dsl-verizon.net) |
| 20:09.50 | merlin{work} | w00t! |
| 20:10.05 | merlin{work} | I love the wonders of decent networking |
| 20:10.06 | merlin{work} | =D |
| 20:10.19 | merlin{work} | hey darth_iBook |
| 20:10.26 | file | I just dared to venture through my closet |
| 20:10.38 | darth_iBook | file: did you end up in Narnia? |
| 20:10.49 | merlin{work} | I just download putty, "run from present location" and I'm on my home computer |
| 20:10.50 | merlin{work} | =D |
| 20:11.01 | file | darth_iBook: I ended up in 386 laptop heaven! |
| 20:11.02 | *** join/#picogui crimson{laptop} (~devnull@12-239-103-228.client.attbi.com) |
| 20:11.17 | merlin{work} | I've installed moria for those days of boredom |
| 20:11.18 | merlin{work} | =D |
| 20:11.46 | crimson{laptop} | hey scanline, gonkulator, file |
| 20:11.50 | gonkulator | hi crimson{laptop} |
| 20:11.51 | darth_iBook | that doesn't sound quite as fun as Narnia, but it works |
| 20:11.54 | gonkulator | hi darth_iBook |
| 20:11.59 | merlin{work} | hi gonkulator |
| 20:12.04 | merlin{work} | wazup p33ps? |
| 20:12.08 | merlin{work} | lol |
| 20:12.13 | darth_iBook | gonkulator: guess what |
| 20:12.22 | gonkulator | darth_iBook: wazzat? |
| 20:12.34 | darth_iBook | gonkulator: I've decided to learn opengl |
| 20:12.41 | gonkulator | darth_iBook: cool |
| 20:12.45 | darth_iBook | however, it is hating me right now and won't work on my laptop |
| 20:12.48 | gonkulator | darth_iBook: I'll be learning openGL later this semester |
| 20:13.03 | darth_iBook | piping through that stupid java program for class doesn't work |
| 20:13.12 | darth_iBook | so I'm just gonna use opengl if I can figure it out |
| 20:13.17 | merlin{work} | I'm pondering upgrading this server |
| 20:13.18 | gonkulator | yeah |
| 20:13.22 | merlin{work} | only an AMD K6-2 |
| 20:13.26 | gonkulator | and Java's openGL implementation blows ass |
| 20:13.31 | merlin{work} | but it seems to get the job done |
| 20:14.26 | merlin{work} | Is there any Java plugin for MOzilla 1.2? |
| 20:14.32 | merlin{work} | Arggghhhh, I can't get it working. :( |
| 20:15.47 | *** join/#picogui plfiorini (~kopeteuse@62.211.165.112) |
| 20:16.03 | plfiorini | hi |
| 20:17.11 | merlin{work} | hi plfiorini |
| 20:17.22 | merlin{work} | Where'd the bot go? |
| 20:17.24 | gonkulator | darth_iBook: I hope you enjoy your stuff |
| 20:17.30 | merlin{work} | PicoBot: seen scanline |
| 20:17.31 | darth_iBook | ? |
| 20:17.34 | gonkulator | hi plfiorini |
| 20:17.44 | merlin{work} | bleep! |
| 20:17.55 | merlin{work} | ibot: seen scanline |
| 20:17.56 | | scanline is currently on #picogui (6h 27m 15s) #openembedded (6h 27m 15s) #elinux (6h 27m 15s) #bzflag (6h 27m 15s). Has said a total of 48 messages. Is idling for 3h 28m 15s |
| 20:17.56 | darth_iBook | gonkulator: what stuff? |
| 20:18.02 | gonkulator | darth_iBook: openGL |
| 20:18.06 | darth_iBook | oh, right |
| 20:18.12 | darth_iBook | if I can ever make it work |
| 20:18.45 | merlin{work} | OpenGL is easy, and fun |
| 20:18.47 | merlin{work} | =D |
| 20:18.54 | darth_iBook | it won't compile on my iBook |
| 20:19.00 | darth_iBook | it keeps giving me a linking error |
| 20:19.02 | merlin{work} | Drawing, however, is hard. |
| 20:19.04 | merlin{work} | =( |
| 20:19.30 | darth_iBook | now you have some |
| 20:20.30 | merlin{work} | w00t! |
| 20:20.50 | darth_iBook | lol |
| 20:20.52 | merlin{work} | where's Xentac? |
| 20:21.36 | darth_iBook | lol |
| 20:21.58 | gonkulator | darth_iBook: thats some of the best irony I have ever done |
| 20:22.15 | darth_iBook | ? |
| 20:22.22 | darth_iBook | kicking my ass if I don't give you fighting skills? |
| 20:22.30 | gonkulator | thats it |
| 20:22.30 | darth_iBook | lol |
| 20:23.16 | file | I wish |
| 20:23.25 | gonkulator | they arrest people for stuff like that |
| 20:23.45 | gonkulator | look at what happened to Pee Wee Herman (the actor thereof) |
| 20:23.59 | plfiorini | hi gonkulator merlin{work} |
| 20:24.34 | merlin{work} | what's da matter file? |
| 20:25.08 | file | merlin{work}: I said "bah" because a person didn't ship my mobo and CPU yet, and they came back with about 5 sentences virtually yelling at me |
| 20:25.26 | merlin{work} | ugghhh |
| 20:25.29 | merlin{work} | bad people! |
| 20:25.32 | gonkulator | file: ever seen a duck get out of dirty water? |
| 20:25.43 | gonkulator | the duck seems to come out of the dirtiest water clean |
| 20:25.48 | gonkulator | well, this is just water off of a ducks back |
| 20:26.11 | merlin{work} | welp, I've gotta get back to work |
| 20:26.15 | gonkulator | later merlin |
| 20:26.15 | merlin{work} | see yall in a couple of hours |
| 20:26.16 | merlin{work} | =D |
| 20:41.17 | darth_iBook | damned opengl |
| 20:41.20 | darth_iBook | why can't I figure it out? |
| 20:41.35 | darth_iBook | does anyone know any good tutorials? |
| 20:42.36 | file | peeps |
| 20:44.04 | gonkulator | darth_iBook: captain_proton knows a lot of openGL stuff |
| 20:44.47 | captain_proton | hm |
| 20:44.49 | captain_proton | tutorials.... |
| 20:45.02 | captain_proton | look at the stuff at http://nehe.gamedev.net - its as good as anything else |
| 20:45.21 | darth_iBook | thanks |
| 20:48.03 | darth_iBook | hmmm... the mac os x tutorial there is for cocoa |
| 20:48.11 | darth_iBook | I need one for c++ |
| 20:49.13 | file | mmm... chocolate coca |
| 20:49.15 | file | er cocoa |
| 20:49.23 | gonkulator | objective C is as good as C++ :) |
| 20:49.25 | gonkulator | darn |
| 20:49.28 | gonkulator | later darth |
| 20:57.07 | gonkulator | dak |
| 20:57.08 | gonkulator | ok |
| 20:57.10 | gonkulator | if you have to |
| 21:07.33 | gonkulator | hmm... |
| 21:07.42 | gonkulator | my professor said "fuck" when we said his code doesn't work |
| 21:07.51 | *** join/#picogui Vivyan (~liquidang@wireless-242.cs.colorado.edu) |
| 21:07.51 | file | lol |
| 21:08.25 | file | gonkulator: what was his code supposed to do and why wouldn't it work? |
| 21:11.48 | *** join/#picogui Tader (~thomas@stan.xs4all.nl) |
| 21:16.07 | gonkulator | file: it technically isn't his fault |
| 21:16.09 | gonkulator | but still |
| 21:16.22 | file | details!!! |
| 21:24.31 | *** join/#picogui Tader (~tader@stan.xs4all.nl) |
| 21:26.34 | plfiorini | just a question... how can i create a new window? should i create a new application? (i am using python bindings) |
| 21:29.30 | *** join/#picogui Vivyan (~liquidang@wireless-242.cs.colorado.edu) |
| 21:29.52 | Vivyan | blah |
| 21:30.35 | scanline | ~particle du jour |
| 21:30.36 | | inverse pseudoneutron |
| 21:31.48 | Tader | goodmorning picogui! :P |
| 21:32.31 | Tader | ibot: hi |
| 21:32.31 | | niihau |
| 21:38.46 | *** join/#picogui Vivyan (~liquidang@wireless-242.cs.colorado.edu) |
| 21:39.12 | Vivyan | grrrrrr |
| 21:40.59 | Tader | :P |
| 21:41.40 | Vivyan | stupid wireless kicking me off |
| 21:41.57 | Tader | Wireless is cool :) |
| 21:42.14 | Vivyan | not when it kicks me off all the time |
| 21:42.21 | Tader | Hehe... |
| 21:43.06 | Tader | Not when it's not installed... But hey I'm a student at a technical university... That's the last place in the world you'll expect an 802.11 access point... :s |
| 21:46.49 | Vivyan | time to go, class over see you |
| 22:38.42 | gonkulator | pl_suicide: whats up? |
| 22:38.53 | gonkulator | why for is your name now pl_suicide? |
| 22:39.59 | gonkulator | yay for column vectors |
| 22:41.09 | gonkulator | captain_proton: I must have |
| 22:41.13 | gonkulator | captain_proton: care to fill me in? |
| 22:41.17 | captain_proton | so this guy works at an aquarium |
| 22:41.17 | captain_proton | and his boss comes to him and tells him that the dolphins are feeling very amorous |
| 22:41.17 | captain_proton | which is a problem, because they have a bunch of third graders coming in an hour |
| 22:41.17 | captain_proton | so she tells him to go get some baby seagulls, because they're the only anti-aphrodesiac for dolphins |
| 22:41.17 | captain_proton | "but be careful!" she says, "a lion escaped from the zoo this morning, and while heavily tranquilized, is still very dangerous" |
| 22:41.19 | captain_proton | so this guy goes out to the beach and picks up some baby seagulls |
| 22:41.21 | captain_proton | but when coming back, he finds the lion, lying across the sidewalk |
| 22:41.23 | captain_proton | as he gingerly steps over it, he gets arrested |
| 22:41.25 | captain_proton | for transporting young gulls over a sedate lion for immoral porpoises |
| 22:42.31 | gonkulator | captain_proton: I have seen worse |
| 22:42.37 | gonkulator | captain_proton: when is a mole a mule? |
| 22:42.52 | gonkulator | captain_proton: when it burros |
| 22:44.12 | gonkulator | this computer graphics course has made me wish I payed better attention in linear algerbra |
| 22:44.19 | captain_proton | heh |
| 22:44.21 | gonkulator | twice now |
| 22:45.09 | gonkulator | paid sucks? |
| 22:45.23 | gonkulator | that could be really bad... |
| 22:48.01 | gonkulator | my prof reminds me of Dan |
| 22:48.04 | gonkulator | er, lurgyman |
| 22:49.15 | gonkulator | well, class is over |
| 22:53.00 | scanline | we were talking in here a bit, and she just left for class |
| 22:53.18 | scanline | choir class |
| 22:53.30 | captain_proton | ah |
| 22:53.30 | scanline | gonkulator: michael is being an ass whenever jen's around |
| 22:54.29 | scanline | hehe |
| 22:54.59 | scanline | And now for the thrilling conclusion of "Further adventures of Captain Proton in IRIX land" |
| 23:01.09 | njs | scanline: I'm not sure that I understand your "virtual machine RPC layer" idea |
| 23:01.30 | scanline | what about it? |
| 23:02.13 | njs | scanline: is it, "we invent an itty-bitty language L, and then send bits of L code down the wire. L code is powerful enough to run a command, save the result in a variable, make invocations on that variable, etc., all within the server (i.e., latency free)"? |
| 23:02.36 | scanline | yeah, pretty much |
| 23:02.41 | njs | ah, okay |
| 23:02.46 | njs | convergent evolution of ideas is cute |
| 23:03.22 | scanline | I'd like a simpler solution, but I really don't want to have to expose the ID generation algorithm to the client |
| 23:03.25 | njs | (I was rolling around some ideas like that a bit ago, probably letting L = a baby Lisp) |
| 23:04.06 | njs | do you know how E promises make it work? I've been meaning to track down more technical documentation/read the code for that, but haven't had a chance yet. |
| 23:04.24 | scanline | I haven't read the docs for E's wire protocol yet |
| 23:04.26 | njs | E promises seem a much simpler and more elegant way to avoid latency than a baby language |
| 23:04.54 | njs | oh, you say this in the next line of your email :-) |
| 23:05.11 | scanline | I don't see how they avoid latency- if you create an object and then set properties in that object, you still have to wait for a round trip, it just doesn't block the rest of your program |
| 23:06.41 | njs | hmm, have I managed to get confused about how promises work? |
| 23:07.01 | scanline | hmm.. or can you call an object's method using a promise rather than a reference? |
| 23:07.18 | scanline | I thought it had to resolve into a reference first |
| 23:07.47 | njs | I'm not sure. I thought you could, but this might be because that's how I'd design it, not because they actually did design it that way. |
| 23:08.22 | scanline | if that was the case, then making a call from a promise would be equivalent to pushing that call onto a queue of operations for the remote machine to execute |
| 23:08.32 | njs | nod |
| 23:08.44 | scanline | so you could still write the wire protocol as a tiny language, and just use that as a tool to make the client interface easy to understand |
| 23:09.30 | scanline | if expressions were evaulated using both references and promises, those would have to be sent as code also |
| 23:09.31 | njs | well, the virtual machine solution will certainly be able to emulate any other solution, that's why languages are cool; the question is whether you actually need the complexity |
| 23:09.53 | scanline | now here's another thought- |
| 23:10.04 | scanline | if you're putting a tiny VM into the protocol, you can reuse that for other things |
| 23:10.29 | scanline | for example, a widget template could simply be a chunk of data already written in the wire protocol, that uses this language to glue everything together |
| 23:10.41 | njs | you know, a simple way to make the clients agnostic wrt how ids are generated is to let clients request them from the server in bulk. |
| 23:11.04 | scanline | simple yes.. rather inelegant :) |
| 23:11.46 | njs | well, it's the obvious and generic way to keep id allocation policy in the server :-) |
| 23:12.12 | njs | yes, languages are very general things. Let's resurrect NeWS while we're at it :-) |
| 23:13.13 | njs | I was talking to a friend about some of this stuff, and we realized that the natural, glaringly obvious thing to use is, of course, Display Lisp! |
| 23:13.15 | scanline | well, there are a couple practical reasons I don't like bulk allocation |
| 23:13.52 | scanline | I'm still envisioning picogui clients that can run in on the order of 32 bytes of RAM |
| 23:14.02 | scanline | :) |
| 23:14.21 | njs | hmm. and that need pipelining for performance? :-) |
| 23:14.54 | scanline | well, you could have it request only one object ID- but there goes half your RAM, and you have to wait on that for every object you create |
| 23:15.03 | njs | one thing about a VM in the wire protocol is that it'd take some work to figure out how to encode it efficiently |
| 23:15.14 | scanline | true, but that'd be up to the client |
| 23:15.33 | scanline | It's important to me that the wire protocol be like checkers- easy to learn, hard to master ;) |
| 23:16.20 | scanline | the more widely used client libs like C/C++ and Python could be smart enough to encode their requests in an optimal way |
| 23:17.00 | scanline | clients that are really low on memory or bandwidth could even keep their object IDs in the languages stack/registers and reference them from there |
| 23:18.46 | scanline | s/languages/language's/ |
| 23:19.25 | njs | That brings up an interesting point -- another problem with languages is that they're hard to sandbox. |
| 23:20.09 | scanline | well, this one would be damn simple- and using capabilities surely wouldn't hurt |
| 23:20.30 | scanline | for example, branching probably wouldn't be allowed |
| 23:20.48 | njs | protecting against memory exhaustion isn't _that_ hard, especially since you want some sort of general quota system anyway, but designing your language so that infinite loops don't hurt might be tricky. |
| 23:21.09 | njs | (though it depends on how powerful your language is -- if you have no flow control, it wouldn't be as much of a problem :-)) |
| 23:21.14 | scanline | :) |
| 23:21.29 | scanline | A pattern a lot like what we use now for the theme system might work nicely |
| 23:22.01 | scanline | the theme language is stack based, with a fixed size stack- has operators to manipulate the stack, perform lookups in the theme database, do simple arithmatic, and a few special operators for manipulating colors |
| 23:22.21 | njs | (but flow control is necessary for any sort of generality -- it'd be really _nice_ to be able to compress a few dozen equivalent commands (e.g., "unbind; delete" for each widget in a window being closed) into a single loop that you send across and that then gets executed on the server) |
| 23:23.09 | scanline | the server could prove the program will terminate before running it ;) |
| 23:23.43 | scanline | having no flow control in the theme language made a lot of things easy- for example, you could statically determine how much stack space the function used |
| 23:23.46 | njs | yeah, proof-carrying code is fun, but I haven't been able to find any references to actually making it practical at the moment, and inventing the tools from scratch could be time consuming :-) |
| 23:23.52 | scanline | indeed |
| 23:24.17 | scanline | and that still wouldn't prove that it doesn't call some function that allocates memory 2 billion times |
| 23:24.28 | scanline | though quotas would fix that I suppose |
| 23:24.32 | njs | (have we talked about this before, or does it count as more slightly-eerie convergence?) |
| 23:24.43 | Xentac | gonkulator: why do you want fighting skills? |
| 23:24.46 | Xentac | ibot: seen gonkulator |
| 23:24.47 | | gonkulator is currently on #picogui (5h 48m 20s). Has said a total of 60 messages. Is idling for 35m 32s |
| 23:24.59 | gonkulator | hmm... |
| 23:25.04 | scanline | njs: probably eerie convercence between this and the theme language... I think I remember talking with you about the theme code before |
| 23:25.05 | gonkulator | I must have said something before that |
| 23:25.11 | gonkulator | Xentac: I was making a joke |
| 23:25.18 | Xentac | just curious |
| 23:25.26 | gonkulator | Xentac: darth_balls has lots of training in that area |
| 23:25.26 | Xentac | I noticed that merlin262 gave someone artistic skills... |
| 23:25.35 | njs | yeah, if you have one VM per connection (which you presumably would, since you don't want messages from different clients stomping on each other's stacks), then you just make your VM use your pool-allocator, and everything's easy. |
| 23:25.35 | Xentac | then you said you wanted fighting skills |
| 23:25.47 | scanline | njs: yup |
| 23:25.47 | gonkulator | right |
| 23:25.59 | Xentac | wait... darth gave merlin262 artistic skills... |
| 23:26.03 | gonkulator | that way I could tell darth_iBook that I would kick his ass if he didn't give them to me |
| 23:26.05 | Xentac | then you wanted the fighting skills |
| 23:26.17 | Xentac | and for some reason merlin262 asked where I was ;o) |
| 23:26.19 | Xentac | (that's why I read it) |
| 23:26.23 | gonkulator | which is the ultimate in empty threats :) |
| 23:26.28 | Xentac | hehehe |
| 23:27.22 | scanline | njs: capabilities are nice.. even if there was a filesystem object in the kernel (and there presumably might be, for things like the app-server) and the VM could make calls to object methods, it would be secure as long as the client could not get access to a filesystem object instance :) |
| 23:27.51 | scanline | s/filesystem/file/ |
| 23:28.42 | njs | scanline: yeah, capabilities are wonderful, not because they're such a huge increase in security or something, but because the resulting system is _so_ easy to reason about |
| 23:28.54 | scanline | so much simpler than ACLs :) |
| 23:29.20 | scanline | I should still do some math wrt the 128-bit object IDs though |
| 23:29.25 | njs | nod |
| 23:29.48 | scanline | >>> 2**128 |
| 23:29.48 | scanline | 340282366920938463463374607431768211456L |
| 23:29.49 | scanline | :) |
| 23:29.54 | njs | I know GUID/UUID's use 128-bits, and don't even bother to preserve all of them |
| 23:30.58 | njs | I tried calculating the collision probability at some point, but done the natural way you end up taking (2**128)! or something, which Maple refused to do. |
| 23:31.24 | scanline | haha |
| 23:32.20 | scanline | if you had 1 million 1GHz computers generating a new 128-bit number each clock cycle for one year, you would only have used 1/1079028307080601th of the key space |
| 23:32.41 | njs | I think there's a cleverer iterative way, but never got around for it. |
| 23:33.24 | njs | right, 128 bits is plenty :-) (and in two years that'll be two words...) |
| 23:34.08 | scanline | on some machines |
| 23:34.11 | njs | you can probably get away with a pretty small key size for these applications, though, since you can seriously penalize wrong guesses |
| 23:34.21 | lurgyman | hmm |
| 23:34.30 | lurgyman | captain_proton seems to have gone to sleep |
| 23:34.50 | scanline | njs: I'd like the protocol to be such that the client can work with whatever key size the server generates |
| 23:35.13 | njs | if the server has a "1 bad capability request and you get kicked out" policy... |
| 23:35.39 | scanline | the client could just reconnect though |
| 23:35.55 | scanline | at least it would prevent you writing a script in this VM to randomly guess keys ;) |
| 23:36.05 | njs | that does make sense; key sizes are notoriously prone to change. |
| 23:36.13 | njs | makes protocol design nastier, though :-( |
| 23:36.23 | njs | heh, that it would. |
| 23:36.51 | scanline | well, 128 bits is too big to consider a simple data type anyway, so you'd want to consider it more like a string |
| 23:36.54 | njs | if you make people pause 10us while connecting, that would work too |
| 23:37.24 | njs | nod |
| 23:37.37 | njs | of course, everything's like a string when you're pushing and pulling stuff off the wire :-) |
| 23:37.48 | scanline | true |
| 23:37.51 | scanline | unless it's SHM :) |
| 23:38.03 | njs | heh, right :-) |
| 23:38.24 | scanline | but more importantly, descbibing the key as a string rather than a number avoids endian issues |
| 23:38.44 | njs | but you, uh, probably shouldn't be leaving capabilities lying around in SHM, eh? |
| 23:38.53 | scanline | hehe, probably not |
| 23:39.04 | scanline | err, security |
| 23:39.31 | njs | you could make SHM more secure by using the "mmap /dev/zero + offset" approach, and treat the offset as a capability! |
| 23:39.46 | scanline | njs: hahaha |
| 23:39.50 | njs | ... limited by your machine's word size, of course :-) |
| 23:39.55 | scanline | and page size |
| 23:40.05 | njs | nod |
| 23:40.11 | scanline | 20-bit capability |
| 23:40.16 | Xentac | lurgyman: what code are you writing? |
| 23:41.42 | lalo | scanline: so, what's new? :-) |
| 23:41.53 | scanline | lalo: discussing the VM idea with njs |
| 23:42.11 | lalo | I noticed, but was too busy with work to follow |
| 23:42.25 | lalo | me too. |
| 23:42.31 | lalo | current one is sketchy |
| 23:42.52 | lalo | just one more thing I copied from CPython |
| 23:42.55 | scanline | well, the strings as return codes would work if each string always had the same address... but the address will be different in each file |
| 23:43.09 | scanline | plus I'd really like to support backtraces |
| 23:43.32 | lalo | backtraces are definitely in the wishlist |
| 23:44.02 | lalo | man, the specs are what counts, feel free to throw away anything you don't like about the code :-P |
| 23:44.04 | scanline | It would probably be in the form of a macro you'd be forced to call at the beginning of each function, that would push an identifier for that function onto a stack |
| 23:44.06 | scanline | or something :) |
| 23:44.12 | scanline | I know |
| 23:44.20 | scanline | I read the specs, just reading the code too :) |
| 23:44.25 | lalo | current excuse for "code" is a prototype |
| 23:44.41 | lalo | we could investigate how CPython gets backtraces |
| 23:45.13 | lalo | altough I suspect functions written in C don't participate on the backtrace machinery... which would be a bit useless for us |
| 23:45.58 | scanline | of course, pushing a function identifier on to the stack at every function call is bad overhead |
| 23:46.25 | scanline | I'd rather have a platform-specific but optional way to walk the stack at the point the backtrace is needed |
| 23:46.44 | lalo | I figured string exceptions were interesting because if you're in the same process it's efficient (pointer compare) but if you're not it's flexible (strcmp) - but I'm not very sure how to implement this |
| 23:47.39 | njs | libbtrace? |
| 23:47.41 | lurgyman | captain_proton has promised to write an interface for mandolin once it's rewritten |
| 23:47.41 | lurgyman | mwa ha |
| 23:48.03 | lalo | in short |
| 23:48.03 | lalo | starting a project like this sucks :-) |
| 23:48.09 | njs | lalo: making pointer compares work across shared library boundaries is seriously hard. |
| 23:48.29 | lalo | for a few weeks you have to get your head to think in high and low level at the same time, if you want the design to make sense |
| 23:49.37 | scanline | from what I understand of how ld works, identical strings in separate files will have different addresses |
| 23:49.47 | njs | yes |
| 23:49.54 | scanline | you'd want one file that defines all the strings, then the header reference them all with extern |
| 23:50.06 | scanline | but then you still have the shared lib problem |
| 23:50.40 | scanline | you could use an ID-based system like picogui does now, but add a way to allocate new IDs (same way you'd get new IDs for types, widgets...) |
| 23:50.53 | scanline | but I'd like to be able to have non-static messages |
| 23:51.01 | njs | one of the areas that's seen a lot of thrashing in g++ over the last few releases is exactly this problem -- certain objects in the C++ runtime are supposed to exist exactly once in the process image, and they went through a number of revs trying to get it right |
| 23:51.07 | scanline | things like "protocol error, expected byte 0x%02X instead of 0x%02X" |
| 23:51.15 | lalo | the only concrete thing I really like about the current prototype code is the translator (get/set) api |
| 23:52.19 | scanline | I should probably take another nap now if I'm going to have the energy to do homework... |
| 23:52.21 | scanline | bbiafh |
| 23:52.27 | lalo | wait a bit |
| 23:52.30 | scanline | ok |
| 23:52.40 | lalo | I have to go too, so let's close this subject in a few minutes :-) |
| 23:52.58 | scanline | I looked at the 'translator' code but wasn't sure what it was supposed to do |
| 23:53.27 | lalo | suppose the picogui kernel is a translator. |
| 23:53.43 | lalo | it manages a collection of objects (the widgets and their properties) |
| 23:54.06 | njs | oh, and making exception handling thread-safe is another area that I hear is lots and lots of fun. Just to warn you. |
| 23:54.10 | lalo | when anyone tries to get or set something on them, this is at low-level done by a function implemented by the translator |
| 23:54.26 | njs | is your translator ~~ what the HURD calls a translator? |
| 23:54.38 | lalo | njs: I know. it's one of the few areas where I think python kludged. |
| 23:55.04 | lalo | remotely similar, at least similar enough that I felt it fair to borrow the word |
| 23:55.40 | lalo | what both have in common is that they are called to get and set things on the objects they manage... and they may use this, if they wish, to implement persistence |
| 23:56.13 | scanline | this is a heck of a lot like what I did with JetCOW in jetstream :) |
| 23:56.17 | lalo | I think the exception thing has three aspects |
| 23:56.30 | lalo | (a) how to signal the fact that an exception happened |
| 23:57.04 | lalo | I think it's probably fine to return or not return NULL... then it would be a matter of what to return |
| 23:57.10 | lalo | (b), how to backtrace |
| 23:57.37 | lalo | (c), where to store the exception info; the current sollution is to use the return value (store in the stack) |
| 23:58.18 | lalo | as to what an exception is... instead of strings... it would probably make sense to go all the way to oop and python, and make exceptions real objects |
| 23:58.33 | scanline | how about exceptions causing an exception object to be created, which is network transparent? |
| 23:58.41 | lalo | this would make the catch clauses slower, I figure... |
| 23:58.54 | lalo | heh, that's exactly what I'm thinking :-) |
| 23:59.20 | scanline | well, exception speed isn't a big deal IMHO, since by definition exceptions don't normally happen |
| 23:59.31 | lalo | in fact this could be done keeping pretty much the current code - just instead of returning the exception text in the string, return its om id |
| 23:59.48 | lalo | "exceptions are by definition exceptional" - Guido |
| 23:59.53 | scanline | so explain the Om id? |
