| 00:00.44 | brlcad | shape's direction normal |
| 00:00.49 | brlcad | but yeah |
| 00:01.38 | pacman87 | i'm just thinking 2D to keep it simple for now |
| 00:01.50 | pacman87 | get the terminology straight |
| 00:06.44 | *** join/#brlcad iraytrace (n=iraytrac@cocoa.sci.utah.edu) | |
| 00:06.58 | starseeker | Howdy iraytrace :-) |
| 00:07.13 | iraytrace | Good evening! |
| 00:07.20 | starseeker | How's life? |
| 00:07.25 | iraytrace | Fun! |
| 00:07.35 | starseeker | Getting into trouble? ;-) |
| 00:07.36 | iraytrace | I'm a bachelor this weekend! |
| 00:07.43 | iraytrace | Lots o' trouble. |
| 00:07.51 | iraytrace | Doing ray-tracing homework assignemtn. |
| 00:07.57 | iraytrace | -ment |
| 00:08.08 | starseeker | fun |
| 00:08.46 | iraytrace | Mostly. There's a certain element of "how many times in my life do I have to implement perlin noise?" that is Meh. |
| 00:09.05 | starseeker | mmm. |
| 00:09.09 | starseeker | I can see that |
| 00:10.13 | iraytrace | Anything good on your end? |
| 00:10.45 | starseeker | tangling with implementing matrix logic |
| 00:10.54 | *** join/#brlcad Elperion (n=Bary@p5487512E.dip.t-dialin.net) | |
| 00:11.02 | starseeker | nothing fancy, but good for getting up to speed on C again |
| 00:11.49 | iraytrace | I'm going the opposite direction. Everything is in C++ for the class here. |
| 00:12.52 | starseeker | Just doing a proc-db, so C should be enough |
| 00:12.56 | *** join/#brlcad Twingy (n=justin@74.92.144.217) | |
| 00:13.42 | iraytrace | proc-db's are always fun. |
| 00:15.43 | pacman87 | "mged> exec fbserv 1 /dev/sgip &" fails with "no such device '/dev/sgip'" |
| 00:16.00 | iraytrace | Are you on an SGI? |
| 00:16.25 | starseeker | Cool, here we go - this dude has both side view and cross sections of an engine: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=110236614029 |
| 00:16.38 | *** join/#brlcad Elperion (n=Bary@p5487512E.dip.t-dialin.net) | |
| 00:16.43 | iraytrace | pacman87: You probably want /dev/X |
| 00:17.13 | yukonbob | has been looking at airplanes a lot lately (having been flying too much, and also because the 50th aniversary of the Canadian Avro Aero happened this week)... |
| 00:17.16 | pacman87 | iraytrace: thanks :) |
| 00:18.58 | iraytrace | Blueprints are cool, but the model I would build from it is available online for $15 |
| 00:19.56 | starseeker | True :-) |
| 00:20.09 | starseeker | likes the combination of history and technical geekery |
| 00:21.25 | iraytrace | It is sweet. |
| 00:27.09 | starseeker | yukonbob: Heh - wonder if a propeller could be modeled with our primitives |
| 00:28.41 | yukonbob | well, the (dare I say it) havoc has props... I haven't inspected them, though |
| 00:29.07 | yukonbob | should spend some time and get a model of that famous (around here) impeller modelled. |
| 00:29.16 | iraytrace | With the new spline support it should be possible |
| 00:29.35 | pacman87 | iraytrace: :) |
| 00:29.42 | starseeker | that could be cool yukonbob :-) |
| 00:29.56 | yukonbob | there's a primitive called sketch, right? (/me doesn't have his manuals around, nor a copy of brl-cad)? |
| 00:30.03 | starseeker | then for bonus points have gcam take the model and make one ;-) |
| 00:30.22 | yukonbob | starseeker: do we have access to a CAM device? |
| 00:30.33 | starseeker | I don't |
| 00:30.35 | yukonbob | (ie: CNC, i guess) |
| 00:31.24 | yukonbob | starseeker: have you seen the video-clip of that 4-axis device making that impeller? |
| 00:31.40 | starseeker | is that the one you linked on youtube? |
| 00:31.59 | yukonbob | I wasn't the original linker to it, but I believe it was off youtube, ya.. |
| 00:32.22 | starseeker | that was neat |
| 00:32.25 | yukonbob | I linked to a frame from it, which gives a good detail of the complexity of the shape, but watching the process is really impressive. |
| 00:36.54 | starseeker | brlcad: did you see a compile warning about macro "DM_OPEN" passed 4 arguments, but takes just 3? |
| 00:38.49 | pacman87 | anyone know any good reference books for solid modelling algorithms? |
| 00:45.09 | starseeker | starts digging into DM_OPEN, then breaks for supper... |
| 00:47.35 | *** join/#brlcad iraytrace (n=iraytrac@c-67-172-239-24.hsd1.ut.comcast.net) | |
| 01:45.48 | *** join/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 02:21.09 | *** part/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 02:21.33 | *** join/#brlcad jra (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 02:22.47 | brlcad | heya Daytona |
| 02:25.04 | Daytona | hi |
| 02:26.01 | Daytona | Looks like the GSoc application deadline might get extened for a week |
| 02:29.15 | starseeker | is a bit confused - the only definition of dm_open grep finds, in dm.h, seems to have 3 args. All the uses that compile correctly have 4, and none seems to include dm.h directly. However, mged_dm.h does include dm.h... |
| 02:29.23 | brlcad | Daytona: yeah, it's really odd |
| 02:29.40 | brlcad | the rate of students is distinctly different from this time last year |
| 02:29.55 | brlcad | we were talking about it in the gsoc irc channel earlier today before she sent out the message |
| 02:30.43 | Daytona | I guess the early Easter my be the cause |
| 02:30.54 | brlcad | most of the returning projects across the board are seeing drastically lower rates (like 1/5th usual) -- the comment from that one guy about easter and holidays is probably spot on |
| 02:31.06 | brlcad | there have been fewer US/EU visitors |
| 02:31.10 | starseeker | and mged_dm.h seems to be the only .h file to also include dm.h |
| 02:31.45 | brlcad | also last year, submission week was a week earlier, went for two weeks instead of one, and easter was much later of course |
| 02:33.42 | Daytona | starseeker: are you seeing this error: pl-dm.c:1040:48: error: macro "DM_OPEN" passed 4 arguments, but takes just 3 |
| 02:33.49 | starseeker | yes |
| 02:33.51 | brlcad | last year there were 6300 total submissions, we're at roughly 1600 this year with three days to go -- so if 4k don't flood in this weekend, it'll probably get extended I'd imagine |
| 02:34.03 | brlcad | yeah, that's my bad |
| 02:34.06 | brlcad | fixes |
| 02:34.21 | brlcad | pl-dm.c was disabled (since we went to autoconf) |
| 02:34.35 | brlcad | just re-enabled it last night .. worked here, but must have missed some things |
| 02:35.04 | starseeker | is somewhat confused as to why other dm_open calls ARE working... must be one of those obvious things |
| 02:35.13 | Daytona | yeah, I just noticed it tonight |
| 02:35.52 | brlcad | huh |
| 02:36.00 | brlcad | must be another DM_OPEN .. |
| 02:36.09 | starseeker | Not that grep can find |
| 02:36.09 | brlcad | read/fixed based on include/dm.h #define DM_OPEN(_interp, _type, _argc, _argv) dm_open(_interp, _type, _argc, _argv) |
| 02:37.04 | starseeker | #define DM_OPEN(_type, _argc, _argv) dm_open(_type, _argc, _argv) is what I'm seeing... |
| 02:37.32 | brlcad | ooh, I see -- missed committing the header since it is in include/ |
| 02:37.49 | starseeker | Ah :-) |
| 02:38.25 | brlcad | that line is wrong in dm.h -- the interp was added years ago, but pl-dm is the only user |
| 02:38.47 | CIA-33 | BRL-CAD: 03brlcad * r30589 10/brlcad/trunk/include/dm.h: need the interp |
| 02:41.17 | Daytona | brlcad: hate to be a pain, but now: rlcad/src/util/pl-dm.c:1046: undefined reference to `dm_configureWindowShape' |
| 02:41.29 | brlcad | eep |
| 02:42.37 | Daytona | brlcad: I'm not in any rush to get it compiled |
| 02:42.48 | brlcad | still, my mistake -- i'll fix |
| 02:43.51 | CIA-33 | BRL-CAD: 03brlcad * r30590 10/brlcad/trunk/src/util/Makefile.am: turn pl-dm off until I can get a full clean build to verify all flags are set |
| 02:44.56 | hippieindamakin8 | hey brlcad and pacman87 and others :) |
| 02:45.01 | Daytona | tgat fixed it :-) |
| 02:45.09 | brlcad | :) |
| 02:45.25 | brlcad | could probably rip the whole thing out, but it's the smallest libdm example (albeit weak) |
| 02:46.13 | Daytona | hippieindamakin8: hi |
| 02:50.48 | hippieindamakin8 | how reliable is this information that the deadline might be extended for the GSoC applications |
| 02:51.13 | brlcad | hippieindamakin8: it entirely depends how many apps are received over the weekend so I wouldn't rely on it |
| 02:51.34 | brlcad | if you procrastinate and don't submit, there's no recourse |
| 02:51.36 | hippieindamakin8 | even i wouldnt :) |
| 02:52.19 | Daytona | I believe the decision won't be made till Monday |
| 02:53.13 | hippieindamakin8 | hey i have been going through the codes man.. OO programming has cut down the code length too :P |
| 02:54.10 | brlcad | it's not complete, it's a start |
| 02:54.46 | hippieindamakin8 | yaya i have been chking individial codes and comparing them' |
| 02:54.55 | hippieindamakin8 | for example Arbs |
| 02:57.40 | hippieindamakin8 | composed a new song last night :) |
| 03:03.51 | *** part/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 03:09.39 | *** join/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 03:09.51 | CIA-33 | BRL-CAD: 03brlcad * r30591 10/brlcad/trunk/src/proc-db/tire.c: started in 2008, new header format |
| 03:16.41 | CIA-33 | BRL-CAD: 03brlcad * r30592 10/brlcad/trunk/src/util/pl-dm.c: don't need the glx headers and dm_configureWindowShape no longer exists |
| 03:36.00 | starseeker | brlcad: Oops, sorry |
| 03:44.53 | CIA-33 | BRL-CAD: 03brlcad * r30593 10/brlcad/trunk/src/util/Makefile.am: try turning pl-dm back on.. works here clean-built |
| 03:45.02 | brlcad | nothing to be sorry about |
| 03:57.53 | starseeker | Confirmed - builds here too |
| 03:58.21 | brlcad | cool |
| 03:59.19 | Daytona | brlcad: builds clean here, too (on ubuntu) |
| 03:59.43 | brlcad | cool |
| 04:00.19 | Daytona | brlcad: is my nick showing up at the start of my messages? |
| 04:00.24 | brlcad | Daytona: so I forget where things left off with ubuntu build .. does it work for you again? you'd upgraded libtool iirc |
| 04:00.35 | brlcad | 00:00 < Daytona> brlcad: is my nick showing up at the start of my messages? |
| 04:00.46 | brlcad | uses irssi |
| 04:01.01 | brlcad | similar to your bitchx |
| 04:01.09 | hippieindamakin8 | does irssi have proxy support ? |
| 04:01.21 | brlcad | ooh, you're on irssi now |
| 04:01.30 | Daytona | Thanks, I'm just trying irrsi, but my nick shows up blank here |
| 04:01.31 | brlcad | ~Daytona++ |
| 04:01.40 | brlcad | blank? |
| 04:01.54 | brlcad | sure it's not colored white or something? |
| 04:01.59 | *** mode/#brlcad [+o Daytona] by brlcad | |
| 04:02.09 | Daytona | likely |
| 04:02.16 | brlcad | do you see the @? |
| 04:02.20 | Daytona | yes |
| 04:02.25 | brlcad | but nothing after? |
| 04:02.28 | Daytona | yes |
| 04:02.41 | Daytona | like <@ > |
| 04:02.44 | brlcad | try hilighting/selecting it .. i'm guessing it's white |
| 04:02.57 | Daytona | you'r right |
| 04:02.57 | brlcad | or whatever the color of your background |
| 04:04.11 | Daytona | Back to ubuntu, yes I upgraded libtool to the latest release |
| 04:04.59 | Daytona | now using libtool 2.2 |
| 04:05.13 | brlcad | and everything then worked? |
| 04:05.27 | brlcad | hmm |
| 04:05.41 | Daytona | yes, but the defaul that is installed with ubuntu does not work |
| 04:05.55 | brlcad | nods |
| 04:06.16 | brlcad | I started working through that with one of the guys in here earlier in the week, to see if I could get a work-around |
| 04:07.00 | brlcad | I did trace down part of the problem -- they've modified it to strip out the -L linker paths, with messages to set -rpaths (which is wrong for pre-installed) |
| 04:11.47 | Daytona | Well, I know the night is just getting started for all you young coders, but it's late for this dinosaur |
| 04:12.04 | Daytona | Have a nice night, bye all |
| 04:12.26 | *** part/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 04:13.33 | brlcad | :) |
| 05:57.25 | pacman87 | brlcad:ping |
| 08:42.43 | *** join/#brlcad vedge (i=vedge@vedge.org) | |
| 08:47.24 | cosurgi | ok, so later probably I'll try to install libtool 2.2 |
| 09:02.29 | alex_joni | brlcad: maybe you can pass this to Daytona (if he get's back): Ubuntu's terminal by default has a white background, which makes irssi kinda odd to use. One of the first things I do is change the profile for terminal to use a *standard* white on black :) |
| 09:50.11 | hippieindamakin8 | brlcad or any body here .. can somebody tell me where are the gui files located ? |
| 10:04.01 | *** join/#brlcad elite01 (n=elite01@dslb-088-070-100-087.pools.arcor-ip.net) | |
| 10:58.25 | MinuteElectron | hippieindamakin8: binaries or source? |
| 10:58.39 | hippieindamakin8 | source |
| 10:58.48 | hippieindamakin8 | and can u tell me wat are the .1 files ? |
| 10:58.55 | MinuteElectron | no, sorry |
| 10:59.08 | hippieindamakin8 | ohh np :) |
| 11:00.59 | hippieindamakin8 | brlcad or d_rossberg are u there ? |
| 11:02.52 | hippieindamakin8 | can u tell me wat are the .1 files for example the file rtarea.1 in the rt folder of the sources |
| 11:02.58 | hippieindamakin8 | ? |
| 11:04.19 | alex_joni | sounds like a manpage name :D |
| 11:06.55 | hippieindamakin8 | ohk wat does it mean ? |
| 11:07.11 | hippieindamakin8 | they are in some scripting language i guess |
| 11:31.56 | *** join/#brlcad clock_ (n=clock@77-56-95-194.dclient.hispeed.ch) | |
| 12:19.53 | *** join/#brlcad elite01 (n=elite01@dslb-088-070-002-219.pools.arcor-ip.net) | |
| 12:40.17 | brlcad | cosurgi: nooooo, not yet :) |
| 12:40.25 | brlcad | you were my test bed to get default working :) |
| 12:44.33 | hippieindamakin8 | hey brlcad u back.. ? good morning |
| 12:45.18 | brlcad | morning |
| 12:45.47 | brlcad | man man |
| 12:46.18 | hippieindamakin8 | if i am not buggin u early in the morning :P tell me wat are these ascii troffs doing in the package and wat do they do |
| 12:46.28 | brlcad | man man :) |
| 12:46.47 | alex_joni | hippieindamakin8: 'man' is a system for documentation |
| 12:46.53 | alex_joni | comes from manual |
| 12:47.01 | hippieindamakin8 | ohh :P |
| 12:47.11 | hippieindamakin8 | u meant man as manual |
| 12:47.14 | hippieindamakin8 | :P |
| 12:47.28 | alex_joni | type it into a terminal 'man man' |
| 12:48.12 | hippieindamakin8 | it is manual on man :) |
| 12:49.31 | hippieindamakin8 | and brlcad i am unable to understand what do i put in the proposal |
| 12:50.05 | hippieindamakin8 | as in apart from the idea |
| 12:52.08 | brlcad | did you read the application guidelines? |
| 12:52.22 | brlcad | that should give you a big range for how to get started |
| 12:52.30 | hippieindamakin8 | yaya |
| 12:57.52 | hippieindamakin8 | incase i start rewriting the entire libraries in OOP language even the applications and programs dependant on those have to be rewritten.. |
| 12:58.22 | hippieindamakin8 | or it has to be written in such a way that nothing else get disturbed ? |
| 13:01.49 | brlcad | "rewriting" is most certainly not how I'd describe the task |
| 13:02.03 | brlcad | refactoring and reorganizing existing code is the underlying idea |
| 13:02.20 | hippieindamakin8 | ya rather porting it to the OOP form |
| 13:02.39 | brlcad | creating a new layer that leverages what is written as much as possible, cleaning it portions while working on the OO layer |
| 13:03.34 | brlcad | the jbrlcad interface was an "imported rewrite" simply because it's a different language -- for a C++ layer, it could utilize librt/libbn/libbu routines directly |
| 13:03.52 | hippieindamakin8 | ya |
| 13:10.49 | *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz) | |
| 13:11.04 | brlcad | g'morning d_rossberg |
| 13:11.06 | hippieindamakin8 | wat exactly do u xpect the student to do in this project ? |
| 13:11.14 | hippieindamakin8 | good morning d_rossberg |
| 13:11.20 | brlcad | hippieindamakin8: I expect them to do what they propose ;) |
| 13:11.30 | hippieindamakin8 | :)))) |
| 13:11.35 | brlcad | it's your job to propose something, not mine |
| 13:11.45 | brlcad | i already provided the general ideas |
| 13:11.54 | d_rossberg | waves from home |
| 13:11.57 | hippieindamakin8 | i am in a mood to change everything rt from the libraries into java |
| 13:11.57 | brlcad | you're supposed to refine those ideas |
| 13:12.02 | hippieindamakin8 | :P |
| 13:12.09 | hippieindamakin8 | ya |
| 13:12.21 | hippieindamakin8 | i shall be coming up with a refined idea |
| 13:13.01 | brlcad | d_rossberg: so I presume you've been watching the list -- the student rate is drastically lower this year for some reason (like 1/5th) probably due to the early easter, so there "may" be a student extension |
| 13:13.28 | hippieindamakin8 | there are many friends of mine who are planning to apply on the last day :) |
| 13:13.43 | brlcad | "Applications to PHP and Gentoo haven't been very encouraging so far - only 10 each" .. wow |
| 13:13.52 | hippieindamakin8 | so there might be a sudden rise in the applications on 31st |
| 13:13.58 | brlcad | gentoo in particular has more than a hundred by this time last year iirc |
| 13:14.06 | d_rossberg | brlcad: yes, i've read the mails |
| 13:14.32 | brlcad | debian guys reported something similar .. they're at about 20 and usually are 200-300 by now |
| 13:16.03 | brlcad | i thought that through most of the week too as cad and bz have been nearly identical interest, and bz's is *much* lower than last year (about 1/5th) |
| 13:16.12 | brlcad | turns out it's almost all orgs |
| 13:16.43 | d_rossberg | hippieindamakin8: your application should show us that you know what your are talking about, that means BRL-CAD but object oriented design too |
| 13:17.34 | hippieindamakin8 | ohk.. u mean i should specify my plan of action |
| 13:17.39 | d_rossberg | how do you think should an oo interface to a database look like, how should the ray-trace interface be implemented |
| 13:18.17 | hippieindamakin8 | ohk.. |
| 13:19.31 | d_rossberg | the database is a file, how would you handle multiple access to this file? (for example) |
| 13:19.40 | hippieindamakin8 | ohk.. |
| 13:20.44 | hippieindamakin8 | .. i always have been a c++ and java programmer mainly but its taking me time to understand how the system works.. given a problem i can solve that.. |
| 13:21.03 | hippieindamakin8 | i have been understanding the system |
| 13:27.04 | d_rossberg | has some guests today |
| 14:14.12 | hippieindamakin8 | brlcad how many applications did u recieve so far |
| 14:16.25 | hippieindamakin8 | brlcad and as far as multiply accessing the file is concerned how do u do it.. i have no idea,, i just read that u have to use some virtual base classes and stuff |
| 15:13.18 | *** join/#brlcad Twingy (n=justin@74.92.144.217) [NETSPLIT VICTIM] | |
| 16:15.22 | *** join/#brlcad Elperion (n=Bary@p548737C3.dip.t-dialin.net) | |
| 17:09.45 | *** join/#brlcad clock_ (n=clock@217-162-110-69.dclient.hispeed.ch) | |
| 17:14.10 | *** join/#brlcad mib_niv7mz6n (i=3c342c9b@gateway/web/ajax/mibbit.com/x-65471bfecec121f5) | |
| 17:14.16 | mib_niv7mz6n | Hello all |
| 17:26.33 | brlcad | hum hum |
| 17:26.46 | brlcad | hippieindamakin8: told you before, don't worry about the other apps, worry about yours... |
| 18:27.58 | pacman87 | brlcad: is the src/librt/g_*.c code used for the CSG/BREP too? |
| 18:32.24 | brlcad | pacman87: what do you mean? |
| 18:32.42 | brlcad | the g_files are the definitions of each primitive |
| 18:33.03 | brlcad | CSG is handled is a different layer above the primitives for ray-tracing |
| 18:33.11 | brlcad | s/is a/in a/ |
| 18:33.46 | pacman87 | so the CSG uses the rt calls to do its evaluations? |
| 18:34.08 | brlcad | when you have CSG operators regardless of the representation, you have evaluated and unevaluated CSG forms |
| 18:35.07 | brlcad | ray-tracing causes the CSG expressions to get evaluated as the shot-line is traversed and primitives segments are encountered |
| 18:35.20 | brlcad | what you're proposing has nothing to do (directly) with ray-tracing |
| 18:35.41 | brlcad | you're evaluating CSG operators on BREP surfaces going from unevaluated BREP to evaluated BREP |
| 18:36.35 | brlcad | CSG isn't a system that "does" anything -- they are simply combination operators (union, subtraction, difference) that describe how shapes interact spatially |
| 18:37.09 | brlcad | does that make sense? |
| 18:37.26 | pacman87 | mostly |
| 18:38.21 | brlcad | so take a simple case of two overlapping spheres that are intersected |
| 18:39.10 | pacman87 | an intersection would be where the ray has entered both spheres |
| 18:39.20 | brlcad | with implicit geometry (i.e. what we use for ray-tracing), the CSG expression of "sphere1 intersected with sphere2" is evaluated during ray-trace |
| 18:39.27 | brlcad | no |
| 18:39.44 | brlcad | http://en.wikipedia.org/wiki/Constructive_solid_geometry |
| 18:40.25 | pacman87 | i meant the surface that's drawn is the second one the ray hits |
| 18:40.47 | pacman87 | so it has to enter sph1 first, then the surface of sph2 that's inside sph1 is actually drawn |
| 18:40.56 | brlcad | think of them independently |
| 18:41.07 | brlcad | you compare a ray against sphere1 .. you end up with a segment |
| 18:41.19 | brlcad | you compare a ray against sphere2 .. you end up with a different segment |
| 18:41.54 | brlcad | with an intersection CSG operation, there is an evaluated result only where the two segments coexist |
| 18:42.03 | brlcad | i.e. where they overlap and share space |
| 18:42.17 | pacman87 | ah, now it's all clear, thanks |
| 18:43.17 | brlcad | now with *brep* on brep csg evaluation, that's a totally different approach |
| 18:43.42 | brlcad | there you have surfaces.. one surface for sphere1 .. one surface for sphere2 |
| 18:44.13 | brlcad | you actually compute the intersection of those two surfaces by performing pairwise trimming operations of the actual surfaces |
| 18:45.45 | brlcad | so in the picture in the wikipedia link, the intersection result ends up with a brep that has 4 surfaces stiched together (created from the 6 + 2 surfaces for the box and sphere) |
| 18:46.45 | pacman87 | sph has 2 surfaces? |
| 18:49.36 | brlcad | yeah |
| 18:50.21 | brlcad | could conceivably have just one, I don't know how the brep control points would be like for that, if it's numerically stable |
| 18:50.43 | pacman87 | could you recommend a good reference book on the math behind csg/brep? |
| 18:51.18 | pacman87 | i'm was planning to stop by the engr lib this afternoon |
| 18:53.47 | brlcad | oof, there's a lot of material on the subject but maybe find a few papers |
| 18:54.22 | pacman87 | right, i was wonder if there were any 'classic texts' in the field |
| 18:54.24 | brlcad | http://sca2002.cs.brown.edu/~jfh/papers/Laidlaw-CSG-1986/paper.pdf is kinda one of the original canon papers on how to implement csg evaluation of polygonal brep models |
| 18:54.50 | brlcad | i'm frankly not sure that will work for spline surface brep models |
| 18:57.20 | pacman87 | allowing partial revolutions of 3D objects adds a *lot* of complexity |
| 18:59.29 | pacman87 | i'm assuming i should set the revolve up to be usable with any primitive, combination, or region? |
| 18:59.46 | pacman87 | keep it generalized |
| 19:01.24 | brlcad | erm, I'm confused |
| 19:01.31 | brlcad | what are you actually working on proposing? :) |
| 19:01.41 | brlcad | csg evaluation of breps or a new primitive? :) |
| 19:02.28 | brlcad | fully implementing a revolution and/or sweep primitive is in itself enough scope for a project |
| 19:02.57 | brlcad | otherwise, yeah, revolve and sweep should be generalized to any primitive |
| 19:04.14 | pacman87 | i'm proposing new primitive(s), and i was trying to find the point at which a new primitive was 'done', ie could be used by the rest of the system |
| 19:06.02 | pacman87 | combs and regions aren't primitives, though (are they?) |
| 19:13.41 | brlcad | ahhh |
| 19:13.49 | brlcad | then you don't need to worry about csg evaluation of breps then :) |
| 19:13.59 | brlcad | combs and regions are not primitives |
| 19:14.49 | brlcad | a new primitive is done when all of the hook functions in g_*.c are implemented librt, it has creation routines in libwdb, and it has editing features implemented in mged |
| 19:17.18 | brlcad | the big hook functions that take time are prep/shot, import (v5), export (v5), tess, plot, nurb(optional), and to a lesser extent norm, curve, and uv |
| 19:17.22 | brlcad | the rest are trivial |
| 19:17.51 | brlcad | prep/shot, tess, and nurb are usually the hard ones -- shot often being the hardest |
| 19:18.04 | pacman87 | brlcad: right, i've been looking in-depth at the g_xxx.c and g_*.c files |
| 19:20.06 | pacman87 | found rt_xxx_class() in g_xxx.c, does that just return whether the xxx is valid? |
| 19:20.36 | pacman87 | rt_*_class() returns 0 for the real primitives i've checked |
| 19:21.04 | brlcad | if you want to see all of the places a primitive is hooked, you can follow one of the existing primitives |
| 19:21.09 | brlcad | you can ignore class |
| 19:21.17 | brlcad | you'll see everyone just returns 0 |
| 19:21.41 | pacman87 | ok, just double checking |
| 19:22.43 | pacman87 | should i include algroithms for those methods in my application? |
| 19:24.09 | brlcad | example: find . \( -name \*.c -o -name \*.h \) -exec grep -I SUPERELL {} /dev/null \; | cut -f1 -d: | sort | uniq |
| 19:24.29 | brlcad | <PROTECTED> |
| 19:24.48 | brlcad | those are two relatively new primitives |
| 19:25.01 | brlcad | that shows you all the files that are edited that involve them |
| 19:25.21 | brlcad | with slight differences -- superell has mirror and v4 supported, metaballs don't |
| 19:25.28 | brlcad | you'd need mirror, but not v4 |
| 19:26.08 | brlcad | you should instill confidence that you're aware of what's involved in the algorithms, regardless of whether you know the algorithm itself yet |
| 19:26.16 | pacman87 | where is the generic interface for the rt_*_* methods? (assuming there is one) |
| 19:26.39 | pacman87 | so a very high-level description of the steps |
| 19:27.16 | brlcad | yeah, explain what you think you have to do, maybe include "why" (i.e. wth do those functions do) |
| 19:28.08 | pacman87 | the revolve will probably be a lot more detailed than the sweep, as i've only been thinking about rev |
| 19:28.21 | brlcad | what and how are more important than why, but can be high-level descriptions with maybe some technical references (who's algorithm/approach are you going to use? your own? some paper?) |
| 19:28.29 | brlcad | that's fine |
| 19:29.09 | brlcad | you could simply state that if you get enough done, you would then move on to the sweep |
| 19:29.33 | pacman87 | s/enough// :) |
| 19:29.44 | brlcad | it's entirely conceivable that you could hack out a new primitive in a matter of just two weeks .. or it could take months |
| 19:30.22 | pacman87 | yeah, i was trying to get a clearer idea of scope earlier so i could figure how long it would take |
| 19:30.27 | brlcad | and there's nothing wrong with either timeline |
| 19:30.37 | pacman87 | i'm thinking the rev could take a few weeks, maybe a month |
| 19:30.55 | pacman87 | then the sweep would probably take a bit longer |
| 19:31.07 | brlcad | care more about having devs working on cool stuff for the project that we need, attracting new long-term guys excited about this stuff :) |
| 19:31.35 | brlcad | s/guys/guys n' gals of course/ |
| 19:32.18 | pacman87 | yeah, working on primitives fits with my approach to problem solving |
| 19:32.28 | pacman87 | i work as a tutor for the college of engineering |
| 19:33.00 | pacman87 | and pretty much every problem/concept i teach, i start with the basic integral/derivative forms of the equations (i teach physics) |
| 19:35.18 | pacman87 | heads off to the engr library, back later |
| 19:43.31 | brlcad | cya! |
| 20:59.19 | *** join/#brlcad Twingy (n=justin@74.92.144.217) [NETSPLIT VICTIM] | |
| 21:32.24 | *** join/#brlcad cad79 (n=7aa2a947@bz.bzflag.bz) | |
| 22:25.42 | *** join/#brlcad spike_ (n=spike@eridani.acm.jhu.edu) | |
| 22:26.15 | spike_ | hey, is anyone in here a mentor for the GSOC project? |
| 22:26.28 | brlcad | hello spike_ |
| 22:26.35 | spike_ | hey :-D |
| 22:26.40 | brlcad | several, best to just speak up :) |
| 22:27.05 | b0ef | it would be nice if NURBS was put on the gsoc page |
| 22:27.28 | brlcad | b0ef: it is in several ways |
| 22:27.29 | spike_ | i don't know too much about BRL-CAD but it seemed like a cool project, just curious what projects need to get done |
| 22:27.41 | brlcad | spike_: have you seen the ideas list? |
| 22:27.48 | spike_ | yup |
| 22:28.05 | spike_ | a lot of them look cool, i just figured i should talk to someone instead of just 'picking' one |
| 22:28.13 | brlcad | then... those are the projects that need to get done ;) |
| 22:28.23 | brlcad | actually you picking one is best |
| 22:28.31 | brlcad | they're all great ideas, lots of possibilities |
| 22:28.42 | spike_ | haha oh alright then |
| 22:29.06 | brlcad | the ones that are "most important" are identified as high priority, but a really good proposal will trump that in just about any category |
| 22:30.00 | brlcad | same goes for a project not even on the list if it fits with our goals |
| 22:55.20 | CIA-33 | BRL-CAD: 03brlcad * r30594 10/brlcad/trunk/BUGS: resolved ogl crash, removed duplicate context invalidate bug |
| 23:44.32 | brlcad | woot, but fixed |