| 00:11.20 | *** part/#brlcad misteriexous (n=ubuntu@217.118.79.39) | |
| 00:12.39 | csanyipal | so long |
| 00:34.55 | brlcad | there is also a "plane" primitive .. the "half" halfspace primitive will make an infinite plane |
| 01:10.06 | Ralith | brlcad, ah, cool. What happens when you try to intersect something with it? |
| 01:25.53 | brlcad | it does what you'd expect -- it's "half" of infinity, so if you intersect a vehicle with it, you'll get a cutting plane |
| 01:26.11 | Ralith | oh |
| 01:26.14 | Ralith | not the kind of plane I had in mind |
| 01:26.29 | brlcad | what did you have in mind? |
| 01:26.38 | brlcad | it's not infinitely thin |
| 01:26.53 | Ralith | infinitely thin was what I was going for |
| 01:26.57 | brlcad | ah |
| 01:27.10 | brlcad | solid modeling, everything has to have a thickness ;) |
| 01:27.17 | Ralith | that's what I thought |
| 01:28.10 | brlcad | it would be interesting/useful to have a user-specific non-zero thickness parameter, but then that wouldn't be a halfspace any longer unless they specified infinity |
| 01:28.35 | Ralith | nothing wrong with a rename. |
| 01:28.52 | brlcad | you can get the same effect with two halfspaces that overlap with whatever thickness |
| 01:29.02 | Ralith | thought that might work; neat |
| 01:29.37 | brlcad | it's not an optimized primitive, so depending on what you're doing -- you might be much better off using arb8's anyways |
| 01:29.47 | Ralith | optimized? |
| 01:30.23 | brlcad | it'll slow down a ray-trace rendering |
| 01:31.12 | Ralith | ah. |
| 01:31.58 | brlcad | most of the primitives are optimized, but not the way that one is implemented given how it extends to infinity -- every ray has to test against it |
| 01:32.31 | Ralith | seems pretty quick to test against, though |
| 01:32.37 | brlcad | so like if you wanted to chop an object in half, it'd usually be much faster to use a box instead of a half |
| 01:35.01 | Ralith | so I think we've discussed this before but I don't recall the details and I have more questions anyway; what is/will be the most effective way to get a 2D slice through an object with as little data loss as possible, for purposes such as g-code generation? |
| 01:35.56 | Ralith | the best I can come up with right now is make a very thin slice and do a orthographic render at very high from directly overhead, then fit splines to it. |
| 01:36.06 | Ralith | but that's really cheating. |
| 01:36.15 | Ralith | at very high resolution* |
| 01:45.04 | brlcad | that's a perfectly viable approach actually, not far off what some of the commercial engines do effectively |
| 01:45.40 | Ralith | yes, it would *work* |
| 01:46.14 | Ralith | but I'd feel so much better about a workflow that didn't deliberately drop information and then desperately try to reconstruct as much as possible |
| 01:46.49 | brlcad | naturally ;) |
| 01:47.24 | Ralith | not to mention the complexities and accuracy issues of path tracing algos in and of themselves. |
| 01:48.41 | brlcad | you could get that pretty accurate just because you'd have knowledge of object boundaries |
| 01:48.57 | brlcad | mostly limited by your sampling resolution, which could be pretty intense as needed |
| 01:49.03 | brlcad | but like you said, less than ideal |
| 01:49.50 | brlcad | the alternate approach requires brep support to be complete |
| 01:50.25 | brlcad | you'd request a brep outline for all objects and then project them onto a plane |
| 01:50.59 | Ralith | so is brep support something that's being evaluated, or even developed? |
| 01:51.22 | brlcad | it's the top-priority of the year, been under development for over a year now |
| 01:51.34 | Ralith | oo! |
| 01:51.37 | Ralith | encouraging! |
| 01:51.42 | brlcad | there's basic support in there already |
| 01:51.50 | Ralith | how far along would you say things are? |
| 01:51.54 | brlcad | some forms will even ray-trace correctly |
| 01:52.04 | Ralith | forms? |
| 01:52.29 | brlcad | different types of surface evaluations |
| 01:53.11 | brlcad | if you fire a ray at an object, do you get 1, 2, 3, 4, ... N intersections with the surface, when are you inside/outside, how to deal with numerical instability, etc |
| 01:53.30 | Ralith | ah. |
| 01:53.32 | brlcad | there are a lot of problems that need to be solved still, but this is being driven by several priorities |
| 01:53.40 | Ralith | great! |
| 01:53.41 | Ralith | :D |
| 01:54.02 | Ralith | this is one of the things I like about BRL-CAD |
| 01:54.14 | brlcad | the brep/nurbs work is the foudation of interactive (opengl) visualizations, step conversion support, and more |
| 01:54.41 | Ralith | as a project, it has a great tendency to already be working on all the really neat ideas that you generally don't see bandied about outside of theory. |
| 01:55.52 | Ralith | so that's how it relates to the tesselation stuff. Remembered you saying something about that last time this came up but didn't follow the connection. |
| 01:56.21 | Ralith | thanks for the info :) |
| 01:56.43 | brlcad | the same thing that would let us ask for a contour for projection onto a plane to get splines is basic to the brep form |
| 01:57.17 | brlcad | from there, you can easily tessellate as it (presuming we have implemented csg evaluation of brep-on-brep objects) |
| 01:57.36 | Ralith | what field of study most completely encompasses this sort of thing? |
| 01:57.48 | brlcad | http://brlcad.org/BRL-CAD_Priorities.png <-- bottom-left priority |
| 01:57.56 | Ralith | computational gemoetry? |
| 01:58.27 | brlcad | er, loosely/sorta/not-really |
| 01:58.48 | Ralith | didn't think so. |
| 01:59.03 | brlcad | it does and it doesn't -- CG covers some of the fundamental algorithms and generally assumes clean numerics |
| 02:00.05 | Ralith | 'clean numerics' meaning no loss of precision due to fp inaccuracies etc.? |
| 02:00.17 | brlcad | from there you can build up the more complicated algorithms, but without resorting to fixed-point arithmetic (and taking a massive performance penalty -- like two orders of magnitude) there are lots of practical implementation problems that still have to be addressed |
| 02:02.37 | Ralith | so *is* there such a thing as a formal field of study covering this sort of real-world implementation? |
| 02:02.58 | Ralith | (also: that poster is beautiful. Who put it together?) |
| 02:04.46 | brlcad | there's lots of work in the area of solid modeling research -- an acm conference dedicated to it as well as a few others |
| 02:05.50 | brlcad | there are just literally hundreds of algorithms and issues in any given CAD or solid modeling system, boundary structure evaluation just being a small set of them |
| 02:06.37 | Ralith | of course. |
| 02:09.26 | brlcad | oh, and I put that poster together |
| 02:09.59 | Ralith | nice! |
| 02:10.01 | Ralith | what'd you lay it out in? |
| 02:10.09 | brlcad | hm |
| 02:10.41 | brlcad | probably either Pages or InDesign, don't recall |
| 02:11.07 | Ralith | pages? |
| 02:11.38 | brlcad | it never went out to the list because it got a less than favorable review by one the editors at arl, said it was too confusing and too many words |
| 02:11.55 | Ralith | bah. |
| 02:12.01 | brlcad | http://www.apple.com/iwork/pages/ |
| 02:12.11 | Ralith | it's a technical subject. Walls of text are par for the course. |
| 02:13.26 | Ralith | unless it was supposed to be some sort of general-audience PR material, but I can't imagine what most of a general audience would want to do with BRL-CAD in the first place. |
| 02:16.32 | brlcad | audience was general user community, anyone that would care what our high-level roadmap looks like |
| 02:17.24 | brlcad | so they did have a little bit of a point, whether it matters for this was arguable, but it was enough to not send it out without thinking about reworking some parts |
| 02:21.40 | Ralith | ah well. |
| 03:01.19 | *** join/#brlcad IriX64 (n=IriX64@bas2-sudbury98-1177871582.dsl.bell.ca) | |
| 03:01.49 | IriX64 | http://rafb.net/p/ushWLt28.html <---- 3 or 4 of these went by, do you know or care about them? |
| 03:04.00 | IriX64 | or is my compiler in error :) |
| 03:08.06 | IriX64 | http://rafb.net/p/NItopl53.html <--- this too my compiler presented me with |
| 03:12.02 | IriX64 | sigh... i wish i'd never dropped out of special ed :) |
| 05:05.02 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) | |
| 06:17.33 | *** join/#brlcad WhiteCalf (i=WhiteCal@whitecalf.net) | |
| 08:23.40 | *** join/#brlcad csanyipal (n=csanyipa@91.102.231.33) | |
| 10:08.56 | *** join/#brlcad mafm (n=mafm@172.Red-83-45-253.dynamicIP.rima-tde.net) | |
| 10:09.14 | mafm | hi |
| 10:23.48 | csanyipal | hi |
| 10:25.11 | csanyipal | Ralith: when I create a plane as arb8 shape with zero hight, then I can't raytrace it, in the framebuffer it can't see. Why? |
| 10:26.23 | Ralith | csanyipal: as far as the raytracer is concerned, if it doesn't have volume, it doesn't exist. |
| 10:38.00 | csanyipal | Ralith: OK thanks! |
| 11:01.33 | Axman6 | ha, set brlcad to checkout about 12 hours ago... came home, it was asking me to if i want to accept the ssh key -_- |
| 11:02.19 | Axman6 | s/set brlcad to checkout/set svn to checkout brlcad |
| 11:02.42 | *** join/#brlcad Elrohir (n=kvirc@p5B14EFC3.dip.t-dialin.net) | |
| 11:14.09 | *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) | |
| 11:37.51 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) | |
| 14:01.49 | brlcad | csanyipal: I was saying yesterday, there is an infinite-plane primitive -- it's the "half" halfspace primitive |
| 14:07.26 | csanyipal | brlcad: thank you, I find it, and tried it out. |
| 14:09.58 | csanyipal | Still I don't know how to create the shape that is on the picture "Azimuth, Elevation, and the xyz Coordinate System" in the tutorial "Volume II - Introduction to MGED" at page 18. |
| 14:12.10 | csanyipal | That shape shows the angle of the Azimuth. |
| 14:13.03 | brlcad | I modeled that, what would you like to know? |
| 14:13.21 | brlcad | it's a pretty simple shape |
| 14:24.08 | csanyipal | It's difficult to explain because of my "weak" English. |
| 14:26.15 | brlcad | I went to some intricate lengths to get exact cuts and no overlaps with very specific shapes in mind -- but the basic shape is that of a torus with arbs cutting off the portion I didn't want and a tgc on the end |
| 14:26.19 | csanyipal | The shape looks like a part of a torus and have in front a trc Truncated Right Cone |
| 14:26.37 | brlcad | basically |
| 14:27.05 | brlcad | it could be done a little more easily using a pipe, but I didn't know it back then |
| 14:27.20 | brlcad | one pipe with a tgc, no need for subtracted arbs |
| 14:28.25 | csanyipal | But can one bend a pipe? |
| 14:30.31 | csanyipal | And how to direct the tgc in the right direction? |
| 14:31.13 | brlcad | of course you can bend a pipe, that's what they consist of -- bend segments and straight segments |
| 14:31.33 | csanyipal | Aha, there is a bend point! |
| 14:31.49 | brlcad | trc's are the same as a tgc, just a specific subcase |
| 14:32.01 | csanyipal | I try the command in pipe.s pipe |
| 14:32.05 | brlcad | all the conics end up being a tgc as it's a general case |
| 14:32.23 | csanyipal | yes |
| 14:32.29 | brlcad | the tgc is specified with a point and a vector -- that puts it in the right direction |
| 14:33.10 | brlcad | there are just additional vectors for orienting and sizing the top ellipse and base ellipse caps |
| 14:33.27 | csanyipal | tgc has a Vertex and a high vector and radius of base and of top |
| 14:33.53 | brlcad | few more parameters |
| 14:34.18 | brlcad | sounds about right for the trc |
| 14:36.21 | csanyipal | I know that but when I want to give the values (x,y,z) for the high vector then .. OK, I'll try it out now. |
| 14:37.23 | csanyipal | thanks for the advice about pipe! |
| 14:37.30 | brlcad | start with something simple so you understand the shape, and turn on your model axes so you can see what direction is what |
| 14:37.55 | csanyipal | yes, I did that alreday.. |
| 14:38.01 | csanyipal | already |
| 14:38.26 | csanyipal | o |
| 14:38.28 | csanyipal | ok |
| 15:17.12 | csanyipal | brlcad: If you have i little time: I don't understand at pipe the points for the pipe and bend points |
| 15:18.03 | csanyipal | mged> in pipe.s pipe 2 0 0 0 0.1 1 100 100 0 0 0.1 1 50 |
| 15:18.41 | csanyipal | give to me a right pipe and not a bended pipe |
| 15:23.53 | csanyipal | brlcad: OK I find a tutorial for the pipe in the Volume III - Principles of Effective Modelling |
| 15:49.25 | *** join/#brlcad redvsblue (i=Pandora@c-69-247-220-102.hsd1.mo.comcast.net) | |
| 15:57.16 | *** join/#brlcad DanielFalck (n=dan@pool-71-111-64-53.ptldor.dsl-w.verizon.net) | |
| 20:05.57 | *** join/#brlcad alex_joni (n=juve@emc/board-of-directors/alexjoni) [NETSPLIT VICTIM] | |
| 20:07.35 | *** join/#brlcad redvsblue (i=Pandora@c-69-247-220-102.hsd1.mo.comcast.net) [NETSPLIT VICTIM] | |
| 20:07.35 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) [NETSPLIT VICTIM] | |
| 20:07.35 | *** join/#brlcad mafm (n=mafm@172.Red-83-45-253.dynamicIP.rima-tde.net) [NETSPLIT VICTIM] | |
| 20:07.35 | *** join/#brlcad archivist (n=archivis@host81-149-119-172.in-addr.btopenworld.com) [NETSPLIT VICTIM] | |
| 20:07.42 | *** join/#brlcad brlcad (n=sean@bz.bzflag.bz) | |
| 20:07.42 | *** join/#brlcad d-lo (n=claymore@bz.bzflag.bz) [NETSPLIT VICTIM] | |
| 20:07.42 | *** join/#brlcad ChanServ (ChanServ@services.) | |
| 20:07.42 | *** mode/#brlcad [+oo brlcad ChanServ] by irc.freenode.net | |
| 20:09.20 | *** join/#brlcad DanielFalck (n=dan@pool-71-111-64-53.ptldor.dsl-w.verizon.net) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad Elrohir (n=kvirc@p5B14EFC3.dip.t-dialin.net) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad bjork_ (n=bjork@ip72-204-40-138.fv.ks.cox.net) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad mon8 (i=yoda@CPE0016d35dfacc-CM000f9f7f1258.cpe.net.cable.rogers.com) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net) | |
| 20:09.21 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad MinuteElectron (n=MinuteEl@unaffiliated/minuteelectron) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad kanzure (i=bbishop@66.112.232.117) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad SWPadnos (n=Me@emc/developer/SWPadnos) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] | |
| 20:09.21 | *** join/#brlcad poolio (n=poolio@bz.bzflag.bz) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad DanielFalck (n=dan@pool-71-111-64-53.ptldor.dsl-w.verizon.net) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad Elrohir (n=kvirc@p5B14EFC3.dip.t-dialin.net) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad bjork_ (n=bjork@ip72-204-40-138.fv.ks.cox.net) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad mon8 (i=yoda@CPE0016d35dfacc-CM000f9f7f1258.cpe.net.cable.rogers.com) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net) | |
| 20:12.30 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad SWPadnos (n=Me@emc/developer/SWPadnos) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad poolio (n=poolio@bz.bzflag.bz) [NETSPLIT VICTIM] | |
| 20:12.30 | *** join/#brlcad kanzure (i=bbishop@66.112.232.117) [NETSPLIT VICTIM] | |
| 20:12.31 | *** join/#brlcad MinuteElectron (n=MinuteEl@unaffiliated/minuteelectron) [NETSPLIT VICTIM] | |
| 20:12.32 | *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) | |
| 21:42.48 | *** join/#brlcad mafm_ (n=mafm@172.Red-83-45-253.dynamicIP.rima-tde.net) | |
| 22:18.55 | *** join/#brlcad misteriexous (n=ubuntu@217.118.79.45) | |
| 22:19.30 | misteriexous | i have a grammar check languagetool.org + new java6.0 engine |
| 22:19.57 | misteriexous | (just want to know an exact time at ibot) |
| 22:21.36 | *** part/#brlcad misteriexous (n=ubuntu@217.118.79.45) | |
| 23:19.51 | *** join/#brlcad ``Erik (i=erik@c-68-54-174-162.hsd1.md.comcast.net) | |
