| 00:00.15 | andrecastelo | thanks, brlcad, i'll do some research, take a look at the code, write the application and see how it turns out ;) |
| 00:00.24 | brlcad | sounds good |
| 00:15.14 | mafm | so brlcad, are you available now to talk about that? |
| 00:15.27 | brlcad | probably |
| 00:16.22 | mafm | good |
| 00:16.53 | mafm | so what about the distributed approach, can you please elaborate or link to some document? |
| 00:17.16 | brlcad | well we have a raytracer in brl-cad |
| 00:17.19 | brlcad | lots of them, actually |
| 00:17.34 | brlcad | all based around our librt ray-trace library |
| 00:18.49 | brlcad | we support parallel ray-tracing pretty pervasively |
| 00:19.14 | brlcad | 'rt' is the main ray-tracer |
| 00:19.42 | brlcad | we also have distributed ray-tracing via 'remrt' (remote rt) and 'rtsrv' (remote compute daemons) |
| 00:20.30 | brlcad | remrt and rtsrv basically set up manual communication pipes and negotiate packets as needed, and you have to specify all machines you want to work on |
| 00:21.25 | brlcad | what we don't do easily though is 'automatic distributed' using something like mpi, pvm, or grid interfaces |
| 00:21.44 | brlcad | it would be nice if rt inherintly had distributed capacities |
| 00:22.45 | mafm | are the computations heavy enough that it compensates to distribute them via network? |
| 00:22.50 | brlcad | so that's the basic idea, either adding the functionality to rt or giving remrt/rtsrv major updates to improve them |
| 00:23.18 | brlcad | that depends entirely on the model being rendered |
| 00:23.41 | brlcad | some are blazingly fast, others swimmingly slow |
| 00:23.58 | mafm | slow as in minutes? |
| 00:24.19 | brlcad | anywhere from subsecond to hours |
| 00:24.51 | brlcad | so it really needs to auto-detect and load-balance accordingly, know when to farm out the job and when not to |
| 00:25.17 | brlcad | the models might be just a few k or might be several GB in size too |
| 00:26.17 | mafm | typical grid jobs, for scientific grids that I'm used to, usually take hours/days |
| 00:26.31 | brlcad | the two extremes where we usually want distributed are when someone's making a movie and rendering thousands of frames or someone rendering a really high-quality detailed complex model rendering |
| 00:26.44 | mafm | and in most them the turnaround time for "hello world" is like 3-5 minutes :D |
| 00:27.14 | mafm | so the idea would be to run them withing clusters I guess, with LAN setups? |
| 00:27.28 | brlcad | pretty much |
| 00:28.25 | mafm | I have programmed in PVM a bit, only at univ for test programs |
| 00:28.29 | brlcad | as well as loosely connected LAN-based setups over slow connections (a typical office with maybe 10-100 high-end workstations available to run compute nodes with an mpi config for example) |
| 00:28.45 | mafm | and in MPI at work but only for testing, no real applications |
| 00:29.16 | mafm | anyway the theory I think that I'm OK with it |
| 00:29.25 | brlcad | MPI or a solid grid framework would probably be ideal |
| 00:30.06 | mafm | but I guess that for doing that from a ray tracer, I would have to have solid knowledge of the techniques of raytracing too? |
| 00:30.52 | brlcad | not necessarily |
| 00:30.59 | brlcad | at least not beyond understanding the data model |
| 00:31.17 | brlcad | ray-tracing generally requires the entire model for example since rays can diverge in any direction |
| 00:32.00 | brlcad | so you either have some sort of master dispatch and distribute portions of data or distribute all the data to all nodes and dispatch rays/scanlines or some hybrid |
| 00:33.31 | mafm | say that you want to render a village |
| 00:33.47 | mafm | the input are the solids represenging houses and so on |
| 00:33.59 | mafm | the output would be what, just the 2D image? |
| 00:34.20 | brlcad | yep |
| 00:34.27 | brlcad | or a series of 2D images in the case of an animation |
| 00:35.19 | mafm | say that you want an animation |
| 00:35.28 | andrecastelo | hey brlcad, the download page is a little off |
| 00:35.55 | andrecastelo | it points to the 7.10.2 release (at least the source) |
| 00:35.56 | mafm | the output or intermediate calculations of one frame could be used for the second? (similar to what MPEG do) |
| 00:36.32 | brlcad | andrecastelo: thanks |
| 00:36.40 | brlcad | it needs to be wikified |
| 00:36.50 | brlcad | or made a module that auto-points to the latest |
| 00:37.05 | brlcad | mafm: yes |
| 00:37.18 | brlcad | rt already has inherint support for multiframe rendering |
| 00:37.33 | andrecastelo | (also, if i want to check the code, i get the 7.12.0 source, right?? how hard is it to compile in windows?? ) |
| 00:37.40 | brlcad | you feed it transactioned command lists and it deals with what needs to be reprepped accordingly |
| 00:38.00 | brlcad | andrecastelo: there's an msvc studio project in misc -- should be click and run |
| 00:38.25 | andrecastelo | great, thanks |
| 00:39.29 | mafm | if frames are interdependent, that turns it a bit difficult it seems |
| 00:39.42 | brlcad | they're independent |
| 00:40.09 | brlcad | it's just you can speed things up if you don't need to reprep, so rt does so accordingly |
| 00:40.52 | mafm | well yes, but if rendering them sequentially is much less effort than the sum() of separate efforts, the speedup of distributing it is not very big |
| 00:41.05 | mafm | with aqsis there was a similar problem |
| 00:41.19 | mafm | buckets (portions of the image) were in principle independent |
| 00:41.35 | mafm | but constructing the octree was very expensive |
| 00:42.23 | mafm | so if you didn't take advantage of that, and construct the octree from scratch each time, the result was that it was more expensive when using several threads |
| 00:42.56 | mafm | I think that there could be similar problems in brlcad, or you already have that part sorted out? |
| 00:43.09 | brlcad | you're thinking thread-level, we have that |
| 00:43.26 | brlcad | parallel is a done deal |
| 00:43.36 | brlcad | distributed even works, it can just be heavily improved |
| 00:44.15 | mafm | I know the difference between theads and distributed, it was an example, to be extrapolated :) |
| 00:44.26 | brlcad | so yeah, buckets of an impage are independent -- but the model itself isn't |
| 00:44.30 | hippieindamakin8 | andrecastelo and mafm u guys might be interested in this http://www.siggraph.org/education/materials/HyperGraph/raytrace/rtrace0.htm |
| 00:44.45 | andrecastelo | hippieindamakin8: thanks! |
| 00:44.50 | brlcad | mafm: sure, but the latency concerns between parallel and distributed are huge |
| 00:45.03 | hippieindamakin8 | u wanted to know the raytracing fundamentals rt :) np |
| 00:45.22 | brlcad | that's why the real driver is big render jobs where we know implicitly that it's going to be a benefit (huge render or tons of frames) |
| 00:45.31 | andrecastelo | also - is a kd tree implemented?? |
| 00:45.56 | brlcad | effectively, we have a hybrid |
| 00:46.04 | mafm | I see |
| 00:47.16 | brlcad | brl-cad's rt is actually the first parallel distributed ray-tracer (per all published references), an early second-generation ray-tracer |
| 00:48.02 | mafm | that's remrt+rtsrv? |
| 00:48.10 | brlcad | the distributed portion, yes |
| 00:48.11 | mafm | or the librt? |
| 00:48.30 | mafm | I mean, which parts would I have to modify? |
| 00:48.59 | brlcad | depends entirely on whether the idea is to improve remrt usability or add inherint distributed support to rt :) |
| 00:49.04 | brlcad | there are merits to both |
| 00:51.11 | brlcad | could read some of the published papers, this relates specifically to brl-cad: http://ftp.arl.army.mil/~mike/papers/95cadsymp/realtime.html |
| 00:51.14 | mafm | both them are in C it seems (browsing the repositories) |
| 00:52.35 | brlcad | more references in http://graphics.cs.uni-sb.de/Publications/2001/star.pdf |
| 00:52.43 | brlcad | yes, 90% of the sources are in C |
| 00:53.09 | brlcad | only the new BREP code is in C++ plus a few modules, along with the tcl/tk modeler sources |
| 00:53.31 | brlcad | rt stays C, remrt could be redesigned |
| 00:53.31 | mafm | is there a "central" source file in librt where the high level stuff is implemented? |
| 00:53.51 | brlcad | what high-level stuff? |
| 00:54.17 | brlcad | high-level ray dispatching is in the binary front-end, src/rt src/remrt |
| 00:55.09 | mafm | I mean an important file in librt |
| 00:55.21 | mafm | to see how scary is the code :D |
| 00:56.29 | mafm | int is_hackers_night( struct timeval *tv ); |
| 00:56.33 | mafm | :D |
| 00:56.48 | brlcad | you wouldn't really be modifying anything in librt, at least I'd be surprised if you need to |
| 00:57.50 | brlcad | otherwise, some of the core is in shoot.c, bool.c |
| 00:58.18 | brlcad | you really should start from srt/rt though -- librt is the library itself, even parallelism happens higher up |
| 00:59.16 | brlcad | src/rt/view.c do.c and most importantly worker.c |
| 01:00.06 | hippieindamakin8 | librt is all about the geometry representation and manipulations/transformations mostly |
| 01:01.55 | mafm | ok, looking at those :) |
| 01:01.58 | hippieindamakin8 | brlcad .. btw wat is the function of this server ? rtserver |
| 01:04.24 | Daytona | librtserver is a C library that is loaded by a Java class to provide raytracing services for a Java App |
| 01:07.28 | hippieindamakin8 | ohh |
| 01:07.42 | hippieindamakin8 | tx :) |
| 01:08.32 | mafm | hmm, don't know brlcad |
| 01:08.59 | mafm | does this have more priority than the other idea? |
| 01:09.09 | Daytona | hippieindamakin8: If you have some ideas on how to improve that code, that would be great!!! |
| 01:09.11 | mafm | I haven't even seen it in the Project Ideas page |
| 01:09.13 | hippieindamakin8 | daytona can u suggest me an appilcation |
| 01:10.36 | hippieindamakin8 | :) i just opened it daytona :P i wanted to apply for this OO API project so was going through the places where there is a possibility of rendering some geometry :) |
| 01:10.57 | hippieindamakin8 | as i get more familiar with brlcad i sure will :) |
| 01:10.59 | brlcad | mafm: like I said -- a really solid application for any of the ideas trumps the priority aspect |
| 01:12.09 | Daytona | hippieindamakin8: The librtserver is just used in a JNI wrapper |
| 01:13.17 | hippieindamakin8 | ohh there is this jade.jar in the src :) now i am connecting pieces |
| 01:13.35 | hippieindamakin8 | in case it is connected to this |
| 01:14.08 | Daytona | the librtserver and the jbrlcad are not connected |
| 01:14.57 | hippieindamakin8 | nono.. in the src there is this folder called java |
| 01:15.05 | hippieindamakin8 | i aint talking about jbrlcad |
| 01:15.11 | Daytona | jade.jar is a units and measurments jar (now replaced by jscience) |
| 01:15.50 | hippieindamakin8 | sorry.. :P i am mistaken.. i wanted to talk about brlcad server.java |
| 01:16.11 | hippieindamakin8 | sorry.. it was a long time :P |
| 01:16.32 | mafm | brlcad: hmm, honestly, I'm not talking into getting into GSoC with BRL-CAD for being a higher priority project |
| 01:17.01 | brlcad | no comprendo |
| 01:17.02 | mafm | hmm, let me think about how to explain it... (I'm a bit tired at the moment :) ) |
| 01:17.11 | hippieindamakin8 | ~/brlcad7-10-2/src/java/mil/army/arl/brlcad |
| 01:18.43 | Daytona | looks like there is almost nothing in that class |
| 01:19.10 | hippieindamakin8 | ya i shalll go through that :) |
| 01:20.50 | Daytona | the jbrlcad has a bit more functionality :-) |
| 01:21.27 | mafm | <brlcad> mafm: like I said -- a really solid application for any of the ideas trumps the priority aspect |
| 01:21.46 | mafm | when I'm asking about the priority, I don't mean to take it as an advantage |
| 01:22.10 | mafm | I mean if you suggest that because of my relation with distributed computing at work or what |
| 01:22.28 | mafm | or if you came up with that for other reasons |
| 01:22.48 | brlcad | yes, I only suggested it as a consideration because of your background |
| 01:23.05 | brlcad | otherwise, no different than then other listed ideas |
| 01:23.12 | mafm | yesterday we were talking about other proposals, it's you who started to talk about this one (IIRC, I was a bit sleepy) |
| 01:23.59 | brlcad | there was someone (a student) that worked on an mpi layer for rt a couple years ago too, for several months, but they were learning everything from scratch and never finished anything more than job propagation |
| 01:24.36 | mafm | aha |
| 01:24.58 | mafm | so one of the reasons is that maybe I see it as a more doable proposal than the other one? |
| 01:25.18 | brlcad | right, and I firmly believe that you need to be excited about the work --figured with your background that distributed might be of interest, if not then screw it |
| 01:26.20 | mafm | well, certainly distributed computing it is interesting, but I enjoy many areas |
| 01:27.18 | mafm | working with a 3D engine to communicate with commands with a remote server is also another, making a multiplayer game it's similar in concept, I think |
| 01:27.51 | mafm | so probably this project would have more chances to succeed |
| 01:27.54 | brlcad | some loose similarities, yes |
| 01:29.38 | mafm | but I also understand that maybe it would be more valuable for you to make my application for other project, because is less common |
| 01:30.02 | mafm | so if I was accepted and finish the program, there would be more chances for other people to do that OpenGL GU |
| 01:30.03 | mafm | GUI |
| 01:30.04 | brlcad | what's less common? |
| 01:30.23 | brlcad | ah, yes -- it certainly has higher impact potential |
| 01:30.30 | mafm | the knowledge of MPI/distributed stuff |
| 01:30.36 | brlcad | it's also one that will be more critiqued design-wise :) |
| 01:30.57 | mafm | so in the case that I would go ahead, I would be beaten? :) |
| 01:32.15 | brlcad | nope |
| 01:33.45 | mafm | good to know :) |
| 01:33.47 | mafm | well, I see two problems then: |
| 01:34.03 | mafm | 1) it's more difficult for me to come up with a plan, |
| 01:34.21 | mafm | 2) it's a more risky project I think, there are more chances of failure |
| 01:34.43 | brlcad | whether you're not accepted is dependent on the quality of your application (compared to quality of others), number of slots we accept (just 2-4), the topic being proposed, as well as any details about yourself and your abilities |
| 01:34.49 | mafm | oh, and 3) I don't have the infrastructure to test the distributed rendering |
| 01:35.01 | brlcad | okay, so forget it :) |
| 01:35.07 | brlcad | not pressing the idea on you :) |
| 01:35.14 | brlcad | pick whatever suits you |
| 01:37.48 | mafm | I don't feel pressed really, I just want to make explicit that it's more risky, in the case that I would be accepted and go ahead, I think that it could fail more easily |
| 01:38.06 | mafm | and I don't know if you could provide the infrastruture or what? |
| 01:38.40 | mafm | we have clusters (not very big) at my job, but they're in production and we cannot install new things |
| 01:38.45 | brlcad | heh, you're just saying all of the reasons for why you shouldn't submit it |
| 01:38.51 | brlcad | forget about it! :) |
| 01:42.29 | mafm | sighs |
| 01:42.40 | mafm | btw, I could submit more than one proposal, you now :D |
| 01:42.46 | brlcad | that you could |
| 01:43.01 | brlcad | but better to focus on getting at least one really solid |
| 01:43.19 | andrecastelo | have you guys received lots of applications?? |
| 01:43.21 | brlcad | i'd recommend doing one proposal and putting the extra effort on a better patch .. |
| 01:43.34 | brlcad | andrecastelo: not really |
| 01:43.56 | andrecastelo | and also - i see that you are part of the bzflags community as well, will you be a mentor both ?? |
| 01:44.04 | andrecastelo | be a mentor in* both.. |
| 01:45.08 | brlcad | i'm admin for both, I hope to actually *not* mentor for either (officially) if I can get away with it ... but then still mentoring as usual group-mentorshipwise |
| 01:45.52 | brlcad | if I do mentor, it's probably be for brl-cad |
| 01:46.17 | andrecastelo | i see |
| 01:48.07 | mafm | there are no assigned/expected mentors yet, I guess? |
| 01:48.21 | mafm | btw I hope that we have some extended period to submit the patch :) |
| 01:48.49 | brlcad | there's about another week to submit a patch |
| 01:52.44 | mafm | ah, good |
| 01:53.52 | brlcad | it will be required of any serious candidates |
| 01:54.03 | brlcad | i.e. any finalists for certain |
| 01:54.33 | brlcad | if someone can't make a basic patch out of a million line codebase, that bodes very poorly :) |
| 01:58.40 | mafm | sed -i 's/BRL-CAD/mafm-CAD/g' :) |
| 01:58.56 | brlcad | heh |
| 01:59.14 | mafm | that'd be at least 1000 lines changed |
| 02:00.09 | brlcad | that'd hit every file |
| 02:00.16 | brlcad | BRL-CAD is in the file header |
| 02:00.18 | mafm | btw, now I should still submit the proposal to the mailing list, or in parallel, or what? |
| 02:00.24 | brlcad | "yes" |
| 02:00.41 | brlcad | could give a summary to the mailing list |
| 02:00.55 | brlcad | then extra detail to the wiki or as a submission |
| 02:01.30 | mafm | what if I paste the complete submission? |
| 02:01.41 | mafm | in the wiki or so |
| 02:02.44 | brlcad | you can put whatever you like into the wiki |
| 02:03.03 | brlcad | good place for full/more detail |
| 02:04.07 | mafm | going to sell his wikispace for Google ads :P |
| 02:05.36 | andrecastelo | hey brlcad, i was looking at the TODO list and found this in the lower priority part: |
| 02:05.37 | andrecastelo | * center of presented area |
| 02:05.49 | brlcad | yes? |
| 02:06.07 | andrecastelo | what is it exactly ?? |
| 02:06.13 | andrecastelo | ( http://brlcad.svn.sourceforge.net/svnroot/brlcad/brlcad/trunk/TODO ) |
| 02:06.32 | brlcad | in a given 2D projection, where is the center of presented area |
| 02:07.42 | andrecastelo | is it simple enough to implement as the required patch ?? |
| 02:07.43 | brlcad | that task really souldn't take more than a couple days at best |
| 02:07.59 | brlcad | it's possible .. a bit heavy for a patch, might take a day |
| 02:08.17 | brlcad | (i.e. a few hours) |
| 02:08.45 | andrecastelo | hm i see.. |
| 02:10.28 | brlcad | but certainly would be impressive for a patch |
| 02:10.52 | brlcad | folks for bz have already submitted patches that are actually harder |
| 02:11.07 | mafm | I think that my router will reconnect, be right back :D |
| 02:11.14 | andrecastelo | really?? hmm.. |
| 02:11.19 | andrecastelo | we have one week to deliver, right ? |
| 02:11.26 | brlcad | yep |
| 02:11.38 | brlcad | seriously though -- make the application itself good first :) |
| 02:11.57 | brlcad | then work on the patch |
| 02:12.02 | andrecastelo | hm ok.. will do |
| 02:12.13 | andrecastelo | another thing |
| 02:13.56 | andrecastelo | i'm having a problem running the program.. tried compiling from source and at first i've got "Build: 133 succeeded, 1 failed.." |
| 02:14.10 | andrecastelo | then tried again and got " Build: 0 succeeded, 0 failed, 134 up-to-date, 0 skipped" |
| 02:14.21 | andrecastelo | tried running and got some problem with libbu.dll |
| 02:14.36 | brlcad | maybe that could be your patch ;) |
| 02:14.47 | brlcad | a build system patch, figure out what went wrong |
| 02:15.20 | *** join/#brlcad mafm_ (n=mafm@89-180-59-113.net.novis.pt) | |
| 02:16.41 | andrecastelo | hmm |
| 02:16.41 | brlcad | otherwise, that's not enough info to say anything useful, would have to know what failed, what the error was, maybe other factors |
| 02:17.00 | andrecastelo | yes, i was pasting it- http://rafb.net/p/61FfDf17.html |
| 02:17.56 | brlcad | it's trying to run asc2g to build our models and failing to find some library |
| 02:18.13 | brlcad | was that the original 1 failed? |
| 02:18.34 | andrecastelo | it fails to find libbu.dll.. but it compiled this library fine |
| 02:19.13 | brlcad | so needs a path added to the library search paths |
| 02:20.33 | andrecastelo | ok, let me give it a try |
| 02:23.42 | andrecastelo | the library compiled fine, i guess it's just a path problem |
| 02:26.30 | mafm_ | mafm: go away! |
| 02:27.21 | Daytona | hmm, two mafm's |
| 02:27.37 | andrecastelo | brlcad: http://rafb.net/p/jcpiTJ88.html |
| 02:27.43 | mafm_ | Daytona: as if one was not enough! |
| 02:27.49 | Daytona | LOL |
| 02:29.00 | mafm | that'll show him :) |
| 02:29.15 | andrecastelo | brlcad: found the problem - http://rafb.net/p/5XE59Z48.html |
| 02:29.43 | andrecastelo | i guess brlcad is not portuguese-friendly ? |
| 02:30.31 | brlcad | how so? |
| 02:30.57 | brlcad | it's failing because you don't have nsis installed -- that's used to build the installer |
| 02:32.04 | andrecastelo | brlcad: true, but technically it would go to another directory |
| 02:32.42 | brlcad | que? |
| 02:33.47 | brlcad | it'll go wherever it's configured to go |
| 02:34.13 | andrecastelo | hm ok, i'll get nsis |
| 02:36.14 | mafm | "and that full nonexclusive copyright will be assigned." -- what does that mean, that we have to assign copyright as FSF requires for their projects? |
| 02:39.18 | brlcad | mafm: yeah, we use single-copyright |
| 02:41.33 | andrecastelo | brlcad: it's a minor bug perhaps concerning non-english windows |
| 02:42.12 | andrecastelo | brlcad: because english windows defaults installations to C:\Program files, and every other language version has its own "Program Files" |
| 02:43.17 | brlcad | andrecastelo: perhaps |
| 02:43.21 | andrecastelo | brlcad: i installed nsis to the default directory and the same problem occurred.. therefore we need to edit the correct file to use a "program files" system var |
| 02:43.35 | brlcad | yeah |
| 02:44.06 | andrecastelo | but what is the correct file ? :b |
| 02:45.41 | andrecastelo | i don't think there's a way to do that though |
| 02:55.25 | mafm | the 3D engine to use is strongly preferred to be OGRE, anything, or not important to specify in the proposal? |
| 02:56.08 | hippieindamakin8 | wat is it rt now ? |
| 02:56.53 | mafm | hippieindamakin8: that's a question for me? |
| 02:57.12 | hippieindamakin8 | in general :) |
| 02:57.49 | mafm | that's from the project ideas, to make an OpenGL-based GUI instead of MGED |
| 02:58.01 | mafm | and OGRE listed as possible one |
| 02:58.56 | hippieindamakin8 | ohh .. |
| 02:59.10 | hippieindamakin8 | yaya rt now everything is based on MGED |
| 03:03.04 | brlcad | mafm: for the submitter to decide really, from my own experience, testing, and connections, ogre really seems to be one of the best fits |
| 03:04.43 | mafm | well, for me it's about the same if it's with CrystalSpace, OpenSceneGraph or OGRE |
| 03:04.48 | brlcad | but it could just as easily be a java3d interface with the right submittions |
| 03:05.01 | mafm | I see OGRE or OSG more suited for BRL-CAD as we've talked yesterday |
| 03:05.53 | mafm | but since for me it's about the same, it's better to let core developers decide I guess |
| 03:07.57 | mafm | so maybe I leave it open in the proposal and to be decided later? |
| 03:08.26 | mafm | btw andrecastelo, I think that you maybe this proposal is also well suited for you |
| 03:09.52 | andrecastelo | hmm |
| 03:10.16 | andrecastelo | i'm already studying the photonmap.c to see what has been done and what hasn't |
| 03:12.05 | brlcad | mafm: it could be decided later |
| 03:13.12 | mafm | another thing is, what's the functionality to replicate? e.g. 20% of MGED? |
| 03:13.16 | brlcad | more detail is generally better, but particularly if you don't have a strong justification then ogre would still be my preference as the one to start with (a LOT of time has already been spent evaluating the various options available) |
| 03:13.50 | mafm | or it's only the framework and the functionality to be filled later? |
| 03:13.54 | brlcad | yeah |
| 03:14.22 | brlcad | having a debug command console would be useful to have, logging services |
| 03:14.35 | brlcad | but mostly the console |
| 03:15.04 | brlcad | ability to get lists of geometry available, loading them, displaying various representations (wireframe, polygonal, etc) |
| 03:18.10 | mafm | loading them as in drawing a cube inside OGRE? |
| 03:18.47 | mafm | with the vertex data coming from the server backend? |
| 03:19.06 | brlcad | display lists, yes |
| 03:19.45 | brlcad | there is a client/server protocol being developed (concurrently) that will manage the geometry itself |
| 03:20.07 | brlcad | the client is a *thin* client, it just passes commands/events to the backend and gets stuff back |
| 03:20.49 | mafm | yep, but the client would have to convert the marshalled data to the 3d engine representation at list, right? |
| 03:21.21 | mafm | and the same to position the camera and things like that |
| 03:26.26 | brlcad | yep |
| 03:26.48 | brlcad | and/or make the protocol support reporting the data in the format that the engine requires |
| 03:27.40 | brlcad | camera state management is predominantly a client issue |
| 03:28.43 | mafm | brlcad: the protocol doing the translation you mean? that would be part of my project anyway, I guess |
| 03:35.22 | andrecastelo | hey brlcad.. how functional is photonmap.c ? |
| 03:35.50 | brlcad | andrecastelo: if you pray to the gods and sacrifice the right sized chicken heads, it works great |
| 03:36.11 | hippieindamakin8 | :)) paganism :P |
| 03:36.19 | andrecastelo | lol.. but do you know what exactly works and what doesn't? |
| 03:36.27 | brlcad | example photonmap rendering http://brlcad.org/tmp/moss_pm.png |
| 03:37.32 | andrecastelo | hm, that cube shadow is a little off isn't it ? |
| 03:38.18 | brlcad | andrecastelo: usability is pretty poor (it's complicated to use), effects through thin-walled objects is flawed (the photons bleed), performance sucks unless you put the objects in a box (all the photons go off into space) |
| 03:38.42 | brlcad | positionwise, no the shadows are right |
| 03:39.38 | brlcad | there's a question as to whether it should be light due to reflections, but otherwise right |
| 03:39.58 | brlcad | http://brlcad.org/tmp/moss_pm_2.png |
| 03:44.55 | hippieindamakin8 | cya ppl |
| 03:45.00 | andrecastelo | cya hippieindamakin8 |
| 03:45.00 | hippieindamakin8 | later |
| 03:45.40 | mafm | laterz :) |
| 03:48.03 | andrecastelo | brlcad: so, i'd have to rewrite part of it |
| 03:48.15 | brlcad | andrecastelo: possibly |
| 03:48.59 | brlcad | implementing a bidirectional path-tracer would pretty much trump photonmapping though |
| 03:49.05 | mafm | OK, I have everything but the plan :PPP |
| 03:49.22 | brlcad | or a path tracer with photomapping aspects |
| 03:49.33 | brlcad | or a path trace option to rt directly |
| 03:49.37 | andrecastelo | at least the kd-tree functions are already implemented.. those are usable, right ? |
| 03:49.49 | brlcad | yeah, you really don't worry about any of that |
| 03:50.17 | brlcad | even for implementing some other global illumination interface, you wouldn't worry about that |
| 03:50.29 | mafm | is there some talk in the mailing list, wiki or whatever of the main features needed by the GUI? |
| 03:50.54 | andrecastelo | i see.. well, i thought photon mapping was the most realistic method to implement GI |
| 03:51.07 | brlcad | andrecastelo: read up on http://brlcad.org/wiki/Developing_applications .. talks about some of the basics of the shotline interface |
| 03:51.21 | brlcad | no, photon-mapping isn't the most realistic |
| 03:51.26 | brlcad | it's the fastest |
| 03:51.47 | brlcad | at least it can be fast |
| 03:52.26 | brlcad | bidirectional path-tracing is a brute-force solution to the rendering equation, and converges generally to a great answer |
| 03:52.46 | brlcad | that's what most people see when they look at ultra-realistic images |
| 03:53.20 | brlcad | usually MLT |
| 03:54.24 | mafm | brlcad: is there some talk in the mailing list, wiki or whatever of the main features needed by the GUI? |
| 03:54.31 | brlcad | the brazil rendering system is probably one of the best MLT implementations (http://www.splutterfish.com/sf/WebContent/Index) |
| 03:54.43 | brlcad | mafm: not really, at least not yet |
| 03:55.13 | mafm | :S |
| 03:55.15 | brlcad | for now, just think of it as a thin graphical network client |
| 03:55.34 | mafm | yep, but I was trying to think about the milestones |
| 03:55.44 | brlcad | that needs to display stuff obviously, look up geometry, display it, spin it around, etc |
| 03:55.48 | andrecastelo | brlcad: those are amazing! |
| 03:56.11 | mafm | <slashdot> |
| 03:56.20 | mafm | 1) Create hello world |
| 03:56.22 | mafm | 2) ??? |
| 03:56.30 | mafm | 3) GSoC completed!! |
| 03:56.34 | mafm | </slashdot> |
| 03:56.35 | andrecastelo | 4) Profit!!! |
| 03:56.38 | andrecastelo | hehe |
| 03:57.02 | brlcad | andrecastelo: they were one of the first to make detailed pictures that seriously looked real |
| 03:57.27 | andrecastelo | i was trying to write milestones too, that's why i was asking about photonmap.c functionality |
| 03:57.59 | mafm | holy cow! those images are truly amazing |
| 03:58.10 | brlcad | the milestones aren't necessarily set in stone, it's mostly important to see that you've actually thought through a timeline and design process that we can then talk about |
| 03:59.17 | andrecastelo | what are the most important things in an app ? |
| 03:59.49 | andrecastelo | (for you guys) |
| 04:00.34 | mafm | why is the logging thing an important one? it's logging what commands go back and forth for debugging, or what? |
| 04:01.02 | brlcad | especially images like: http://www.splutterfish.com/sf/WebContent/Gallery/Image/283?category=2&perpage=24&page=7 and http://www.splutterfish.com/sf/WebContent/Gallery/Image/436?category=2&perpage=24&page=2 |
| 04:01.50 | brlcad | mafm: logging is one of those fundamental things nice to have sorted out early on, for testing, development, debugging, etc |
| 04:02.17 | brlcad | not critical for the app, I was just thinking of basic infrastructure items |
| 04:02.23 | andrecastelo | brlcad: omg, those are beautiful |
| 04:03.35 | brlcad | you can do some cool stuff with metropolis light transport, global illumination ;) |
| 04:03.55 | mafm | so milestones could be: logging, console, being able to communicate with the back-end, displaying geometries, rotating & positioning |
| 04:04.48 | mafm | displaying various representations (wireframe, polygonal, etc) |
| 04:07.03 | brlcad | plus maybe clean cross-platform build system integration, and input support (trackball, shift-grips) |
| 04:07.23 | andrecastelo | brlcad: so, what would be better, photon mapping or bidirectional path tracing ?? |
| 04:07.52 | brlcad | andrecastelo: if you have experience with neither, path tracing |
| 04:08.24 | andrecastelo | ok, i'll have to do some research to write the application though |
| 04:08.26 | brlcad | the algorithm for forward path-tracing is like 100-200 lines of code |
| 04:08.49 | brlcad | maybe 10x for full bidirectional, and then some for usability and render options |
| 04:10.03 | mafm | be right back |
| 04:12.44 | andrecastelo | brlcad: what's your timezone ? |
| 04:12.51 | brlcad | EDT |
| 04:13.05 | andrecastelo | how many hours from UTC ? |
| 04:13.11 | brlcad | -4 |
| 04:13.26 | andrecastelo | great |
| 04:13.50 | andrecastelo | i'll be back tomorrow in the morning to finish discussing the application ;) |
| 04:14.25 | andrecastelo | good night |
| 04:14.32 | andrecastelo | mafm: good night man, cya |
| 04:17.23 | brlcad | waves |
| 04:34.32 | mafm | sighs |
| 04:34.43 | mafm | maybe I'm lucky and get 3h of sleep today :) |
| 04:41.13 | mafm | brlcad: there it is |
| 04:41.28 | brlcad | there what is? |
| 04:42.43 | brlcad | ah, to the mailing list |
| 04:42.45 | brlcad | got it |
| 04:43.24 | mafm | and to the Google's webapp too |
| 04:43.54 | brlcad | looks good, gracias |
| 04:44.01 | mafm | yw :) |
| 04:48.34 | mafm | it's about time to sleep, night :) |
| 05:00.27 | *** join/#brlcad spik1 (n=spike@eridani.acm.jhu.edu) | |
| 05:56.31 | *** join/#brlcad clock_ (n=clock@217-162-111-241.dclient.hispeed.ch) | |
| 07:16.13 | *** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch) | |
| 07:23.18 | *** join/#brlcad homovulgaris (n=ca3fe93d@bz.bzflag.bz) | |
| 08:09.20 | *** join/#brlcad yukonbob_ (i=1000@s142-179-54-198.bc.hsia.telus.net) | |
| 08:09.20 | *** join/#brlcad homovulgaris (n=ca3fe93d@bz.bzflag.bz) [NETSPLIT VICTIM] | |
| 08:09.20 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] | |
| 08:09.20 | *** join/#brlcad ``Erik (i=erik@c-68-54-174-162.hsd1.md.comcast.net) [NETSPLIT VICTIM] | |
| 08:11.50 | *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz) | |
| 08:19.14 | *** join/#brlcad yukonbob_ (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] | |
| 08:19.14 | *** join/#brlcad homovulgaris (n=ca3fe93d@bz.bzflag.bz) [NETSPLIT VICTIM] | |
| 08:19.14 | *** join/#brlcad ``Erik (i=erik@c-68-54-174-162.hsd1.md.comcast.net) [NETSPLIT VICTIM] | |
| 10:01.05 | *** join/#brlcad Maloeran (n=maloeran@glvortex.net) | |
| 10:51.11 | *** join/#brlcad MinuteElectron (n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) | |
| 10:56.42 | *** join/#brlcad andrecastelo (n=pulse@189.71.62.248) | |
| 10:56.53 | andrecastelo | good morning everyone |
| 11:27.18 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) | |
| 11:58.11 | *** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch) | |
| 12:37.52 | andrecastelo | hey brlcad :) |
| 12:37.54 | andrecastelo | you there? |
| 12:47.03 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) | |
| 12:48.45 | mafm | hallo |
| 13:00.13 | *** join/#brlcad CIA-34 (n=CIA@208.69.182.149) | |
| 13:22.20 | brlcad | hallo |
| 13:34.54 | d_rossberg | brlcad: the tags/7.12.0 won't compile on windows (cpp issue) |
| 13:35.09 | d_rossberg | it looks like bob solved it for the trunk |
| 13:35.26 | brlcad | cpp issue? |
| 13:38.00 | d_rossberg | math.h is included in a cpp file but in an extern "C" section |
| 13:38.30 | d_rossberg | the problem is: math.h defines templates if it is included in a cpp file |
| 13:38.53 | d_rossberg | however, this won't work with extern "C" |
| 13:39.13 | brlcad | ah, yeah I think I actually may have fixed that one |
| 13:39.41 | d_rossberg | yes, it looks like it's fixed on the trunk |
| 13:40.37 | brlcad | the bigger question is why the .exe isn't working, haven't had a chance to poke at that yet |
| 13:41.54 | d_rossberg | which exe? the install.exe and mged.exe work on my computer (MS Windows XP) |
| 13:42.13 | brlcad | the 7.12.0 test binary |
| 13:42.37 | brlcad | ftp://ftp.brlcad.org/incoming/brlcad-7.12.0.exe |
| 13:42.40 | d_rossberg | the is what i was talking about |
| 13:42.50 | brlcad | wim couldn't get it to work |
| 13:43.20 | d_rossberg | what os does he use: xp or vista? |
| 13:43.35 | brlcad | course his e-mail apparently just died too -- him and another just got delisted from the mailing list due to bounced e-mails :) |
| 13:45.11 | d_rossberg | BUT: rt does not work (at least from mged), even a ray-trace to a file fails |
| 13:45.40 | brlcad | i believe wim was on xp |
| 13:45.56 | brlcad | what's rt fail with? runtime linker error? |
| 13:47.19 | d_rossberg | nor error, it says "Raytrace complete" |
| 13:49.22 | brlcad | then it's probably crashing |
| 13:49.34 | brlcad | does it produce any output? like the prep output? |
| 13:50.11 | d_rossberg | no, at least not on the mged window |
| 13:51.39 | brlcad | bob's finally back this week, I'll mention it to him |
| 13:55.11 | d_rossberg | maybe i could be more useful if i get my new VS running |
| 13:55.47 | d_rossberg | i got it today and at the moment i try to get my old projects running with it |
| 13:55.57 | d_rossberg | this could take some time ... |
| 14:00.59 | brlcad | nods |
| 14:02.20 | brlcad | feel free to comment on any of the submissions .. |
| 14:29.51 | *** join/#brlcad andrecastelo (n=pulse@189.71.62.248) | |
| 14:29.59 | brlcad | hello andrecastelo |
| 14:30.01 | andrecastelo | hi everyone |
| 14:30.03 | andrecastelo | hey brlcad |
| 14:30.10 | brlcad | responded to your submission |
| 14:30.14 | andrecastelo | just read the comment on the app |
| 14:30.21 | andrecastelo | yeah, just saw it ;) |
| 14:30.50 | andrecastelo | you want more detail about how MTL would interact with BRL-CAD's existing libraries, right ? |
| 14:31.08 | brlcad | yeah, just a little more detail about how it ties to us |
| 14:31.36 | andrecastelo | hm, ok |
| 14:31.38 | brlcad | you've shown reasonable competency in the task itself, but there's not much in the way of how it integrates with our code |
| 14:32.04 | andrecastelo | hm, ok, i'll try to detail further and it'll be ready in a minute |
| 14:32.20 | brlcad | what library/routines would you possibly be using for example -- does it leverage existing functionality (or not) |
| 14:33.43 | brlcad | emphasis on most tools are tools that *integrate* with brl-cad cleanly, that aren't just stand-alone apps being added to an otherwise already massive codebase, replicating existing functionality for example |
| 14:34.12 | brlcad | also just to show me that you have at least done your homework on what we actually already provide |
| 14:35.08 | brlcad | you also never explicitly mention what language you'd be working in or forsee implementing this in |
| 14:35.14 | brlcad | i.e. more implementation detail ;) |
| 14:35.22 | brlcad | otherwise a good submission |
| 14:35.33 | andrecastelo | hm ok, i thought C was implicit :b |
| 14:35.53 | andrecastelo | i'll include these details, though |
| 14:36.08 | brlcad | C or C++ is preferred, but not exactly required -- that's a design decision |
| 14:36.27 | brlcad | I'd be perfectly willing to consider another language if someone makes a strong valid case (good luck with that :) |
| 14:37.00 | andrecastelo | brlcad: is it possible to implement that in C++ ? |
| 14:37.14 | brlcad | yes |
| 14:37.20 | andrecastelo | i mean, wouldn't it kind of conflict with existing code.. ? |
| 14:37.38 | brlcad | it would have to live in a directory of its own, but no |
| 14:37.45 | andrecastelo | ok, i'll add some justification why C++ would be better then |
| 14:37.45 | brlcad | we already have some C++ interfaces |
| 14:38.11 | brlcad | you should know where we use C++ already :) |
| 14:39.42 | brlcad | if you do propose C++, I'd also expect you to speak of basic performance profiling as well |
| 14:40.43 | andrecastelo | performance profiling ? |
| 14:41.31 | brlcad | sanity checks to make sure it's not spinning its wheels in abstractions |
| 14:43.16 | andrecastelo | hm ok ok.. |
| 14:44.29 | brlcad | path tracers are very expensive, the same really holds true for a C implementation too -- but even more so for C++ |
| 15:01.28 | andrecastelo | hey brlcad - would it be better for me to reuse some of the ray tracer code for the path tracer or just write one from scratch ? |
| 15:01.37 | brlcad | heh |
| 15:01.41 | brlcad | which do you think |
| 15:02.28 | andrecastelo | reuse the code for a prototype version |
| 15:02.31 | brlcad | "10:33 <@brlcad> emphasis on most tools are tools that *integrate* with brl-cad cleanly, that aren't just stand-alone apps being added to an otherwise already massive codebase replicating existing functionality for example" |
| 15:02.33 | andrecastelo | then extend it to become fully functional |
| 15:02.43 | brlcad | we already shoot rays |
| 15:02.57 | brlcad | one of the specific intents is to use librt for ray shooting |
| 15:03.50 | brlcad | see the links I posted yesterday about developing brl-cad applications |
| 15:03.58 | andrecastelo | alright |
| 15:14.15 | *** topic/#brlcad by brlcad -> BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || Channel logs at http://ibot.rikers.org/%23brlcad/ || BRL-CAD is participating in the 2008 Google Summer of Code, see http://brlcad.org/wiki/Google_Summer_of_Code and join the developer mailing list || GSoC application submissions are (STILL) open with a one-week extension, submission deadline is now April 7th | |
| 15:14.52 | brlcad | kudos to those of you that got your submissions in on time :) |
| 15:16.50 | andrecastelo | brlcad: this week will be really useful for me to get familiar with the code and detail it even further ;) |
| 15:17.52 | brlcad | andrecastelo: glad to hear it :) |
| 15:27.04 | andrecastelo | brlcad: hm, so, i updated the application with the details you wanted ;) |
| 15:27.22 | andrecastelo | i'm open to critics and suggestions now |
| 15:32.50 | *** join/#brlcad Jai (n=Jai@202.3.77.38) | |
| 15:37.40 | mafm | 1 week extensions??? |
| 15:39.27 | andrecastelo | mafm: yup :) |
| 15:39.49 | mafm | pff |
| 15:39.56 | andrecastelo | why pff ? |
| 15:39.59 | mafm | and I only slept 3h tonight because of that :P |
| 15:49.59 | brlcad | mafm: hehe, well it does speak well for those that got their applications in on time |
| 15:50.08 | brlcad | it's an edge |
| 15:52.54 | mafm | no need to send a box of Oporto wine to the mentors then? |
| 15:53.02 | brlcad | hehe |
| 15:53.35 | brlcad | never said he could or could not be bribed with women and alcohol |
| 15:56.06 | mafm | let's just keep the alcohol thing, you wouldn't want to meet most portuguese women :P |
| 15:56.28 | andrecastelo | mafm: great, i'll supply the women |
| 15:56.32 | brlcad | :) |
| 15:56.37 | andrecastelo | :P |
| 15:57.56 | *** join/#brlcad Daytona (n=jra@c-68-55-36-65.hsd1.md.comcast.net) | |
| 15:58.22 | Daytona | brlcad: I just saw that the application extension is official |
| 16:00.57 | mafm | that's a real news flash :) |
| 16:01.27 | brlcad | heh |
| 16:01.46 | brlcad | drive by daytona! |
| 16:12.55 | mafm | <lh> brlcad: you are cruel and vicious |
| 16:12.56 | mafm | lol |
| 16:13.11 | mafm | I think that brlcad is served in the women dept. |
| 16:13.21 | brlcad | lubs the leslie |
| 16:13.46 | mafm | no luck for you andrecastelo, I think that you might better try with brazilian coffee or something :) |
| 16:14.52 | brlcad | mafm: the comments also take on quite a different tone when you've met many of them :) |
| 16:15.10 | *** join/#brlcad vedge (i=vedge@vedge.org) | |
| 16:15.36 | mafm | did you go to the mentor summit last year? |
| 16:15.44 | brlcad | yeah |
| 16:15.51 | brlcad | fun stuff |
| 16:15.54 | mafm | btw, how many applications are there? ~20 for CrystalSpace |
| 16:16.42 | brlcad | fewer, but close hold until the deadline :) |
| 16:17.42 | mafm | probably they'll get more too, it was less than 20 yesterday, but probably there will be 5 or 10 today before the former deadline |
| 16:17.51 | brlcad | the interesting thing is that right now there are probably 4 really strong applicants and we were only going to accept at most 4 ... :) |
| 16:18.35 | mafm | how so? don't you want to accept more, if they give you the chance? |
| 16:18.36 | brlcad | hopefully the apps keep improving, it'll be a tight competition |
| 16:18.50 | brlcad | no no |
| 16:19.28 | brlcad | i just saying that 4 are *really* strong, so if there wasn't an extension -- the list of who gets selected is pretty straightforward |
| 16:20.02 | brlcad | now with an extension, it just becomes more competitive .. if more really strong apps come in |
| 16:20.26 | hippieindamakin8 | good morning fellows :) |
| 16:21.12 | mafm | I see :) |
| 16:21.48 | andrecastelo | brlcad: how many mentors does brlcad have ? |
| 16:21.59 | andrecastelo | that seems to be the student cap, right ? |
| 16:22.19 | brlcad | we have 6 mentors at present |
| 16:22.46 | brlcad | with two potential backups |
| 16:22.57 | brlcad | but no, that's not the cap |
| 16:23.16 | brlcad | they can't all be managed effectively with that many -- that's why we're looking at just 2-4 |
| 16:23.42 | andrecastelo | but the organizations do define a max number of students, right ? |
| 16:23.42 | brlcad | takes time to integrate, review code, plan, etc |
| 16:23.46 | brlcad | right |
| 16:23.48 | andrecastelo | i see |
| 16:23.49 | brlcad | our max is 4 |
| 16:24.01 | brlcad | i don't want more than 4 for brl-cad this year |
| 16:24.12 | andrecastelo | i hope i can make my application compete with those 4 strong ones you talked about :) |
| 16:24.45 | brlcad | that's also why having students speak to what interests them in the long run becomes useful, we'd like new long-term developers not just projects worked on |
| 16:25.47 | brlcad | here's an example of a really strong application received for bzflag, http://my.bzflag.org/w/User:Jude- |
| 16:26.33 | brlcad | detailed, careful thoughts on implementation, ties to our code, timelines, etc |
| 16:31.29 | hippieindamakin8 | mafm are u dawn thomas ? |
| 16:32.25 | hippieindamakin8 | ohh its somebody else from india |
| 16:39.01 | mafm | nope, I'm Manuel A. Fernandez Montecelo (hence the acronym of my nick, not the most creative thing really) |
| 16:41.45 | hippieindamakin8 | :) |
| 16:43.49 | yukonbob_ | waves in |
| 16:44.00 | yukonbob_ | ohh --- I'm a ghosty |
| 16:52.23 | brlcad | howdy yukonbob_ |
| 16:52.28 | yukonbob_ | :) |
| 16:52.44 | yukonbob_ | doesn't have his tail-removing instructions with him atm :P |
| 16:53.07 | yukonbob_ | it's a msg to nicserv, but I don't remember what... |
| 16:56.42 | yukonbob | what's the status of the GSoC -- on schedule, or extended? |
| 16:59.38 | brlcad | points at topic ;) |
| 17:20.11 | *** join/#brlcad dtidrow (n=dtidrow@c-69-255-182-248.hsd1.va.comcast.net) | |
| 17:23.03 | ``Erik | erm |
| 17:23.06 | ``Erik | why? |
| 17:23.18 | ``Erik | lack of solid submissions? :( |
| 17:24.45 | ``Erik | <-- thought the expectation was ~50 "good" submissions |
| 17:25.13 | ``Erik | no, I have not been reading my email. |
| 17:25.14 | ``Erik | :D |
| 17:47.18 | mafm | 50 good submissions?? |
| 17:48.06 | mafm | that's a lot |
| 17:51.51 | *** join/#brlcad Elperion (n=Bary@p54874DCE.dip.t-dialin.net) | |
| 17:54.58 | ``Erik | mebbe I had pie in the sky aspirations |
| 18:21.11 | hippieindamakin8 | i am done making a proposal |
| 18:21.48 | mafm | about what? |
| 18:36.27 | mafm | hippieindamakin8: ? |
| 18:48.35 | hippieindamakin8 | OOP geometry |
| 18:49.23 | hippieindamakin8 | its about OOP geometry engine mafm |
| 18:57.09 | *** join/#brlcad clock_ (n=clock@77-56-79-63.dclient.hispeed.ch) | |
| 19:01.05 | mafm | ah ok |
| 19:01.21 | mafm | I had thought about that project too |
| 19:04.01 | yukonbob | re: topic, me nods, points out that the irrsi topic-line across top doesn't show novels ;) |
| 19:07.46 | hippieindamakin8 | i might put in a proposal for the web repository as well |
| 19:17.48 | yukonbob | wonders, is there an official msg from Google re: this extension -- on the gsoc (code.google.com/...) the timelines still show the deadline as 31Mar |
| 19:18.09 | brlcad | ``Erik: given how similar i'm seeing our rate is to bz's, that would have resulted in about 25 "good" ones |
| 19:18.44 | brlcad | thinks ``Erik should comment on the submissions more |
| 19:19.38 | mafm | yukonbob: do you mean if the extension is for real? |
| 19:20.31 | yukonbob | mafm: see brlcad's note to me re: topic |
| 19:20.33 | brlcad | yukonbob: depends how big your window is too :) |
| 19:20.52 | brlcad | sees it all |
| 19:21.08 | yukonbob | brlcad: heh -- not all of us are developing on 16-way multi-gpu chromium-equipped workstations ;) |
| 19:21.22 | brlcad | mmm.. 16-way |
| 19:21.34 | yukonbob | laughs |
| 19:21.39 | yukonbob | what a bunch of nerds we are. |
| 19:26.41 | yukonbob | looks for discussion brlcad and self had re: other projects... |
| 19:27.47 | brlcad | thinks yukonbob should mentor |
| 19:28.38 | mafm | why do you keep saying "re: something"? I don't understand at least half of your phrases |
| 19:28.53 | brlcad | re: == regarding |
| 19:29.45 | yukonbob | brlcad: my heart has always been 100% agreeable to that, but I've held off just for logistic reasons... |
| 19:29.52 | yukonbob | bets he _will_ mentor |
| 19:29.59 | yukonbob | do I need to register or anything? |
| 19:30.08 | brlcad | you would need to, yes |
| 19:30.08 | yukonbob | (by some deadline?) |
| 19:30.12 | mafm | ah ok |
| 19:30.24 | brlcad | by then end of the week |
| 19:30.36 | yukonbob | fine, fine ;) |
| 19:30.39 | brlcad | er, by the time we start ranking actually |
| 19:30.47 | brlcad | which is probably the day apps close |
| 19:30.58 | brlcad | I can't add you though, you have to do it |
| 19:31.23 | yukonbob | ok -- I'll do that now... |
| 19:31.33 | mafm | I guess that they'll extend it for another two weeks or something :P |
| 19:32.11 | yukonbob | mafm: 1 week |
| 19:33.02 | brlcad | yeah, just one week, mentors then still have their 12+ days to deliberate, rank, and ask for more detail |
| 19:33.25 | brlcad | then there's a community bonding time frame for about 5 weeks |
| 19:34.11 | mafm | just kidding, I was saying that by the time that the new deadline is reached, maybe they extend it for another 2 weeks :) |
| 19:34.16 | brlcad | where students get familiarized, hopefully start committing, learn what they need to |
| 19:34.24 | brlcad | ahh, heh, not likely |
| 19:35.04 | brlcad | 4300 or so applicants was probably enough to stay on schedule -- adding another week it just going to make it a heck of a lot more competitive |
| 19:35.22 | brlcad | as there will likely be 6000-8000 applicants I bet |
| 19:36.04 | brlcad | which means for a group like ours, we're going to have anywhere from 8-16 people competiting for 2-4 slots |
| 19:37.03 | yukonbob | finishes mentor application |
| 19:39.07 | brlcad | wonders wtf Avinash Dubey is |
| 19:39.18 | brlcad | rejects |
| 19:44.32 | mafm | I go now, night |
| 19:45.21 | *** join/#brlcad spike_ (i=[U2FsdGV@centaur.acm.jhu.edu) | |
| 19:56.22 | yukonbob | has to log-off temporarily (battery) -- chat later, folks |
| 20:51.05 | *** join/#brlcad Elperion (n=Bary@p54874DCE.dip.t-dialin.net) | |
| 21:33.20 | yukonbob | returns, like he never even left (thx, screen(1)) |
| 21:50.16 | yukonbob | wonders (speaking of meta-projects like GSoC and BRL-CAD) -- is there any progress on Coverity and BRL-CAD? |
| 21:53.25 | brlcad | i emailed dave yet again, no response |
| 21:57.08 | hippieindamakin8 | hey brlcad are u there |
| 21:57.09 | hippieindamakin8 | ? |
| 21:57.30 | brlcad | nope |
| 21:57.38 | brlcad | ~ask |
| 21:57.39 | ibot | Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there, just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. |
| 21:58.58 | yukonbob | likes the "against our will" description |
| 21:59.51 | hippieindamakin8 | ya wanted to submit my proposal :) |
| 21:59.54 | yukonbob | receives cattle prod, and is instructed to refrain from meta-comments |
| 22:00.25 | yukonbob | hippieindamakin8: submit it, and it will be commented on |
| 22:00.47 | hippieindamakin8 | :) ya |
| 22:01.29 | hippieindamakin8 | brlcad can the user define his primitives ? |
| 22:01.54 | brlcad | huh? |
| 22:02.06 | yukonbob | hippieindamakin8: in what sense? Primitives are "primitive" -- they're the core bits that "stuff" is made from... |
| 22:02.32 | hippieindamakin8 | like in AUTOCAD inventor 3d modelling and solidworks .. i add 2 already existing primitives and they can define it as another primitive |
| 22:02.36 | hippieindamakin8 | for his later worl |
| 22:02.46 | hippieindamakin8 | *work |
| 22:03.00 | yukonbob | read vol II, iirc about regions, etc. |
| 22:03.17 | hippieindamakin8 | smthing like.. a template |
| 22:06.46 | brlcad | yes, you can do that -- they're not called primitives, though -- they are combinations |
| 22:07.01 | brlcad | primitives are the most fundamental object type |
| 22:10.57 | hippieindamakin8 | ohk.. like when i used AUTOCAD, it calls it user defined primitive |
| 22:13.22 | hippieindamakin8 | just sent the proposal |
| 22:14.13 | yukonbob | s/vol II/ Vol III/; s/regions/combinations and regions/ -- (http://brlcad.org/wiki/Documentation) |
| 22:15.37 | hippieindamakin8 | thanks yukonbob |
| 22:15.39 | yukonbob | np |
| 22:15.49 | yukonbob | ~doc |
| 22:15.49 | ibot | Documentation can be found at http://digium.com/index.php?menu=documentation or http://www.digium.com/handbook-draft.pdf or http://www.voip-info.org/wiki-Asterisk or http://www.asteriskdocs.org or http://www.oreilly.com/catalog/asterisk |
| 22:16.00 | yukonbob | hrm... |
| 22:16.12 | yukonbob | ^-- not for you, hippieindamakin8 |
| 22:16.14 | yukonbob | : |
| 22:16.17 | yukonbob | : |
| 22:16.21 | hippieindamakin8 | ya got it :) |
| 22:16.25 | yukonbob | :), rather |
| 22:16.44 | hippieindamakin8 | hates lab reports and open office the most |
| 22:16.57 | yukonbob | hates bad coffee |
| 22:20.21 | hippieindamakin8 | is waiting for comments from brlcad and yukonbob |
| 22:20.38 | brlcad | hippieindamakin8: be patient |
| 22:20.47 | hippieindamakin8 | ohk :) i am |
| 22:20.49 | brlcad | it might take hours or even days |
| 22:21.08 | hippieindamakin8 | that was just a message .. i shall get back to my lab reports |
| 22:21.11 | brlcad | work on a patch in the meantime :) |
| 22:21.27 | hippieindamakin8 | yaya.. hey is the nsis installer working properly? |
| 22:21.47 | brlcad | the patches will probably be a huge deciding factor given the extension |
| 22:21.53 | hippieindamakin8 | i wanted to do that though :P |
| 22:21.53 | brlcad | works for me |
| 22:22.05 | hippieindamakin8 | ya i understand :| |
| 22:22.15 | brlcad | could do the localization fix that was mentioned yesterday |
| 22:22.22 | brlcad | so it doesn't just assume "Program Files" |
| 22:22.30 | hippieindamakin8 | ohk... |
| 22:22.34 | hippieindamakin8 | i ll have to chk.. |
| 22:22.56 | hippieindamakin8 | i have no windows rt now.. i shall get to it in a day or 2 |
| 22:23.22 | hippieindamakin8 | wats the new deadline for the submission of patches ? |
| 22:24.59 | yukonbob | hippieindamakin8: you mean for GSoC proposals? 7 Apr. |
| 22:25.09 | hippieindamakin8 | no the patches |
| 22:25.59 | brlcad | technically they're optional, but you'll not likely get selected if you don't provide a patch given how competitive it's going to be |
| 22:26.35 | brlcad | given we start ranking on the 7th, it's in your best interest to have something done by then or very shortly thereafter |
| 22:31.48 | hippieindamakin8 | ya i sure want to submit a patch |
| 22:32.27 | yukonbob | go for it!! |
| 22:32.28 | brlcad | you sure do |
| 22:32.43 | yukonbob | cheers "submit a patch! submit a patch!" |
| 22:34.14 | yukonbob | also gets a glass of water... |
| 23:10.09 | poolio | hippieindamakin8: Submit a documentation patch :) |
| 23:10.18 | poolio | brlcad: how's the gsoc going? |
| 23:11.27 | hippieindamakin8 | cool :) |
| 23:12.09 | hippieindamakin8 | ya seen them in the list of TODOs |
| 23:12.17 | brlcad | code patch would probably be preferred since one of the points is to see how you code, but yeah, even a documentation patch would work :) |
| 23:12.25 | brlcad | poolio: fun fun |
| 23:12.33 | brlcad | some good proposals to evaluate |
| 23:12.36 | hippieindamakin8 | i shall try for a technical patch first and then i shall come up with docs when i am unable to do that |
| 23:14.01 | hippieindamakin8 | i shall be submitting it by 9th .. will it do ?? too many academic committments rt now with the end semester examinations approaching |
| 23:14.01 | brlcad | looking at the bug and feature requests on sf.net is also recommended |
| 23:14.06 | hippieindamakin8 | i am trying to keep a track of them |
| 23:43.44 | hippieindamakin8 | brlcad: Sean w.r.t ur slashdot post |
| 23:44.04 | brlcad | que? |
| 23:44.05 | hippieindamakin8 | <PROTECTED> |
| 23:44.19 | hippieindamakin8 | :) |
| 23:44.20 | brlcad | doesn't mean computer science degree students |
| 23:44.34 | hippieindamakin8 | i aint :) and it is open to all |
| 23:45.07 | brlcad | the fact that they are working on code as stucents makes them "computer science students" over the course of the summer |
| 23:46.00 | hippieindamakin8 | :) thats true |
| 23:46.08 | poolio | brlcad: at some schools that's an insult |
| 23:46.42 | brlcad | the wording is actually based off of google's own original announcements from chris |
| 23:46.56 | hippieindamakin8 | ya my school considers my contribution to the firefox and imageprocessing community a big waste of time :P |
| 23:47.18 | brlcad | more power to them if they're insulted, less rediculous people to have to deal with during the review process |
| 23:47.29 | hippieindamakin8 | its ohk.. :) wat Sean wrote is actually rt but literally not :) |
| 23:47.40 | brlcad | you're missing the point |
| 23:47.47 | brlcad | literally it is the case, it doesn't say degree |
| 23:47.59 | hippieindamakin8 | aah yeah |
| 23:48.12 | hippieindamakin8 | ya that is ppl interested in CS |
| 23:49.33 | hippieindamakin8 | my school is completely ridiculous in the point that when enrolled in mechanical engg u are just supposed to take mechanical engg courses when given an option for an elective |
| 23:50.49 | yukonbob | isn't that point neither here not there, though -- the only way the school (maybe) needs to be involved is to verify you're enrolled in school... you could get a job selling ice-cream for the summer if you like, too... |
| 23:51.19 | yukonbob | ...whether it's related to your studies or not, or whether the school likes it or not... |
| 23:51.23 | hippieindamakin8 | hehe |
| 23:51.33 | brlcad | yep, in fact that's exactly what one of the bzflag gsoc students did last year |
| 23:52.17 | brlcad | (and he was ultimately failed due to lack of progress and inadequate time put forth) |
| 23:52.45 | yukonbob | heh |
| 23:53.25 | yukonbob | off-topic: how suitable would the bzflag framework be to building a virtual world, a la second life... perhaps one could have a project setting up a virtual ice-cream shop? |
| 23:53.39 | brlcad | heh, doable |
| 23:53.50 | yukonbob | makes note |
| 23:53.50 | poolio | yukonbob: Just out of curiousity, for what purpose would this virtual world exist? |
| 23:54.01 | brlcad | there's no dynamic geometry right now, that'd require modifications |
| 23:54.04 | yukonbob | feh -- same reason second life exists, I guess... |
| 23:54.13 | poolio | Although there are probably a ton of better frameworks out there. |
| 23:54.19 | yukonbob | sure |
| 23:54.27 | poolio | yukonbob: entertainment? There are some AGI people trying to hook up to virtual worlds for the purpose of learning |
| 23:54.31 | brlcad | blasphemy! |
| 23:54.41 | poolio | ... |
| 23:54.42 | brlcad | nothing is better than bzflag |
| 23:54.45 | brlcad | except brl-cad |
| 23:54.46 | yukonbob | laughs -- brlcad is agitated!!! |
| 23:54.51 | hippieindamakin8 | :)) |
| 23:54.52 | poolio | brlcad: Can I quote you on this? |
| 23:55.01 | brlcad | hehe, sure |
| 23:55.34 | brlcad | gets craving for chinese food |
| 23:56.26 | yukonbob | is interested in machinema (sp?) -- perhaps brl-cad and bzflag could provide the bulk of the framework to make a video... |
| 23:56.37 | yukonbob | meow |
| 23:57.30 | yukonbob | does bzflag support any easy-to-do articulation? (/me thinks "no") |
| 23:57.48 | brlcad | bzflag doesn't support any articulation |
| 23:57.58 | brlcad | "no dynamic geometry" |
| 23:58.04 | yukonbob | GSoC ;) |
| 23:58.19 | brlcad | not something I'd rank high on our priority list :) |
| 23:58.56 | yukonbob | can hardly wait to get a 'puter w/ hardware acceleration to get bzflag installed again... |
