Re: [Edu-sig] The Fate of VPyton
Did a litttle googling on the build issues... enough to determine that projects using boost.python seem to be finding that Scons makes the best alternative to a building mechanism of bjam, make or some combination thereof. For example: http://mapnik.org/documentation/install/ Of course with the additional advantage that Scons is Python, so that whatever difficulties are posed, at least it is not learning a new syntax, as someone like myself faces with make or bjam. Willing to invest in learning something about C++ and the Python C API. Enough already. So we are adding Scons build to the next, so far imaginary, generation of vpython. "We" are? Art
ajsiegel@optonline.net wrote:
So we are adding Scons build to the next, so far imaginary, generation of vpython.
"We" are?
Art, I'm impressed by your determination... Not knowing much about this at a code level, but I thinking more than ever that it may be a *really* good idea you look hard at Blender. I've posted edu-sig several times previously, little or no feedback. No matter, but given the lengths you now seem to prepared to go to to rescue VPython, I'd be amazed if that effort might not be better applied to FIRST researching Blender, then perhaps brining your considerable skills to that table. It's a sophisticated active platform to build upon - and it renders beautifully ! Lots of "we" there -- good community including in France. OpenGL, cross platform, GPL license, For many Blender-ers it's their determined creative auto-didact motivation to 'do stuff in 3D' that drags them into learning applied Python. For example documentation help and recently a distributed network renderer :-) Python is embedded in Blender. Plus GameKit and 3D web plug in [windows] FEATURES http://www.blender3d.org/cms/Features.155.0.html FORUM http://blenderartists.org/forum/forumdisplay.php?f=11 and much more.. good luck Jason
From: Jason Cunliffe
Not knowing much about this at a code level, but I thinking more than ever that it may be a *really* good idea you look hard at Blender. I've posted edu-sig several times previously, little or no feedback.
I have looked at Blender, and am aware - in broad terms - at the strength of the surrounding community. And have some across Blender used in serious educational settings as well. It's just a different beast than vpython. My commitment to vpython is in some sense a commitment to lightweight simplicity where lightweight simplicity will do. It's the same song I have been singing here from day one. Maybe I am a moralist, puritan on these matters. There is something gluttonous, and in contradiction to the kind of principles we should be promoting in an educational setting, by not emphasizing the just enough to get the job done, and nothing more. And our tools should be setting some example. Less *is* more would make a good norm, with the exception to that rule needing to find a justification for itself. Seems like we have migrated to the opposite of that way of thinking in many ways. I celebrate technology less than you do, I think, because I hold it responsible for a good amount of this particular kind of damage.. There are solutions to limiting that damage, but damn it, none of them are themselves technology based. Which seems, almost by definition, to make them uninteresting, off-topic, anti-progressive, etc. When of course I see just the opposite being true. How did we get here? I am not contrarian just ot be so. Its coming from belief, and that belief is coming from deep within my sensibility. I am stuck with my sensibility, but have no reason to disrespect it. So that is the commitment I think you are seeing with me and vpython. And I am surprised you don't understand that after hearing me spout off here for as long as I have. The geometry with which I am fascinated is all about the universe of exploration that can be done essentially limited to the concepts "point", "line", "plane". It has kept me intellectually busy for some time, with some help from python, vpython, and a good deal of reading and a good deal of working at it. Those tools are less than adequate for those exploring what exactly? Art
Art Thanks for your clear reply.. Yes. Having a reasonable sense of your sensibilities from edu-sig dialogues, I hesitated 3 times before even suggesting it. <personal memo> I range wide in my sensibilities and preferences. btw, I think I was perhaps the original poster to edu-sig about VPython. Was greatly influenced by a magnificent shell+3D window on Amiga c.1987 called "PageRender3D". Am fairly sure you would approve highly of that one. And then live GEOMETER implemented on SGI c.1993 [minimal and mind-blowing] Or REBOL for example for its less-is-more philosophy. In general prefer one-to-one encounters than groups anytime. Signal to noise is usually so much better. </personal memo> However we are entering the great RW [read-write] cooperative, collaborative epoch. Ideas and process having many roots in the world of [Open]software, networks of networks, virtuality [modeling/simulation/contemplation... in every order] These dynamics are now bubbling up out into social financial and political spheres [ah see - Pure Geometry purifies] meme-nets are thriving as never before My subversive proposal to you is along those lines. Optimistically, I imagined that maybe a VPython version of Blender is quite possible AND a good idea. The key paradigm is for a pure geometric minimal shell with live 3d viewer and modular architecture. small fast cheap beautiful user-controls I think using the Blender Python API it is possible to configure Blender to run as shell based Geometry [Game] Engine. You can ignore all the dancing girls and singing seals if you want. What you gain however may be like a laboratory for geometry [with physics::: particles, light, gravity, etc] Geometry meme -> Kepler elipses -> lead to orbiting planets. or geometry behind this http://en.wikipedia.org/wiki/Image:Milky_Way_2005.jpg And heres a video clip of experimental interface where minimal 2D geometry symbolizes physical mechanics [draw = think = link] [ancient Greeks making marks in the sand with sticks seashells and string] * MIT live drawing* http://www.youtube.com/watch?v=NZNTgglPbUA Won't be surprised if you tell me that you love it OR hate it :-) I understand why you don't want all the bells and whistles of Blender as it installs now. But if it turned out that all that stuff was in proactive very easy to turn-off ,and instead present a VPython face. Something installs well quickly cross-platform, how would you feel? Jason
Jason, I've looked into Blender a number of times, and while I would like to learn it, I have never had time for the learning curve. VPython is something my students and I can use immediately to solve problems at hand. If my goal were to become a 3D artist, then something like Blender would be the way to go. You suggest some sort of "streamlined" tool built on Blender is possible. But my recollection of Blender is that it embeds a Python intepreter, but is not itself a Python extension. If that is true, then I cannot use such a tool (even if streamlined) to augment _my_ Python environment. When I need a quick 3D visualization for something I'm working on, I would have to somehow "import" all my work, libraries, etc. into this Blender-based environment. I just don't see Blender and VPython in the same "niche." If I am wrong about this, please let me know. Perhaps I have not looked hard enough, or perhaps Blender has changed since I last visited it. --John On Friday 13 October 2006 4:04 am, Jason Cunliffe wrote:
Art
Thanks for your clear reply.. Yes. Having a reasonable sense of your sensibilities from edu-sig dialogues, I hesitated 3 times before even suggesting it.
<personal memo> I range wide in my sensibilities and preferences. btw, I think I was perhaps the original poster to edu-sig about VPython. Was greatly influenced by a magnificent shell+3D window on Amiga c.1987 called "PageRender3D". Am fairly sure you would approve highly of that one. And then live GEOMETER implemented on SGI c.1993 [minimal and mind-blowing] Or REBOL for example for its less-is-more philosophy. In general prefer one-to-one encounters than groups anytime. Signal to noise is usually so much better. </personal memo>
However we are entering the great RW [read-write] cooperative, collaborative epoch. Ideas and process having many roots in the world of [Open]software, networks of networks, virtuality [modeling/simulation/contemplation... in every order] These dynamics are now bubbling up out into social financial and political spheres [ah see - Pure Geometry purifies]
meme-nets are thriving as never before
My subversive proposal to you is along those lines. Optimistically, I imagined that maybe a VPython version of Blender is quite possible AND a good idea. The key paradigm is for a pure geometric minimal shell with live 3d viewer and modular architecture. small fast cheap beautiful user-controls
I think using the Blender Python API it is possible to configure Blender to run as shell based Geometry [Game] Engine. You can ignore all the dancing girls and singing seals if you want. What you gain however may be like a laboratory for geometry [with physics::: particles, light, gravity, etc]
Geometry meme -> Kepler elipses -> lead to orbiting planets. or geometry behind this http://en.wikipedia.org/wiki/Image:Milky_Way_2005.jpg
And heres a video clip of experimental interface where minimal 2D geometry symbolizes physical mechanics [draw = think = link] [ancient Greeks making marks in the sand with sticks seashells and string] * MIT live drawing* http://www.youtube.com/watch?v=NZNTgglPbUA
Won't be surprised if you tell me that you love it OR hate it :-)
I understand why you don't want all the bells and whistles of Blender as it installs now. But if it turned out that all that stuff was in proactive very easy to turn-off ,and instead present a VPython face. Something installs well quickly cross-platform, how would you feel?
Jason _______________________________________________ Edu-sig mailing list Edu-sig@python.org http://mail.python.org/mailman/listinfo/edu-sig
-- John M. Zelle, Ph.D. Wartburg College Professor of Computer Science Waverly, IA john.zelle@wartburg.edu (319) 352-8360
On Friday 13 October 2006 4:04 am, Jason Cunliffe wrote:
My subversive proposal to you is along those lines. Optimistically, I imagined that maybe a VPython version of Blender is quite possible AND a good idea.
Guess I am not seeing it. Vpython needs compatibility with numpy, and a few committed maintainers building distributions - Windows, deb, Mac X11 - and it is a viable project. A crossplatform windowing solution - like glut - would bring Mac native into the fold, and make ongoing development/maintenance easier. Eliminating the boost dependency would have advantages, if not too much is sacrificed. Either way, move on with the 4.xxx branch that Jonathan had begun bringing transparency, spot-lighting - etc. into the fold. Again, I simply don't see it as that far off. The fact that the community surrounding it is rather thin technically is its biggest vulnerability. I can contribute to trying to solve that by upgrading my own lower-level skills, and by trying to recruit some interest by others who have those skills already. That's *my* plan. The potential connection with blender I see is through the yafray rendering engine www.yafray.org wihich is connected to the blender project, but not exactly sure how. Yafray is a high quality rendering engine with an XML scene description language. It would be interesting if vpython could import and export from Blender via this XML. I had explored trying to implement this via PyGeo, but yafray XML is notoriously underdocumented at this point. If that improves, I would definititely try to do something. PyGeo already exports nicely to Pov-Ray, so the infrastructure to do this kind of thing is already there. Would be relatively trivial to add the yafray XML alternative. To me, among other things, it's just an excuse to get hands on with XML functionality - but maybe it would lead to something interesting.
MIT live drawing* http://www.youtube.com/watch?v=NZNTgglPbUA
Don't have the bandwidth available to me to give it a look from here. Hoepfully when I get back. Art
participants (3)
-
ajsiegel＠optonline.net -
Jason Cunliffe -
John Zelle