[Pharo-project] what are the frameworks to develop 3d interactive world like codecity in Pharo
Hi guys we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo? - Lumiere? - opengl binding? Stef
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this. Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf Cheers, Alexandre On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose. Stef On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
Hi Guys, I think that depending on your goals both approaches are good: - binding to OpenGL it would be very cool to have a working/solid solution in this area. As Henrik and Marcus said, native support would be great. If I remember correctly, Lumiere is not very usable. Perhaps with NativeBoost and its GL bindings could be a way to go. - software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow... Cheers, #Luc 2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr>
Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
I think that depending on your goals both approaches are good:
- binding to OpenGL it would be very cool to have a working/solid solution in this area. As Henrik and Marcus said, native support would be great. If I remember correctly, Lumiere is not very usable. Perhaps with NativeBoost and its GL bindings could be a way to go.
Would be cool, but frankly I do not think the Pharo community can provide a reliable, multi-platform and smooth integration of OpenGL in a near future. Many have tried before and there is not many success.
- software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow...
I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture. Mondrian handles 20 thousands nodes and it does not use quadtrees. @Stef: We studied it and we could not make lumiere work. Alexandre
2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr> Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
On 21.09.2011 13:50, Alexandre Bergel wrote:
I think that depending on your goals both approaches are good:
- binding to OpenGL it would be very cool to have a working/solid solution in this area. As Henrik and Marcus said, native support would be great. If I remember correctly, Lumiere is not very usable. Perhaps with NativeBoost and its GL bindings could be a way to go. Would be cool, but frankly I do not think the Pharo community can provide a reliable, multi-platform and smooth integration of OpenGL in a near future. Many have tried before and there is not many success.
- software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow... I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture.
Hey, it WOULD be cool way of visualizing a programs execution though ;) Think squares as classes, height of buildings as amount of instances, and some sort of traffic being message sends, camera panning automatically to catch the activity. For GC, you could have a tetris-effect, where some of the buildings height start blinking then disappears! The possibilities are endless! Hmm, it kind of sounds strangely familiar when I think about it... http://www.youtube.com/watch?v=x3XzPhdBx9g#t=143s Maybe that wasn't as ludicrous as I've always thought :) Cheers, Henry
Hey, it WOULD be cool way of visualizing a programs execution though ;) Think squares as classes, height of buildings as amount of instances, and some sort of traffic being message sends, camera panning automatically to catch the activity. For GC, you could have a tetris-effect, where some of the buildings height start blinking then disappears! The possibilities are endless!
Hmm, it kind of sounds strangely familiar when I think about it... http://www.youtube.com/watch?v=x3XzPhdBx9g#t=143s Maybe that wasn't as ludicrous as I've always thought :)
:-) Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
I prefer rez on PS2 This is a really cool game http://www.youtube.com/watch?v=A4EFNWe4mCc Stef On Sep 21, 2011, at 2:06 PM, Henrik Sperre Johansen wrote:
On 21.09.2011 13:50, Alexandre Bergel wrote:
I think that depending on your goals both approaches are good:
- binding to OpenGL it would be very cool to have a working/solid solution in this area. As Henrik and Marcus said, native support would be great. If I remember correctly, Lumiere is not very usable. Perhaps with NativeBoost and its GL bindings could be a way to go. Would be cool, but frankly I do not think the Pharo community can provide a reliable, multi-platform and smooth integration of OpenGL in a near future. Many have tried before and there is not many success.
- software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow... I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture.
Hey, it WOULD be cool way of visualizing a programs execution though ;) Think squares as classes, height of buildings as amount of instances, and some sort of traffic being message sends, camera panning automatically to catch the activity. For GC, you could have a tetris-effect, where some of the buildings height start blinking then disappears! The possibilities are endless!
Hmm, it kind of sounds strangely familiar when I think about it... http://www.youtube.com/watch?v=x3XzPhdBx9g#t=143s Maybe that wasn't as ludicrous as I've always thought :)
Cheers, Henry
On 21 September 2011 13:50, Alexandre Bergel <alexandre.bergel@me.com> wrote:
 I think that depending on your goals both approaches are good:
 - binding to OpenGL  it would be very cool to have a working/solid solution in this area.  As Henrik and Marcus said, native support would be great.  If I remember correctly,  Lumiere is not very usable.  Perhaps with NativeBoost and its GL bindings could be a way to go.
Would be cool, but frankly I do not think the Pharo community can provide a reliable, multi-platform and smooth integration of OpenGL in a near future. Many have tried before and there is not many success.
why so pessimistic? many applications using opengl without problems. why pharo has to be any different? i doubt that its hard to make opengl up and running. what is hard is to make sure that you can be on having certain features (opengl version), supported across all platforms. But that's not our fault. If someone using 10 years old OS with 10 years old opengl driver and hardware, why he would expect that it can run things which are written for today's machines?
 - software 3D (no hardware support)  As Alex said, for simple stuff it could be enough and it is easier to ensure portability.  But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow...
I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture. Mondrian handles 20 thousands nodes and it does not use quadtrees.
@Stef: We studied it and we could not make lumiere work.
Alexandre
2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr> Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
   - Lumiere?    - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel  http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel  http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- Best regards, Igor Stasenko.
why so pessimistic? many applications using opengl without problems. why pharo has to be any different?
i doubt that its hard to make opengl up and running. what is hard is to make sure that you can be on having certain features (opengl version), supported across all platforms. But that's not our fault. If someone using 10 years old OS with 10 years old opengl driver and hardware, why he would expect that it can run things which are written for today's machines?
There is nothing that prevent it from running. It just requires a lot of effort. I am the first one who wants to see a nice integration of OpenGL in Pharo. Alexandre
- software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow...
I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture. Mondrian handles 20 thousands nodes and it does not use quadtrees.
@Stef: We studied it and we could not make lumiere work.
Alexandre
2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr> Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- Best regards, Igor Stasenko.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
On 21 September 2011 15:22, Alexandre Bergel <alexandre.bergel@me.com> wrote:
why so pessimistic? many applications using opengl without problems. why pharo has to be any different?
i doubt that its hard to make opengl up and running. what is hard is to make sure that you can be on having certain features (opengl version), supported across all platforms. But that's not our fault. If someone using 10 years old OS with 10 years old opengl driver and hardware, why he would expect that it can run things which are written for today's machines?
There is nothing that prevent it from running. It just requires a lot of effort. I am the first one who wants to see a nice integration of OpenGL in Pharo.
no, you will be second one, after me :)
Alexandre
 - software 3D (no hardware support)  As Alex said, for simple stuff it could be enough and it is easier to ensure portability.  But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow...
I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture. Mondrian handles 20 thousands nodes and it does not use quadtrees.
@Stef: We studied it and we could not make lumiere work.
Alexandre
2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr> Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
   - Lumiere?    - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel  http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel  http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- Best regards, Igor Stasenko.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel  http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- Best regards, Igor Stasenko.
On Wed, 21 Sep 2011, Alexandre Bergel wrote:
I think that depending on your goals both approaches are good:
- binding to OpenGL it would be very cool to have a working/solid solution in this area. As Henrik and Marcus said, native support would be great. If I remember correctly, Lumiere is not very usable. Perhaps with NativeBoost and its GL bindings could be a way to go.
Would be cool, but frankly I do not think the Pharo community can provide a reliable, multi-platform and smooth integration of OpenGL in a near future. Many have tried before and there is not many success.
What are your problems with existing OpenGL bindings? Levente
- software 3D (no hardware support) As Alex said, for simple stuff it could be enough and it is easier to ensure portability. But be carreful of the size of your scenes because I think that Moose graphs could be huge and it could be really really slow...
I would not be that pessimistic. Visualizing software does not require animation, moving colored cameras and texture. Mondrian handles 20 thousands nodes and it does not use quadtrees.
@Stef: We studied it and we could not make lumiere work.
Alexandre
2011/9/20 Stéphane Ducasse <stephane.ducasse@inria.fr> Ok what is your experience with lumiere? I would like to see if we can have Plotz or another solution to display metrics in Moose.
Stef
On Sep 20, 2011, at 1:53 PM, Alexandre Bergel wrote:
I think that recoding a 3d engine is easier and much less frustrating that making Pharo talk to OpenGl. I am not aware of the last progresses made on that front, but I burnt 1 student on this.
Codecity visualizations are not complex, it is just a matter of displaying triangles and multiplying matrixes. A student here learnt Smalltalk that way: http://bergel.eu/download/papers/Jaca11a-Klotz.pdf
Cheers, Alexandre
On 20 Sep 2011, at 08:45, Stéphane Ducasse wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
On 20 September 2011 12:45, Stéphane Ducasse <stephane.ducasse@inria.fr>wrote:
Hi guys
we would like to propose a fun project and the idea would be to build a little codecity. Now what are the frameworks to develop 3d interactive world in Pharo?
- Lumiere? - opengl binding?
Stef
a lurker pops up!! Aren't OpenCobalt, http://www.opencobalt.org/ , http://en.wikipedia.org/wiki/Open_Cobalt, and Croquet, http://en.wikipedia.org/wiki/Croquet_Project, projects running along those lines? According to wikipedia OpenCobalt is the continuation of Squeak. -- Frank Church ======================= http://devblog.brahmancreations.com
participants (7)
-
Alexandre Bergel -
Frank Church -
Henrik Sperre Johansen -
Igor Stasenko -
Levente Uzonyi -
Luc Fabresse -
Stéphane Ducasse