Using real open source projects in the context of a University course can be a quite successful experience. Both from the didactic point of view and from a software production perspective.
Five years ago I decided to involve my students of the advanced computer graphics course at the Pisa University in the collective writing of an extensible mesh processing system that we called MeshLab with a plugin based architecture that allowed a easy to be managed compartmental development. Students get very involved and, beside the computer graphics hard core topics of the course, they learned a lot in term of collaborative development, ethics of sw development, open source licenses.
At the end of that course a working system was successfully distributed under a GPL.
Since then, every year, I have repeated this approach extending it, and with the helps of tens and tens of willing students the system evolved into a serious complex mesh processing system, GPL'd, multiplatform, included in ubuntu and that is used worldwide by ten of thousands of users in academic and commercial environments.
I found that the knowledge of participating to the development of a **real** system gives the students a really strong motivation to give their best.
The fact that their contribs will be released publicly with their name linked to the commits and listed in the official developer page was a strong incentive to do not cheat.
For most of them it was the first time that they were making something real (not only exercises) something with a purpose that was quite different from the standard "get the score" approach.
Many of the students continued to maintain their portion of code well beyond the course terms (some even after graduating).
I cannot but thanks my students for the dedication that they have shown in the projects.
Just google for MeshLab for more details of it.
P.
In the context of archiving games, accurate software emulation the whole HW underlying each game is the only solution.
Obviously it should be done in a open, portable, multiplatform way to ensure that it is a long term solution.
Mame and Mess ( mamedev.org ) has already shown that this approach is viable and practical.
I do not completely agree with the above post.
While it is true that there are fields where the cost of the sw is not a real issue it is also true that some of these software niches are filled by FOSS tools. For example in 3D printing/rapid prototyping one of the best tools around for cleaning up models is an open source tool, called MeshLab ( http://meshlab.sourceforge.net/ ) .
The concept is not completely new.
There is a vast literature on the reconstruction of 3D objects from image sequences, ( http://en.wikipedia.org/wiki/Structure_from_motion )
Beyond the well known Photosynth (that does NOT reconstruct a dense surface) there are other free tools that allow to create complete 3D objects from photo sequences.
Arc3d and MeshLab can provide tools for reconstructing full 3D objects like a href="http://meshlab.sourceforge.net/SnapMeshLab.AO2.png.
"> this one.
Another 5cent;
not only the story is a bit old, but there exist free/open source tools that provide the same services, i.e. reconstruction of 3D models from your uncalibrated photos.
Look at http://www.arc3d.be/ and to http://meshlab.sourceforge.net/;
The first one (arc3d) is a free-to-use service that allow people to upload photos and get back a set of 3D depth maps well as the
relative vantage points from where the uploaded images were taken and the
settings of the cameras.
The second one (MeshLab) is an open source tool that allow to load, inspect, process and integrate the data produced with the arc3D service, so that you can obtain a 'clean' 3D model and export it in a variety of common 3D formats.
So if someone wish to try by himself this kind of technologies, the tools are there.
The raised issue is not completely unfounded, but focusing on a more particular sector of the modeling scene the field offer even less choices. The current panorama of 3D modeling packages is quite tailored towards the needs of the majority of users (low and high poly modelers) and most of these packages offers less than satisfying experiences when used to manage the large unstructured meshes that come from 3D scanning technologies or used for rapid prototyping needs.
The meshes produced by these automatic technologies are typically huge (millions of triangles per object) and not organized into scene graphs; most modeling packages simply sit down on the specs of these objects. 3D scanning and rapid prototyping hardware is becoming more and more popular, devices very low priced are already on the market ( http://hardware.slashdot.org/article.pl?sid=07/04/ 05/144212 ) while on the software side there are almost no alternative to the high end commercial package in the 50k$ range (20 times the HW cost).
The only open source alternative on this sector of 3d modeling is Meshlabhttp://meshlab.sourceforge.net/.
Sorry for what could seem a shameless plug, but I would like to hear your comment on this side of 3D modeling scene.
Collada Academic, not only games
on
Collada
·
· Score: 1
Beside big players, like Maya, 3dsmax and Blender, I would like to remark that Collada is gaining a momentum even in the academic environment.
MeshLab ( http://meshlab.sourceforge.net/ ) , for example, is a Collada enabled open source tool aimed to the processing of large 3D scanning meshes for Cultural Heritage applications, like documentation, reproduction, restoration planning and didactical uses of 3D technologies.
There is a European Community funded project, called Epoch, that is offering a somewhat similar solution,
i.e. reconstruction of 3d models starting from just a set of uncalibrated, hand-taken photos.
The main target is Cultural Heritage documentation and preservation, so there are not so many bells and whistles...
You can find more info on this 3d reconstruction service and the open tools developed/supported by Epoch for this purpose
Slashdot Mirror
Yes it is true. A personal experience below.
Using real open source projects in the context of a University course can be a quite successful experience. Both from the didactic point of view and from a software production perspective.
Five years ago I decided to involve my students of the advanced computer graphics course at the Pisa University in the collective writing of an extensible mesh processing system that we called MeshLab with a plugin based architecture that allowed a easy to be managed compartmental development. Students get very involved and, beside the computer graphics hard core topics of the course, they learned a lot in term of collaborative development, ethics of sw development, open source licenses. At the end of that course a working system was successfully distributed under a GPL.
Since then, every year, I have repeated this approach extending it, and with the helps of tens and tens of willing students the system evolved into a serious complex mesh processing system, GPL'd, multiplatform, included in ubuntu and that is used worldwide by ten of thousands of users in academic and commercial environments.
I found that the knowledge of participating to the development of a **real** system gives the students a really strong motivation to give their best. The fact that their contribs will be released publicly with their name linked to the commits and listed in the official developer page was a strong incentive to do not cheat. For most of them it was the first time that they were making something real (not only exercises) something with a purpose that was quite different from the standard "get the score" approach. Many of the students continued to maintain their portion of code well beyond the course terms (some even after graduating).
I cannot but thanks my students for the dedication that they have shown in the projects.
Just google for MeshLab for more details of it.
P.
In the context of archiving games, accurate software emulation the whole HW underlying each game is the only solution.
Obviously it should be done in a open, portable, multiplatform way to ensure that it is a long term solution.
Mame and Mess ( mamedev.org ) has already shown that this approach is viable and practical.
I do not completely agree with the above post. While it is true that there are fields where the cost of the sw is not a real issue it is also true that some of these software niches are filled by FOSS tools. For example in 3D printing/rapid prototyping one of the best tools around for cleaning up models is an open source tool, called MeshLab ( http://meshlab.sourceforge.net/ ) .
The concept is not completely new. There is a vast literature on the reconstruction of 3D objects from image sequences, ( http://en.wikipedia.org/wiki/Structure_from_motion ) Beyond the well known Photosynth (that does NOT reconstruct a dense surface) there are other free tools that allow to create complete 3D objects from photo sequences.
Arc3d and MeshLab can provide tools for reconstructing full 3D objects like a href="http://meshlab.sourceforge.net/SnapMeshLab.AO2.png. "> this one.
Disclaimer (I am involved with MeshLab devel)
Another 5cent;
not only the story is a bit old, but there exist free/open source tools that provide the same services, i.e. reconstruction of 3D models from your uncalibrated photos. Look at http://www.arc3d.be/ and to http://meshlab.sourceforge.net/;
The first one (arc3d) is a free-to-use service that allow people to upload photos and get back a set of 3D depth maps well as the relative vantage points from where the uploaded images were taken and the settings of the cameras.
The second one (MeshLab) is an open source tool that allow to load, inspect, process and integrate the data produced with the arc3D service, so that you can obtain a 'clean' 3D model and export it in a variety of common 3D formats.
So if someone wish to try by himself this kind of technologies, the tools are there.
The raised issue is not completely unfounded, but focusing on a more particular sector of the modeling scene the field offer even less choices./ 05/144212 ) while on the software side there are almost no alternative to the high end commercial package in the 50k$ range (20 times the HW cost).
The current panorama of 3D modeling packages is quite tailored towards the needs of the majority of users (low and high poly modelers) and most of these packages offers less than satisfying experiences when used to manage the large unstructured meshes that come from 3D scanning technologies or used for rapid prototyping needs.
The meshes produced by these automatic technologies are typically huge (millions of triangles per object) and not organized into scene graphs; most modeling packages simply sit down on the specs of these objects. 3D scanning and rapid prototyping hardware is becoming more and more popular, devices very low priced are already on the market ( http://hardware.slashdot.org/article.pl?sid=07/04
The only open source alternative on this sector of 3d modeling is Meshlab http://meshlab.sourceforge.net/.
Sorry for what could seem a shameless plug, but I would like to hear your comment on this side of 3D modeling scene.
Beside big players, like Maya, 3dsmax and Blender, I would like to remark that Collada is gaining a momentum even in the academic environment. MeshLab ( http://meshlab.sourceforge.net/ ) , for example, is a Collada enabled open source tool aimed to the processing of large 3D scanning meshes for Cultural Heritage applications, like documentation, reproduction, restoration planning and didactical uses of 3D technologies.
http://www.epoch-net.org/index.php?option=com_cont ent&task=view&id=46&Itemid=88