I think BeerBaron is just trolling all you guys either that or he is bat shit crazy
I think a cleaned up suzanne will be good for glass and sub-surface tests the current one is non-manifold which gives wonky results for those kind of test if I am not mistaken
I am no stranger to rational argument. This is not the place, nor the time. When it comes to art, I demand no less but total dictatorship.
Was it not you who referred to Suzanne as just a shape to test shaders on? Then by your own conviction it cannot be both! Suzanne can be everything. Or nothing. What it can never be is just one thing.
I hope I made myself clear.
Suzanne is the Schrödinger cat ^^
Or the first monkey that had AIDS. Either cure it or leave it alone.
+1 to all quad suzane
I don’t like that the current suzanne’s eyes are not 1 solid shape with the body.
I agree suzanne is not an appropriate primitive to test features that require a manifold mesh.
On the other hand, it is good to test modeling tools that are supposed to work on non-manifold mesh, too.
Suzanne is just a primitive, with the right changes it can become an object capable of really test-driving tools, workflows, and materials. The teapot that comes with Autodesk apps. is already capable of those uses, why can’t.
The Utah Teapot is not specific to Autodesk.
It’s the Blender mascot.
It was an easter for 2.25 by NaN employees.
It became a mascot by its reccurent use in release logs, documentation, tutorials and render tests.
Iterations to Suzanne model by Blender’s community could take signification that as its mascott, Blender really evolves by investment of its community.
But in fact, there is no reason to go against creation of mascott that would take other significations : a Cycles mascott, a sculpt mode mascott, or game engine mascott…
They don’t have to look as Suzanne.
Sorry bro, she only does triangles.
@ pitiwazou, good job, thx!
Push this to dev list, see if it gets into master 
Cedric, je suis mooort de rire sur le débat que tu as involontairement lancé, ça a fait ma journée :")
+1 for the new Suzanne.
Maybe it would be great to also add other test models to try different things
For instance Houdini has useful and nice ones (useful for different situations :water simu, soft bodies…)
http://www.tokeru.com/cgwiki/images/4/49/Emo_pig.gif
C’était juste une proposition, je ne pensais pas que ça allait se passer comme ça.
Great idea to make different primitives.
hi, the all quad Suzanne looks great, no doubt about it, imo it should be an option in the Add Monkey operator, currently Suzanne it great for testing manifold/non manifold, modifiers & many other features, being non manifold it raises exceptions. That said, an option to use the manifold quad version would be useful also.
We already have the teapot model in add mesh extra objects addon, if all else fails, I can put Quad Suzanne into that addon in a prominent place if asked to.
I totally disagree with the policy of adding those kind of primitives “because some user might want to test fluids”.
That’s how software package bloat!
500kb here, why not? 1mb there, sure! And soon you’ll have good’old 100mb Blender compete with big software in terms of useless heavy download size.
If people want an object to test fluids, cloth, hard surface, the internet is full of them. Go blendswap and choose what you need, save it on your HD and take care of it.
I can think of just this alternative solution, if you still go the “official” path: http://www.blendernation.com/2016/04/04/blender-cloud-add-early-preview/
BF might host an official selection of meshes for testing/benchmark purpose, available as a download directly inside Blender.
my 2c
I think Suzanne is intentionally low poly for testing purposes in some situations. they could always have different model states or subdivisions like the teapot model. It doesn’t matter if it is quads, all surfaces are always triangles on the GPU side, that’s how 3D has always been rendered on hardware.
I think this just doesn’t deserve a proper reply.
It’s not a testing object per say but more of an easter egg that was re-purposed for that role. So it’s low poly because that was how ever modeled whipped out at the time.
The lowpoly nature of Suzanne kind of now works against it as a test object especially for ray tracing engines.
It’s so low poly that if you use it directly you are likely to encounter shadow terminator problems. If you sub divide it using catmull-clark subdivision you get poles where are the triangles and because the eyes terminate in a very high valance poles you will get shading errors there(though the are not so noticeable because it’s a flat area).
@ @pitiwazou
you could make it non manifold by adding a few faces inside!
may be add some basic cycles mat like a colored Suzanne
happy cl
Dictionary.com defines art as “the quality, production, expression, or realm, according to aesthetic principles, of what is beautiful, appealing, or of more than ordinary significance.” By definition, Suzanne is art, (aesthetically and appealing.) Notice how the definition says “OR of more than ordinary significance.” Suzanne is not very significant, but she can still be called art, according to the definition.
As for the tris/quads argument, doesn’t dynamic topology fix the tris problem anyway?
For the record, I do appreciate pitiwazou’s Suzanne 2.0 head and I think it should be added to Blender by default or as an addon 
No…
A long time ago I reported certain “errors” in the Suzanne mesh (like being non-manifold) in IRC. Someone (don’t remember whether dev or not) told me that all those errors are there on purpose so developers can test the robustness of new algorithms.
Maybe you can show your new model to the devs and they can tell you which kind of errors/bad practices they need in the mesh so you can integrate them? I’m sure even those requirements change over time.
Here is suzanne made of one single ngon :
Hadrien
Last I checked, the utah teapot model included in max is self intersecting, non manifold, and 3 separate pieces. Just like Suzanne!
