Hi Jason, yes I'd like to make chaining work in the future.

Right now blending in MoI is limited to only function between a pair of edges, so you can't select a chain of them, only 2 at a time.

- Michael

I hope this illustrates, how you may end up with an issue. I'm curious how to be sure we will maintain continuities flowing from 2 directions.

Hi Jason - maintaining continuities in multiple directions tends to be used for a kind of advanced NURBS modeling workflow.

MoI is more focused on kind of more core, easier to use broader construction functions currently.

For example for the shape you show there rather than building it in a kind of individual patch-by-patch method, I would probably build it as a single larger rail revolve or sweep, for example with these 2 curves:



I then used Rail revolve to do a surface of revolution that is squished by that vertical rail:




There tend to be many advantages to defining smooth pieces as larger surfaces like this - there are fewer curves that are being used, that tends to make it easier both to set things up and to fine tune things for instance by adjusting curve control points and having the surface update.

And as you can see, when you make things out of broader surfaces it is easier for the surface to be fully continuous.

Even when you have multi-direction continuity tools available to you as in other NURBS modelers, it doesn't necessarily guarantee you a pleasant shape through the entire breadth of a surface, they are more about maintaining continuity right at the edge area. When you do that patch by patch approach it tends to make for more undulations and undesirable shaping happening even though technically there is continuity enforced at all edges.

- Michael

I appreciate this Michael. I wasn't aiming so much for efficiency perse but rather looking for blending cases. The shape I did wasn't by design but just doodling and looking for ways to break MoI.

Hi Jason - I guess for quick doodles, I'd generally say just don't worry about continuity at all for that kind of very quick model construction...

In that kind of a working mode, you may want to just stick things together however possible just to get some shaded thing on the screen. Then for a higher quality construction with continuity, you could then sit back and think about a more efficient way to build things in larger sheets to get better smoothness.

But also building in larger pieces can tend to make doodling go faster in many cases though too, because you get more stuff happening for each curve you draw. Like that previous example of just drawing 2 curves to make that form rather than a bunch of pieces.


If you do a lot of "patch by patch" things, you're basically kind of leaving the realm of the strongest/quickest areas of NURBS modeling where things are defined more by a smaller number of profiles, and you're moving more into an area that is more sub-d friendly where you're trying to manipulate more localized areas of a smooth final result.

- Michael

I can joke and say I want T-splines for MoI?
T-splines for rhino are so buggy (v2.0) when you work in symettry and delete a face it hang'n'crash.
And when you start to have a complex model (like a coca cola bottle) it gets very slow even on a quad with 8gb of ram.
Also I think that Rhino is much slower than MoI (loading the exact same 3dm and moving the camera around, MoI is faster)

Hi DesuDeus, that's definitely something I want to explore in the future!

But it will be a pretty major undertaking, so it will likely be a while before I can explore it.

- Michael

That would be wonderful, Michael!

-Will

"...I can joke and say I want T-splines for MoI? T-splines for rhino are so buggy (v2.0) when you work in symettry and delete a face it hang'n'crash. And when you start to have a complex model (like a coca cola bottle) it gets very slow even on a quad with 8gb of ram... "

If you find Tsplines to be buggy, it probably has very little to do with Rhino. So I don't think having a MoI plugin will necessarily fix that. But the Tsplines V2 code seems MUCH improved and more stable than V1. If you have this repeatable crash using symmetry, just turn symmetry off and delete the other half as you work. You can always add symmetry later... Also I feel the best workflow is to rough out your model in a poly/subd modeler and then import as Tspline to add local refinements...

talking about TSplines...something from SIGGRAPH 2009 Technical Papers

http://www.siggraph.org/s2009/sessions/technical_papers/details/?type=papers&id=37

http://www.cl.cam.ac.uk/~tc270/subdnurbs.html

Yeah T-splines are awesome, I mean its just the logical continuity of nurbs. Ive played a lot with the Rhino plugin, Its buggy but you can do so complex shapes that are just impossible to do with patches of nurbs! Also from rhino you save your t-splines into 3dm and MoI open it, then you do booleans and fillets between two T-Splines volumes. Seemless integration. But it cost a lot for a plugin (559 if im right) + the cost of Rhino, and while Rhino is rock solid, the plugin is very buggy and crash a lot.
The other sad part is that very few software have them. At least cheap software (less than 2k)
MoI with T-Splines would just nuke any competition :D

And it will gain a hell lot of customers around the world :)

DesuDeus wrote: "The other sad part is that very few software have them. At least cheap software (less than 2k)
MoI with T-Splines would just nuke any competition :D And it will gain a hell lot of customers around the world :)"

TSplines might sell more licenses if a host app were available for $300 instead of $800. But Rhino has a lot more exposure, so it probably balances in the end. Of course any developer would be interested if a plugin could bring new customers to their app. But as Michael mentioned, he's not ready to support plugins yet. Also there are many Rhino functions which TSplines is using behind the scenes. MoI would need to have equivalent functions, otherwise Tsplines has to write their own code for it. They do this sometimes, but I'm sure they rather avoid it. Also MoI would need support for mesh objects. T-Splines uses Rhino's mesh object, but adds an extension to support T-points and creasing. The last (and probably most important) is that TSplines is not a simple plugin such as a render engine. I've been told that Tsplines is doing very sophisticated coding at a very low level in Rhino. Consider how they are able to take over Rhino's interface. McNeel had to develop this to support them. Probably MoI would need to do the same. I think it's a big piece of work and would slow down other development for MoI...

Also my feeling is they want to develop a very solid plugin to Rhino before porting to other applications. It would help them a lot to isolate problems if they only focus on Rhino. Also (this just my opinion) i think TSplines achieved only about 50% of their potential so far. Once they have the ability to support trimming, fillets, etc without collapsing to Nurbs then i think there will be a great interest among the big CAD developers. Also they need to improve the continuity around star points. V2 made a BIG improvement to have a minimum of C1. But for many companies, I'm sure they would need C2 at all points. TSplines is a young technology, and I think they will start expanding in a couple of years...

@ Jason - Sorry for derailing your thread! :)

jonah

Yeah continuity is always hard to have everywhere.
But if you do product visualisation T-Splines are perfect :D

Quoted from T-Slines forum.

Question: T-Splines For SolidWorks

Answer:(TomFinnigan T-Splines developer) I probably shouldn't go into too much detail about our discussions with other companies, but SolidWorks is certainly a company we've spoken with about T-Splines.

Currently the bulk of our development efforts are going towards improving the Rhino plugin and planning for new features. We'll let you know when they're available, but the big features will be exciting for any application that includes T-Splines. Rhino is a good app to develop new features in, because of the simple SDK and McNeel's excellent developer support.

Another thing we're spending time on is improving the documentation and (free) learning materials, which I'm sure you understand the importance of.. :)

We'll be sure to post news here as soon as there is a commitment to include T-Splines in SolidWorks or any other app. Unfortunately, for now I could only speculate what will happen.


ps. Bob McNeel is one of the T-Splines Board of Directors.

I know all that neo.
Im just dreaming about using T-Splines inside MoI :)

I think it would be very interesting to see how Solidworks would implement something like TSplines. One would normally think of Tsplines as free-form. But i think the workflow could really benefit from the use of parametrics. Imagine picking any 2 points on a tspline and then using a dimension to drive it. I'm currently learning Solidworks, and not really a fan of the workflow. But i think with Tsplines it could actually be very cool...