Michael:

Thanks for your thorough reply. Regarding the masking of trimmed surfaces, I momentarily forgot about that (um, very fundamental) aspect of NURBS modeling and how it might impact the implementation of enhanced sub-object control over larger objects. Also, I can't recall if the Maya deformers operate on NURBS geometry-- though I think they must--or only poly-meshes and subdivision surfaces...

Anyway, I appreciate what you suggest about sketching as "drawing" vs. sketching as "sculpting"... But I think that sculpting in a 3D-modeling environment is hard to avoid, especially when dealing with "freeform" and "organic" shapes. :^) The funny thing about Sketchup is that while it threads the drawing metaphor completely through the interface, the "Push/Pull" command combined with the ability to dynamically subdivide/add detail to surfaces leads to a way of modeling that is not drawing-like at all!

Now that I've mentioned Sketchup, I will say that I love that one can make a "component" which has the ability to recognize and reorient to faces as well as boolean automatically when placed. If something like that ever happened in a complex surface modeling environment like Moi...well, now that would be pretty jaw-dropping! (I can only imagine how difficult that would be to implement however...)

On a much simpler note, I also feel that it might be nice if the "Esc" key (which I have programmed as one of my tablet buttons) cleared any current selections much like it cancels out of a command. (Now that I'm almost certain is one line of code that can go in the *.ini file...?!)

Anyway, thanks for creating such a nice app; I'm enjoying exploring the software and particularly like the Shell tool.

Best, Justin

> Also, I can't recall if the Maya deformers operate on NURBS geometry--
> though I think they must--or only poly-meshes and subdivision surfaces...

Well, I'd imagine that they work on NURBS geometry, but only on surface control points, not on an edge, especially a trimmed edge shared between 2 adjacent surfaces.


> But I think that sculpting in a 3D-modeling environment is hard to avoid,
> especially when dealing with "freeform" and "organic" shapes. :^)

There is a certain "flavor" or level of freeform that can be achieved just through drawing and not sculpting - for instance you can create a swoopy "freeform" surface using sweeping by drawing rails and cross-section curves.

This is for a more man-made stylized manner of freeform. Completely organic like a monster's head is a different type of thing - that's when you have to use sculpting. The difference is if you have a type of "high-frequency" detail required.

Man-made stylized freeform surfaces don't tend to have really tiny bumps and wrinkles in them. Lots of tiny bumps and wrinkles means good for sculpting - smoother flowing freeform means good for drawing + construction (MoI) style.

So what I'm getting at is it really depends on the type of models that you're trying to build.


The thing is that there is a lot of new stuff out there right now focused on sculpting - I mean stuff like Silo, Modo, Cinema4D, XSI, ZBrush, Mudbox, ... But there isn't much that is focusing on structured drawing-style modeling and making that easier and quicker. So that's kind of a reason for that different focus for MoI. I just figure that if you want to do something organic with high detail like a monster head, you should just be using one of those other apps for that and not MoI.

Sketchup is definitely an interesting kind of hybrid between drawing and sculpting. They've got it tuned really well for architectural concept modeling, but it is difficult to do other types of stuff with it.


Anyway, I hope you don't think that I'm attacking your ideas or anything, I'm just trying to give a bit of an overview on some of the thinking behind MoI's design.

I'm glad that you like it! I think that one of the tricks to using it is to know when to use it for a particular kind of model that it is suited for, and when to use a different package for stuff that it is not suited for.

Re: Esc key, yeah I think that makes sense, I have been thinking of making it do that. Right now you can't program it because it is a special case key but I'll try to get this in for the next beta.

Thanks!

- Michael

>Anyway, I hope you don't think that I'm attacking your ideas or anything, I'm just trying to give a bit of an overview on some of the thinking behind MoI's design.

No, not at all! Just to explain, I'm coming from an architectural context. I have many years of "real-world" design experience, recently returned to graduate school and now teach part-time at the graduate level. Architecture school is ground zero for software experimentation, and students are quite catholic regarding the tools they use.

In recent years, the affordablity, precision (and, recently, the scripting abilities) of Rhino has firmly ensconced NURBS modeling as a default mode for architecture students everywhere. Professionally, you can see NURBS working overtime in the projects of architects like Zaha Hadid and Ali Rahim of Contemporary Architecture Practice.

>>There is a certain "flavor" or level of freeform that can be achieved just through drawing and not sculpting - for instance you can create a swoopy "freeform" surface using sweeping by drawing rails and cross-section curves.

This aptly describes much of Zaha's current work. You look at the renderings and can easily imagine all the profile curves along which surfaces have been created. This technique seems perfectly suited for her work since it is very very much about creating a sense of momentum through mulitple implied vectors. (Even the overall graphic of her website consists of a series of overlapping, filleted curves moving in various directions...)

>>So what I'm getting at is it really depends on the type of models that you're trying to build.

True. But, I've recently I've witnessed a slight shift away from NURBS modelling back to poly-modeling in the form of Sub-division surfaces (in Maya, Modo and a bit in 3ds Max). Sub-D models are often in the end indistinguishable from NURBS models though the workflow can definitely be more "sculptural" as you have noted. For product design, of course NURBS offer a precision for design profiles that you will not find in poly-based modeling.

>>So what I'm getting at is it really depends on the type of models that you're trying to build.

I also think this question can be turned around to read: "What types of models does the application afford you the possibility of making? [and] What new and unexpected possibilities for design does an application open up?"

With Moi, it currently seems that speed and intuitive informality with which one can work with NURBS geometry may perhaps lead to increased iterations of design study. If you can model complex geometry more painlessly, then you can simply draw more of it to analyze and evaluate from whatever design criteria you are operating on.

However, as a designer, I am always selfishly looking new tools which open up those fundamentally new opportunities for design exploration--if only in very subtle ways. Thus my suggestions will probably be slanted towards the creation of tools that don't yet seem to be available in any package...

But based on its current incarnation, I'm certain that whatever Moi evolves into will be cool and fun to use. The interface just plain rocks!

> True. But, I've recently I've witnessed a slight shift away from NURBS modelling
> back to poly-modeling in the form of Sub-division surfaces (in Maya, Modo and
> a bit in 3ds Max).

Yeah, I think this has been a growing trend for several reasons. For younger people, Sub-D kind of has a greater "cool" factor since it works well for creating monsters and creatures. There's also been more of a general wider surge of innovation and new products focused on sub-d methods.

Newer products can tend to have a more modern user interface and can attract more users because of that.

There hasn't really been any accessible new NURBS modelers come out during the past few years, well except for MoI of course! :)

So I think that those are some of the reasons for this type of a shift. But I'm also seeing a lot of people now sort of "rediscovering" what NURBS modeling is useful for with the introduction of MoI and its focus on ease of use.


> Sub-D models are often in the end indistinguishable from NURBS models though
> the workflow can definitely be more "sculptural" as you have noted.

In certain senses the Sub-D workflow can actually require a higher learning curve, I think it tends to require a somewhat higher degree of spatial visualization since you are often manipulating so many different individual points in 3D space.


> With Moi, it currently seems that speed and intuitive informality with which one
> can work with NURBS geometry may perhaps lead to increased iterations of
> design study.

Yeah, especially for V1 the focus is on speed and fluidity - I like to say making it simple to do simple things.


> If you can model complex geometry more painlessly, then you
> can simply draw more of it to analyze and evaluate from whatever design
> criteria you are operating on.

Well, the first step for MoI is the focus on more simple geometry right now. Complexity can still be achieved in a sort of different sense - by a combination of more simple individual parts. NURBS tends to lend itself to this kind of combinatory approach since it has good support for boolean and trimming type operations.


Eventually there will be a trend in NURBS modeling that will combine some of the workflow of Sub-D modeling, but producing a patchwork of NURBS surfaces from the control cage, rather than a polygon output surface. You can see the beginnings of this trend in Catia's new Industrial Design Sub-d modeler, and the T-Splines stuff. I think that 4 or 5 years from now this will probably be more mainstream and you'll see some more cross-over feature sets.

The only thing I'm not so sure about is how feasible it is to combine both sets of tools (sculptural and drawing) in one application and still maintain a really simple and approachable UI. For the time being I'm intending to solve this problem by focusing more on the drawing + construction type workflow rather than sculptural/squishing type.

- Michael

>>For younger people, Sub-D kind of has a greater "cool" factor since it works well for creating monsters and creatures. There's also been more of a general wider surge of innovation and new products focused on sub-d methods.

Ha, I wish Moi and Modo had been around when I was a teenager! Or, maybe not, I probably would have dropped out of high school modeling god knows what and never ventured into architecture. There are a only few architects who explicitly admit to wanting to make monsters (and they all use Sub-Ds)...

Anyway, thanks for your many insights.

Best, Justin

> There are a only few architects who explicitly admit to wanting to
> make monsters (and they all use Sub-Ds)...

:) Yeah sorry I didn't mean to imply that Sub-Ds are only useful for monster creating and not other stuff.

The monster-creating (or more accurately I guess "character" creating) is just one of the general factors behind the surge of Sub-D popularity. It's a big focus for certain high-visibility industries such as hollywood movies and computer games.

- Michael