Hi Michael,

I just wanted to say thank you for creating a piece of software that is a delight to use. Moi is light, fast and visually appealing. In my experience usability seems to be neglected for other areas, whether it be budgetary or political reasoning. So I just wanted to say as someone who has designed interfaces, you've done a great job with MoI. After installing it I dived in and within a few hours had my first completed model. Coming from SubD modelling I had to think about my approach to the chosen model, which is the thing I was listening to at the time, my MP3 player. However it didn't take long to get a flow going without reading the manual. I had fun just seeing which tools did what.

I've used a lot of software over the years and what you have created is the office of the 3D world. I know how time consuming developing software can be, expecially with limited resources, so I just wanted say that your efforts are appreciated. I look forward to seeing where you take this gem.

Kind Regards
Charles

Modelled in MoI, of course.


Rendered in Thea Render.

Hi Charles, thanks for the kind words - I'm really glad that you like MoI!

Your model turned out great, it looks like you are off to a great start with MoI!


> Coming from SubD modelling I had to think about my
> approach to the chosen model

Yeah, the overall modeling strategy with NURBS modeling is a lot different than SubD modeling, so that will take a bit of getting used to. But actually these differences are useful because it makes NURBS modeling be a nice companion to SubD modeling since it has a different set of strengths and weaknesses.

Which method you would want to use for a particular project depends on the characteristics of the model. If the model is a man-made manufactured object that is defined by 2D plan drawings, then that's the kind of a thing that fits in really well with NURBS, you can use the 2D curve interaction that NURBS brings to make models like that come together very quickly (and more accurate as well). If your model is much more organic in nature and not really defined very well by 2D profile curves (things like faces, characters, monsters, etc...) then those are the kinds of projects that are a better fit for SubD modeling where you'll be more sort of sculpting the shape by 3D point hull manipulation.

One of the main differences between NURBS and SubD modeling is that Booleans and cutting operations are a primary way of working in NURBS, while they are something that you have to avoid with SubD. So that's a major shift in thinking - in NURBS you usually want to incorporate Boolean operations directly into your modeling approach. This can be a really big shift if you have previously spent a lot of time training yourself to avoid them in SubD.

One area where it is pretty easy to think about Booleans is when you are cutting a hole in an object.

But you can also use Booleans or trimming to carve off pieces of the outside of a shape, not just for making holes - it seems that this aspect of booleans tends to be a harder jump for SubD modelers to think about initially.

Here's an example of this kind of a thing - say you want to make a shape like this:



I've found that a lot of people experienced in SubD will try to initially approach models similar to this by building all the exterior wires of it (trying to define the edgeflow), starting with something like this that they then try to surface:



But with NURBS modeling you don't necessarily want to draw in all the final edges right from the start like that, instead if you recognize that there is a more simple underlying shape that is then cut off, you want to model that underlying shape as a whole more simple piece first, and then draw in some other side profile curves and use them as cutting objects.

Some of the final edges in the model then come from intersections and cuts rather than being drawn in manually.

So the NURBS modeling process for this shape would go more like this sequence:







So that concept of looking for a more simple but extended underlying shape and then cutting that back to get the final model tends to be a major strategy for NURBS that is a bit difficult for SubD modelers to think of at first, particularly when it involves the outside part of a shape - interior holes tend to be easier to recognize as cuts.

Sometimes it helps to think of NURBS modeling as more of a physical manufacturing process where you'll start with some basic pieces of more simple shaped stock, which then have chunks cut out of them (not only holes but also exterior pieces trimmed off) to make the final part.

Hope this helps!

- Michael

Hi Michael,

Your most welcome. It's nothing more than the truth.

Thanks for those tips. I'll bear that in mind. Actually the most interesting thing I found was the placement of holes. I kind of worried about their placement as I normally would, thinking how will this complicate the mesh? What I did early on was export very early versions to the renderer to see what would happen. However the exporter did a very clean job, and I love the fact you can play with the export options.

I then actually restarted the model due a small paradigm shift. I didn't worry about the end mesh so much but rather what were the simplest shapes the the unit consisted of. So having the unit in front of me I rotated it and thought simple, if that makes any sense. The model itself appears very simple, however when you look at the four corners you realise there's quite a bit going on, and there's more complications on the underside. Booleans to the rescue!

I then seemed to go into hyperdrive, and within a short space of time I was looking at a finished model. And to be perfectly honest I almost felt cheated, in the sense that, this felt too easy. It should have taken me a lot longer, because that's the process I'm used to. And I realised I'd just had a great experience. Modelling became fun again, it was like a child playing with silly putty. So I just wanted to thank you for creating an experience that allowed me to have my experience.

Also I don't know how your beta testing works, but if need anyone to test new features before going public, I'd be more than happy to be a guinea pig.

Have a great day.

Cheers
Charles

@Blade
Seems till now every users of a Moi's licence can use the numerous next betas ;)
So no doubt that you will be more than happy to be a guinea pig. :)

Hi Charles, that's great! It sounds like you are getting the hang of it rapidly.

Yeah, the nice thing about NURBS modeling is that you can make a lot of stuff happen quickly from profile curves, either by using just a few curves to create shapes (with various commands like extrude, sweep, etc...) or also using profiles to make modifications like boolean cuts.

So the overall process is kind of more like drawing, you draw some guide curves and make things happen from those.

SubD is kind of more like a sculpting process in comparision, where you're working more directly on little bits of the 3D shape at a time pushing points around in 3D space. This is great for the right kinds of circumstances like if you want to make little subtle lumpy details like around the corner of an eye or things like that, but it involves a lot more wrangling of little points than 2D curve drawing does.

It is pretty good to have various kinds of tools like this at your disposal though, so you can use the one that is most suited for the particular model at hand.

re: Beta testing - betas of the new versions are freely available for existing MoI owners to use, each beta version just expires after a few months. At the moment I'm still in a kind of in-between researching period experimenting with a few things. But it won't be too much longer before I switch into the v3 beta period though.

- Michael

Hey Frenchy,

Thanks for the for the information. It sounds like there will be many happy guinea pigs. :-)

Cheers
Charles

Hi Michael,

You are right. They definately seem to be complimenatry to each other. Though a friend of mine uses Rhino and a product called T-Splines I think, which seems to be somewhere in between from what I can tell.

Thanks for the beta information as well. I look forward to V3. Will it have major innovations or more of a refinement of what's already there?

This may come across as a silly question, but how complex would it be to add say that T-Spline type functionality? Do you know the product I'm referring to?

And lastly, do you take requests while in beta? or do you know where your headed from your roadmap?

Cheers
Charles

Hi Charles,

> Though a friend of mine uses Rhino and a product called
> T-Splines I think, which seems to be somewhere in between
> from what I can tell.

Yup, that's kind of a hybrid or bridge type mechanism - T-Splines works with a SubD control cage but can produce NURBS surfaces instead of only producing polygons for output.


> Will it have major innovations or more of a refinement
> of what's already there?

Well, kind of a refinement I guess - I mean I want to put in quite a lot of new features, but at the same time it is a major goal to try and keep the basic essence of MoI intact as I add new features instead of it getting a bloated and complex feel to it.

So hopefully if all goes will it should get new tools and new capabilities but the overall feel should be more like a refinement and close resemblance to the current version instead of something hugely different.


> This may come across as a silly question, but how complex
> would it be to add say that T-Spline type functionality? Do
> you know the product I'm referring to?

Yup, I'm familiar with T-Splines. So basically T-Spline modeling _is_ SubD modeling, just that it can produce NURBS surfaces as the output. But all the toolset for how you make T-Spline models and stuff like that is all the same as a SubD modeling toolset. And the thing is, the tools that make up a SubD toolset are fairly different than the tools that make up a regular NURBS modeling toolset, like I mentioned before the overall strategies and how you do things are fairly different between these modeling styles. So adding in "T-Spline" support into MoI is not about adding in just 1 thing, it would be about adding in an entire different toolset and modeling approach containing a whole variety of different tools with it.

So that's quite a large amount of work to implement an entirely different modeling paradigm within the same software especially without adding in UI bloat and complexity. It may be something that's possible in the future at some point, but there is still quite a bit of stuff in the regular NURBS world that I want to add in still, and I don't really want to go off in some completely different direction too early before MoI's current toolset gets further refined.

> And lastly, do you take requests while in beta?

Yes, definitely! In fact the beta is generally guided by requests and feedback and not so much by any exact predefined plan.


> or do you know where your headed from your roadmap?

Well, I do have quite a few ideas that are shaped by all sorts of previous requests and feedback.

Of course combined with feedback I also have things like feasibility and amount of work required for particular things, so for example if some requests would require several years of full time work for me to implement, it will probably will not be able to happen anytime too soon.

Another thing that tends to happen is requests that would require significant new UI design tend to get done more slowly (adding in T-Splines for example would fit in this category), and on the reverse requests that can fit in well and add functionality without adding a lot of UI tend to get the highest priority.

- Michael

Hi Michael,

Wow, thanks for your honesty, it's quite refreshing.

And thanks for the explanation regarding T-Splines. I didn't realise it was SubD that produced nurbs, I thought it was a nurbs cage of some sort. Knowing that, I would agree with you that adding this feature would be confusing, especially considering what MoI is, and like you say it would be a huge effort to incorporate. I like the idea of maintaining MoI's current philosophy. What you have said makes total sense, and is logical.

I do have one question though, and btw let me know if I'm getting annoying :-)

When you have a solid, say a sphere, you can show points. However when you boolean the object with another, points are no longer available. I know there a reason for this, however would it be possible to show points on a booleaned object? See image below.


Show points.



Cheers
Charles

Hi Charles, it is possible to show points for a booleaned object, but when you do so you'll probably see that the points are not as you are expecting.

To show points on your booleaned sphere, use Edit > Separate on it first to break it apart into individual surfaces, then show points will work on those individual surfaces.

When you do that, you'll see that the surfaces involved there are actually larger than the edges - that's due to the structure of how NURBS solids work, where when things are trimmed or intersected, the underlying surface stays the same and only new trim curves are created. Trim curves are kind of like markers that live on a surface and mark different portions of the surface as being active or inactive, making holes in it.

But the shape of the surface (and therefore what you would alter by pulling points around) comes from the underlying untrimmed surface and not the trim curves directly, except in the special case where the trim curves are the same as the full outer boundary of the surface.

MoI will not turn on control points when you've got joined pieces that do not have surface control points that touch each other because it would be too easy to create gaps and holes in the model since pulling the points around would just pull one part of a surface around and not pull the edges more directly.

This all is another big difference between NURBS and polygon modeling - in poly modeling the edges that you see are always the exact same thing as the edges of the surfaces involved, so pulling edges and pulling the surface are equivalent. With NURBS where you have a trim edge, that's a kind of separate thing from the surface and there is not the same concept of pulling around the edge directly.

However, this entire structure is also the reason why boolean operations work better with NURBS surfaces than they do with polys - with polygons when you do booleans the boolean code has to create a lot of little slivered polygons, and the more booleans you do the more dicing and chaotic little fragmented pieces are generated.

With NURBS surfaces on the other hand, when you do booleans and trimming the underlying surface itself stays the same and only new trim curves are calculated that mark areas of that surface as being holes. No matter how many cuts you do on a surface it does not fragment into a bazillion little shards like poly booleans do. And you can also do stuff like erase the trim curves to get back to the underlying surface.

See this FAQ answer here: Why does show points work for some objects but not others? for some more information and some illustrations about this.

- Michael

Hi Michael,

Thanks for the education. I've had a play with separating solids and re-joining them again after manipulation, and I've managed to came up with something quite complex that is a solid at the end. It's given me another methodoloy, which is always cool. More bows for my arrows so to speak.

Thank you sir, I will continue to play.

Cheers
Charles