NURBS question

Last semester I took an Alias I course, had a great teacher with strict rules. He forbits us from doing the following for his class:

-Use only the tools he specified for that particular assignment
-No pulling of surface CVs, only change curves
-Keep all curves
and more rules.

He has reasons for those rules, and he explains them. I was convinced.

This semester, I am taking Alias II, and took a teacher that my first teacher recommended.

First assignment, we pull surface CVs. When he was doing demos, he throws curves as soon as he’s done making that surface. I asked him if he ever keeps his curves, he said he used to keep them when he was a student.

Ok, that’s not what I am concerned with. I am more concerned with the pulling of surface CVs. The reason given by my first teacher is, for production surfacing, if you pull the surface CVs, no one else will be able to replicate what you have done. So you can really screw the project up by providing insufficient info to the next person who will work on your model.
I don’t know what my second teacher does when he’s at work, maybe he’s just teaching us easier ways to get to things.

So this thread is really more about the proper procedure of production surfacing. I was in a short argument with the proctor( the person who tutors Alias for our class) about pulling CVs, and his argument is if someone else needs to make changes to your surface, then what’s the point of the curves?

To me, that’s a very unreasonable argument. There are thousand and one reasons why someone needs to recreate a certain surface. It’s not just because it’s bad. So I am still going to follow the rules I agree to and make sure that whoever looks at my model knows how I got there.

What’s your take on this?

I agree with your first instructor regarding curves, but if you need to pull Cvs can you do so on a copy of the curve driven surface, so that you have both designs available?

rules come about because people need to agree how to behave in a way that helps everyone. or because breaking a rule will crash the computer, as in, “never try importing IGES data into alias without saving your model first.”

i see two big issues here: technical and social. technically, there’s no reason not to pull cv’s to acheive a solution (unless you’re going to a cad package that totally barfs on anything nongeometric, in which case you’re using the wrong cad package or you’re modeling something that should be modeled in a cad package). in fact, pulling cv’s is how cars are made, at least exterior sheetmetal, even in alias. surfacestudio and icem are full of tools just for sculpting points into satisfactory forms.

as for the social issue of less competent users not being able to handle my models, that’s crazy. if someone else needs to reshape something in my model, they can find a way, whether there are curves there or not, even if i sculpted a form by pulling points. everybody chooses different solutions to modeling issues, let them use the tools they know to get the job done fast.

in production, the name of the game is getting either a rendering or tooling data. that’s it. nobody gives you an extra three days to rename all your nodes, group everything properly, clean out the leftover construction curves and surfaces, etc. usually pieces are changing daily, or hourly; there isn’t always time to do things “the proper way”. what matters is that your model has the right aesthetics, is toolable, and passes cleanly downstream. period.

having said that, i hate when someone trims with shrink surface, or throws away all the construction curves, or wipes all the history. that makes my job harder when i have to redo someone’s work. but that’s different - i didn’t tell them not to use a certain modeling method, i just want as much of the construction aids there as i can find so i can reuse them as needed.

learning is a different ballgame too. using only the tools a teacher specifies will help you learn that method. this doesn’t mean you can’t learn other methods too. do it their way, then do it another way, then another. find out which is fastest, easiest, simplest, most accurate, etc. then you can solve more modeling problems faster than knowing only one set of rules to follow, stumbling when a problem doesn’t fit your rules and their assumptions.

I agree with both of you. I agrees with pulling surface CVs, but duplicate that specific isopam and re-rail the surface with the newly added curves, so technically, you are still providing enough info to let the next person know what you have done.

What I disagree with is throwing away the curves when you can have them kept, and pulling surface CVs without leaving something behind to indicate the change. Nobody says you have to, but it will make someone else’s life a bit easier.

i dont care about original curves for rebuild in CAD. surface data all i need. curves only help me understand how the form started. doesnt mean i’ll rebuild it same way.

In a digital world, to throw out data that can be hidden on a separate layer shows a lack of wisdom. Another sad effect of Ph.Ds who have not proven they understand philosophy.

In fact, not only do I save everything, I save every revision of a model. At a much later time, I can revisit the development history and the “thinking” involved in the project. The files themselves can be proof of timelines for patent issues. And it’s a great way to explain why project billing is so high.

Re: moving points, if the form requires it, do it. NURBS allows this and you often have to mesh before going to RP or mfg, so the math is lost in translation. Besides, you can always export the surface curves and let the downline application deal with them directly.

That said, I prefer to use simpler functions to get the form, if possible. If I can sweep a profile instead of lofting something I will, because it is easier for engineering to machine with rotary tools. The final mfg method should also influence the design. If not, your part costs are higher than they need to be.


I completely agree with orange, the fact is that I also keep everything, which can get a bit messy, but always useful when you have to step back, or when you suddenly lose a model you’ve been working on. (these things happen sometimes).
The thing is I keep the curves, even for myself, In case I need them eventually to redesign, or build over them.
I’m not a friend of pulling surface CVs anyway, It feels too random (I haven’t practised that either, must admit).


“So you can really screw the project up by providing insufficient info to the next person who will work on your model.”

in my experience the “next person” is often in different package. especially if working in Alias. personally i keep curves. like ppd suggests, put on layer and hide. even if history broke. might need later. but if it goes to another app, thats different. and more common.

first off there’ll probly be no layers on import of your file. too many curves can be confusing. misleading. and cause problems. not every IDer has nice clean construction. most have multiple curves and junk stashed.

second, after moving CV’s the surface isnt related to curve anymore. do you really want the engineer to depend on curve no longer used by a surface? i wouldnt.

third, if rebuilding surfaces in CAD (sometimes - oftentimes - necessary) those curves may not produce same surface as original. even if there is no CV change. iirc different apps use their own build equations. can use imported curves, but surface may turn out different. maybe only slightly. but enough. which is more important? using curve and getting different surface? or matching the surface?

fourth - and important imo - alot of IDers dont make very good surfaces. i’ve rebuilt surfaces for IDers. been given direction “make exactly like”. i’ll include their mistakes. once had an IDer do this. looked at my Pro and complained of errors in surfacing. when he turned his back to talk someone else, i turned my object layer off and turned his import on. zoomed in on area. he turned back and continued rant. until i told him it was his file. i then plotted his file and asked him to mark for correction. to avoid a repeat. he made one mark. in disbelief i went back to work. you can guess what happened next.

“So this thread is really more about the proper procedure of production surfacing.”

the “proper procedure” is to communicate what you want. amazing how often that doesnt happen. and sometimes you have to trust person receiving the file is competent. no man is an island and all that…

going straight to the main issue, proper procedure for production surface.

using catia you have multiple options. but the best method for me is always to create a skeleton> then solids> surfaces> solids… these are seperate files that you build your geometry on preferably in an assembly file.

most of the curves i use are either 3d splines, composite or projected. i rarely use curves on a plane, but i do use them for projection.

once the curves are optimized for a loft or sweep (this process is rather complicated, but if you don’t know how to optimize your curves you better ask your teacher, and if he doesn’t know you’re in trouble!!) then i create the surface. you can create different files and try different methods. it might take you longer but you’ll always have the surface you like.

there’re tools in catia that can give you surface tangency and curvature. that way you can rework it or balance it by ironing out the surface.

the hard part is usually tangencies in aca but if your initial geometry is well processed it shouldn’t be a problem.

control points are used to refine the surface. surface geometry is something you should consider when using them. how you use them varies from one surface to another. in cad you can’t think art. that’s the difference between surfacing in cad and surfacing in other type of 3dsoftware using mesh/nurbs.

design intent is really important.