Why are Class-A surfaces such a big deal for cars, but not industrial design / architecture?

Yes Icem. They call it „Patches“, which allows you to combine multiple Bezier or Nurbs patches/surfaces into one single Bezier surface.

1 Like

kids these days, smh …



Wow, I have never heard about that command before! ConvertToBeziers :slight_smile: Is it one of those super useful, but also super hidden test commands that silently exist in Rhino?

I just tried the command with a single span model, yet it split it into a bunch of patches, even though it was perfectly smooth sort of Class-A surface anyway… Interesting…

photo removed by admin- keep it clean folks, cmon…

@ec2638, this made me laugh out loud. We did catch this before the tooling was cut, but it’s exactly what you describe. Design File vs. tooling maker modifications.


’It’ almost 'happened’ to you! @Mark_Landsaat:wink:

And, if you’re doing beautifully surfaced carbon bike frames (or whatever), this kinda crap will drive one nuts. (and potentially hurt sales)

My recollection is ConvertToBezier has been a regular Rhino command since at least V4. It’s not one which gets a lot of use in regular modeling.

Perhaps there is some confusion about the term “single span”. A single span surface has the number of control points equal to one more than the degree in each direction. In mathematical terms there are no internal knots. The object in you illustration does not appear to be a single span surface as it has 24 control points in one direction.unless it is degree 23 in that direction. :slightly_smiling_face: It does appear to be single span in the other direction with 4 control points in that direction.

I have no idea who had the final decision on the Dual shock 4 controller. :slight_smile: But it’s a failure from an ergonomics standpoint, because many people complained about that large radius at the bottom where the middle finger is supposed to be.

Another notorious example of a badly designed modern game controller by a multi-billion corporation was “The Duke” controller for the original Xbox game console. It was so huge that most gamers couldn’t reach comfortably some of the face buttons and the left stick. :smiley: Due to the inability of many kids to play with “The Duke” controller, sales of the Xbox were very slow, many consoles were eventually returned, and “Microsoft” was forced to release a smaller version named Controller S. “The Duke” was designed by a big man with big hands called Seamus Blackley, and was tested by a few (also big) men who were hardware testers for “Microsoft” at the time. It’s one of those cases where the target audience of a consumer product (primarily kids and teenagers) was totally ignored by the engineers of the company.


Haha, yes you are correct, this is an example of a forged aluminum dropout, but we definitely had a WTF moment when they send the files back to us for review.

I can’t share the images, but this has happened and gone to production on some carbon frames as well, and it was exactly what you described. Tooling maker opens IGES/STEP file and cuts out what they deem is “wrong” and patch it together with their MCAD program of choice.

Yup, often out of some misguided thought of easier or faster (for them). Catch and push them, and they ‘find’ a way to do it ‘your way’…

Or they tell you how much more it will cost. :worried:

Looks like it has 22 spans in the V direction… Originally, it was a cylinder that I converted to degree 3 in both directions, but with much higher amount of control points to be able to manipulate the shape afterwards. It’s perfectly smooth and easy to edit at any time, yet it may not fall under the definition of a Class-A surface quality.

The cylinder became a multispan single surface when converted to degree three in each direction with increased number of control points. The What command can be used to check on the number of control points with the number of spans = number of control points - degree.

When the multisurface was converted to Beziers the result should be 22 individual single span surfaces, each with 4 x 4 control points.

Same story here. Sent in a nice, blended transition CAD model, came back replaced with a radius fillet (AND had the audacity to claim “we didn’t change anything” until we sent back proof).

Outside of the automotive industry, some people really don’t get Class-A. :wink:

Not far from the truth imho…

Outside of the design 3D modelling nutshell most people can’t tell the difference between great and perfect… and THEREFORE don’t care either. It’s like the old graphic design joke “If you hate somebody, teach them about kerning…” (It means that once you know how to look for errors you will never stop seeing them, thus ruining your pleasure of enjoying great design for life) I have many times fine tuned design until I am finally satisifed and upon presenting it nobody can tell the difference, not even side by side… while it to me it was a cruzial stop-the-press moment to fix that “sore thumb sticking out”… :wink:

We all are like that though, we see what’s on our radar and are blind to the fidelity of that we have little knowledge about. if you aren’t a trained musician or a devoted hi-fi dude then the highest level of fidelity doesn’t matter. BUT it matters for the creators and thus inspire, gives staisfaction and momentum in the design process resulting in an overall better design. So keeping designers satisfied is important.(that is satisfied enough, but don’t overdo it).

AND it’s important for design critics, automotive sellers, reviewers etc. If they find a flaw on a bad day then it can affect the way they talk about the product for a long time and thus ruin sales badly.


I would only partially agree. Any unskilled person will notice something weird going on, even if he/she is not able to tell you why. They will say its bad design, but maybe its just not well made or its broken. A trained eye will catch that. On the other hand a trained eye will see any little error another person might don’t care about…

But especially on cars, just think of the impact a little fender bender has on the cars appearance and its value.

1 Like

I am sure we agree 100%, it’s just a matter of point of view for the discussions sake.

(But note that I sad between ‘great’ and ‘perfect’. Great is good enough per def, perfect is flawless)

I dare say that on 90% of cars there are so many small and bad design desicions made where top notch class-A surfacing won’t save them, but on AAA cars on the other hand all details matters. So putting a Porche in the same category as a ‘transporter’ and applying same rules to all cars obviously doesn’t make sense though.

@Holo @TomTom

There is this funny thing about perfection:
Only few people are capable of recognizing its presence without the help of others.
But many more realize its absence immediately without being able to point out the nature of what is missing.


I was a perfectionist with “class A” surfacing, obsessing over surface continuity and those G2/G3 transitions. EMap was my best friend. Now that I run my own CNC mill, those G2-G3 continuity are the worst stuff to machine perfectly. Now I use arcs and lines exclusively.

In Class A, G2/G3 is not mandatory. Many smaller fillets have only G1 transitions even on production models. (Technical between G1 and G2. There is no explicit G2 and and G1. Its always a range. We only consider G1 under 0°05 degrees as Tangent, but nothing like this for higher continuity)

Ah-ha, it’s your fault, Snecx… :wink:

Explains why you ‘fellows’ are always looking to zap my, and @Mark_Landsaat ‘fancy-pants’ fillets!

If I had a nickle for every time I heard some derivative of - “fancy…”… :sweat_smile: