Saturday, November 22, 2014

King of the Rocketmen

Finally got back to the CAD software:


This is a "flash hider" that sticks on a stock Luger pistol to make a "King of the Rocketmen" prop. When I've finished the model, it will be freely available to print at my Shapeways store...and I'll look at the numbers for lathing one out of aluminium.

And I really don't understand Fusion 360.

Okay, first off, I'm learning CAD. That's an expected learning curve. CAD is a different way of doing things, and I also have been emphasizing organic and poly modeling techniques in the past: CAD leans much more towards parametric methods.

But I'm having a lot of trouble understanding how Fusion 360 is organized. And I'm not alone in this. It seems to have a collection of semi-hierarchically related structures, among them bodies, patches, components, and parts. How they relate, which can contain which, which can be converted into which, is quite unclear. Furthermore, the possible operations change greatly depending on which one you are looking at; even if it is otherwise identical looking and created in an almost identical method.

The software is being rapidly revised, including even the names of parts and functions, and there exists nothing even slightly like a manual. There are videos, but they use terms and show a workspace which is several versions superseded and in many cases no longer applies.

Above that is a bigger question of; "What they heck is this software for?" I have my own suspicions. I think that Autodesk had a bunch of maverick programmers and marketing people; beanie-wearing, espresso-drinking hipsters who were a little too out there for the flagship products. So they put them in a room and told them to go wild with a sort of mad mix of trial balloons, concepts in search of an application, some solid code, and above all an urge to look as trendy as humanly possible.

This means not just the team, but most of the active user base, are people who have way too much experience with Inventor and so forth. Making the software accessible to outsiders doesn't seem like a priority, because they are working in a self-exclusionary echo chamber.

It is fairly telling that the top threads in their own forum are about how to include videos and personalized graphics in one's forum posts.

In a refreshing alternative to most 3d software, what they seem to think of as their selling points are not the modeling functionality, and certainly not gosh-wow render tricks or a suite of included-with-the-download DAZ dollies, but instead the concept of sharing and portability. They believe strongly in this idea, although the implementation seems nebulous (really, now; software that is buggy on a powerful desktop machine, with a graphics-heavy interface, is not going to translate well to a smart phone. And the cloud storage system they have already for file management is crawling-slow, heavy and unwieldy even on a decent work-based connection. Good luck shoving those files down the pipe you will get at a beach-house!)

As far as I can tell, in fact, all the vaunted "sharing via cloud" could be achieved just as well by throwing a save file at DropBox every now and then. But I could be missing the point. Right now whatever it is, it confuses the existing user base more than it accomplishes anything else (lots and lots of forum threads wondering how they can access their own files after a quit and restore!)



In any case, I struggled for days just to lathe the simple shape above. Practically every standard method I tried ended up with unselectable edges or grayed-out "OK" buttons, with no real explanation why. At least one of them crashed the application completely. And what I finally got, violates the spirit of parametric modeling completely in that it is not easily editable. Half the useful functions require turning history off, and the hierarchy of parts doesn't actually seem to go into a final form that includes sub-forms that can be edited.

Navigation is still broken, although I'm using third-party tools to remap and create a middle mouse button, which should help. (For me, 3d aps live or die on navigation. If I can't zip around an object to view it from different angles, I can't build it properly. But the vast majority of 3d aps start now as Windows native, and tend to take such things as three-button mice as standard. Even power Mac users don't get these easily -- nor do they play well with the rest of the Mac environment. But funny thing; an ap as otherwise ghastly as Carrara can implement a smooth 3d navigation and selection and even basic movement rotation and scaling tools with just a couple of control keys. So it isn't that F1-home key-scroll wheel is necessary to make 3d navigation work. It is just that many 3d ap programmers are, well, stupid.)



Oh, yeah. And sad thing is, the tool path tools are poor enough I'm going to end up exporting and generating tool path and so forth in a different ap anyhow. Which means that I might as well have built the thing in a poly modeler to begin with.

No comments:

Post a Comment