CAD Interaction Advice

I am a CAD teacher in high school. Mechanical engineer turned teacher so i know CAD, but slim on game design. I know there is a big push for interaction for applications such as virtual walkthroughs. My guess is that is the whole reason why Unreal studio and datasmith was created. My question is:

Is it wise to have a student model environments/static assets within the game engine itself? Design the product in technical CAD software, but model assets like chairs, couches, and beds within the game engine.

We have maya and max, but an aspiring architect really does not really want to learn 3-4 different software packages to get a walkthrough…

No, the modeling tools in UE4 are not adequate to do more than model blocky objects, and mostly are there just to block out a level before getting replaced with detailed meshes that you’ve imported in. Also, as far as how UE4 handles that type of geometry it has a bigger performance impact than static meshes that you import.
Ultimately, if it did have better modeling tools then it wouldn’t be any different than learning another 3D software. So if you’re unable to make something in your CAD software, then you’d need to use a different 3D modeling software for that.

Several game engines, including UE4, have been working on improved modelling tools within the editor. Its not been top priority, and they are usually keen to point out that this is not supposed to replace 3rd party modelling software completely, or anywhere close. There are some signs that the UE4 tools might be ready to be included in early preview/experiemntal form in a coming UE4 release, but even when that happens it will take time for learning materials to be available etc.

And yes, there is also a push to add more tools to integrate game engines with 3D modelling and CAD programs, as game engines seek out new markets and industries.

I suppose I might take this further and ask whether they should be modelling props at all, as opposed to picking them from libraries of models others have already made?

Even if the students can dodge learning so many different packages, they are still going to end up having to learn a range of different tools and forms of modelling if they need to make all the ingredients themselves.

Traditionally I’m sure the more common view of which bits a game engine is best suited at are more along the lines of blocking out ideas and constructing the overall scene/level and sorting the other ingredients such as lighting. And there are probably a dizzying array of workflows and variations in exactly what combinations people use. Personally I’m not a huge fan of most established modelling programs and so intend to push the ‘do as much in the game engine as possible’ idea as far as possible, but I have no sense of how many others will take it that far. Not am I an architect or terribly familiar with that world, although I imagine the immediacy that can stem from working with ideas and concepts in a game engine could be a big plus - eg constructing the rough form of the building in game engine and then being able to tweak very quickly and do stuff like mess with artificial lighting and sunlight direction to see the implications of form refinements. But the way some of the bridges between modelling apps and game engine are evolving, there isnt necessarily much more of a delay to this immediacy if the changes to the building model are made in a different app and then automatically updated in the game engine.