The Democratization of 3D
In the past I’ve written about a few topics that I’d like now to pull together with a common thread. One topic is the difference between modeling and design. Another is the idea that artist types sometimes need a translator to get into digital 3D. Yet another is the new non-parametric CAD program SpaceClaim. The last concept I want to pull in with these others is that when you automate a task, often you simplify it, maybe even oversimplify it, and give up some control. The common thread that pulls all of these ideas together is the CAD Specialist.
The mid-90s saw a wave of mid-range 3D modelers flood the market, beginning to erode the share of the high-end systems, and bringing some folks up from the 2D flatlands. When the only people running 3D modelers were priviledged to do it because it cost nearly 6 figures to get into it with hardware and software, the level of the user tended to be pretty high. They were well trained and probably some of the brighter folks in their offices. This was when 3D CAD was a pretty exclusive club. These users tended to be full-time CAD specialists, because it was not a casual thing to pick up and start using these systems.
SolidWorks, Mechanical Desktop and Solid Edge all hit about the same time in 1995. This was when Windows NT 3.51 and 4.0 was trying to win some of the technical computing business away from Unix. These low-cost, modelers with a relatively low cost Windows based PC were attractive because the entry price was now about the same as a full license of AutoCAD. When the price point came down, and 3D started to be used by more users, the average user was no longer the well trained geometric whiz that was running $30k of software on a $40k Unix box, he was running about $4k of software on a $4k box, and in the same proportion, may have also had about 10% of the training of the high end users.
In order to dig deeper into the market in this way, the mid-range products had to be simpler to use and more automated. The interface had to be easy to navigate and pay more attention to details glossed over by the high end systems. The users were maybe only part time users. The price point had come down, the functionality had come down, but the usability had gone up.
I believe the CAD market is going through another one of these adjustments. We have seen the introduction of Alibre, a SolidWorks 95-ish product which has lowered the cost of getting into a parametric 3D modeler even further. And now, with SpaceClaim, both cost and the specialization bars have been lowered yet again. They are marketing this product to non-CAD specialists, and even non-CAD users. You think sending engineering jobs to India is a frightening prospect, imagine leaving the modeling to the machinist, manager or the marketing guy!
Unix CAD comprised the top one or two levels of the pyramid. Mid-range CAD started out in the second and third levels, but has recently been pushing up into the top level as well. Products like Alibre push down into the Ocasional user level. Pushing CAD down further yet into the non-user market vastly increases the number of potential customers, but it also demands less specialization of knowledge (to say it differently, it has to be “dumbed down”). SpaceClaim is aiming at the Ocasional Users and Non-user Technical levels. Non-user Non-technical is reserved for products like Sketchup, although Sketchup is also used higher in the pyramid for fast and easy concepting where final detail is not crucial.
Maybe I’m just protecting my livelihood when saying this, but what happens when you take a tool that can do some pretty complicated things, and you make it very automatic, which enables you to do the work without much forethought, planning or insight. This must be what went through the minds of contract FEA analysts when CosmosExpress became available – allowing anyone with a SolidWorks license to do “analysis”. The Express tool is so dumbed down that literally anyone can get results out of it. Stress analysis, formerly the exclusive domain of specialists, has now been opened to all. The curtain of the Holy of Holies has been rent.
Still, is it really analysis if you’re just going through the motions, not understanding what is going on? How much value is in the answer if you don’t even understand the question? I think the same question is valid when talking about “dumbing down” the controls to editing a CAD model. Sure, SpaceClaim enables machinists to edit models from any CAD package as if they were native SpaceClaim data, but should they? (There’s that awful word again.)
Anyway, your assignment for next time is to think about this. Think about where you are on the pyramid. If you’re high on the pyramid, all of your eggs are in one basket, and in a market which wants to do away with specialists, you are a marked man. If you are low on the pyramid, do you enjoy the other tasks that generalization allows you to be involved with? Wherever you are, are you able to get the job done with the tools you have? Would you trade some control for simpler, more automated tools, or is it the other way around for you?
Comments?
What you haven’t touched on here, although the title hints at it, is where draftspeople and machinists and others fit in all of this. Since the industrial revolution, machinists have taken drawings from non-machinists (designers or drafters), interpreted the 3D intent of 2D projections, and translated that into something producible. Recent access of 3D CAD and modeling has brought about a congruence of artists, technical gamers, manufacturing and marketing technicians who are able to wield 3D design as well as engineers, scientists and software writers/publishers. What is a “CAD professional”, anyway? He or she could have arrived at that position from anywhere else. A specialist, these days, is simply a professional (not necessarily from a formal background, either) who concentrates on a particular area.
Is automation a good thing? Well, I guess that depends.
I (less so than many others here) have certainly spent plenty of time figuring out how to obtain what I wanted from a CAD system. Many of the automation features have added useful functionality for even the “CAD professional,” although some more than others.
My bigger concern is the analogy you’ve drawn between the “dumbing down” of FEA and the “dumbing down” of CAD.
In my company, among others, automated FEA has lead to a false sense of security in bad designs. That’s incompetent engineering, and there have been expensive repercussions in both a financial sense and in the outside perception of engineering professionals.
And yet, many of those same automation features are invaluable tools for the FEA professional, assuming that the manual control is still available. (my main gripe about CosmosXpress) Are there still FEA analysts who never use automeshers? Or for that matter, who don’t use pre/postprocessors?
If the automation of CAD functionality leads to poor quality models, are the products similarly impacted? Certainly poor models can lead to non-producible or less attractive products, or expensive tooling. But is that due to the tool being too automated, or the operator being incompetent?