The Death of Drafting - Model Based Design

Way back in the day, back before the Internet, back before some folks on the Bentley boards were even born, the term CADD was making itself known, promising less paper, less time, less stress, and overall better design. 

Sure, sure.

Anyway, as of 1980 or so it was quickly becoming apparent that the days of the drafting board were bound to end, so superior was the new CADD system. One had to adapt - learn how to use the computer and how to draft using it.  That, or be left behind.  So, most AEC workers - drafters and designers alike - proceeded to do what would let them remain gainfully employed - they learned CADD.  It was probably one of the biggest paradigm shifts in the history of Engineering.

Are you ready for another one? Because it's coming: model based design, or MBD.  And it will make for another adpatation... or will it?  There are some people - and a lot of 'em - that may soon be out of a job.  Think about it: when drawings went from drafting table to computer, the person drawing it merely had to draw on the computer instead of by hand.  The way the whatever-it-was was designed didn't change.  Put another way, the person drawing it didn't necessarily need to know how it worked, but rather what it looked like.  But as computers and the software they run became more powerful, a lot of realizations hit: with the ability to "draw full size" and in 3D, it was now possible to construct a 3D model of the full-size project, and manipulate it as needed.  The advantage of this was self-evident, but this presented a new problem: how did one construct a model that way? 

The explanation for that is the new change - the one that a lot of drafters might unable to make.  Not unwilling, merely unable.  Before you throw stones at me, let me explain.

Consider a part for a car, a widget (kind of explains why you visit your mechanic so much, right?).  Anyway, to create the widget in CADD, one needs to geometrically define it.  So it is drawn in CADD, beginning with the 3D edges to lay it out.  Once one has the edges, then there is the need to assign faces, so the program knows what's a hole, what's not, and so on.  Finally, properties and materials get assigned so the software knows what is actually a voulme, a void, or what material should the thing be rendered as.  But in the end, the widget seen on screen is at best a 3D solid of some sort, constructed from basic geometric entities.

Enter model based design.  Lines, arcs, circles and such and not drawn; rather, one defines and places intelligent objects.  What is displayed on screen is still a wireframe of the model, to be sure, but you can't simply edit it the way you would a CADD drawing.  Delete a line - wait. Can you?  It may be the edge of a part or pipe, and be integral to the object.  It's one big ball o' stuff!  To delete that line, you need to delete the whatever it is and somehow draw it back without the line.  Or make it display differently.  So if you delete the whole thing - wait, should you?  It's in CADD, but what does it represent ?  Are you deleting a supporting beam, a control module, an axle connecting to an engine, a fill slope?  One needs to know what that line represents in order to properly work with it.

I'll tell you what it represents: an idea coming to reality.  And that idea was thought up not with drafting in mind, but designing.  And there is the crux of things: for a meaningful model to be created, someone needs to know how to 1) design it, 2) the use of the modeling software (MBD) itself, and 3) and the environment it is created in (CADD).  A pure drafter - again, nothing against those folks - will almost certainly not know the first thing (if they do, they underrepresent themselves), they may or may not have experience with the second, and will know about the third.  It's very unlikely they will be able to reliably edit or produce models - to do so requires knowledge of the design priciples behind what is being created in the software, to say nothing of how to actually use the software to create that model.

It took me over three years to get my head around this concept of MBD.  With a package like Revit, design became easier for architechts because the entities they now drew were "smart".  Used to be, if they drew a window where an electrical line was going, nothing in the software would let them know that. It was just a bunch of 3D lines and faces.  Now, though, a window is a window and a electrical line is recognized as exactly that. The software, knowing you can't have the two in the same spot, will ring the gong, tell the designer what their design looks like at that particulalr moment, and throw them off the show.  It's a model, almost like working with glorified lego pieces, and the software can identify conflicts, count quantities, and so on. 

The same thing is occuring in the plant industry. AutoPLANT, anyone? And then, too, in my neck of the woods - civil work.  I use InRoads.  I love Inroads.  I like making flythroughs with InRoads... but I'm getting off the subject.  Anyway, I can display my InRoads model but unless I have graphic group turned on, it's displayed as CADD objects that anyone can manipulate.  So the model is there but not in the CADD file, and when I see it I can edit it like ordinary CADD objects. 

However, with Autodesk's Civil 3D (please no stones... please!) this cannot be done.  It simply cannot - Civil 3D is a MBD program yet it makes me just uneasy enough to use - and I say that as a pretty decent user of things CADD and 3D. I know just enough about Civil 3D to call myself barely okay at it. I proceed extremely cautiously with it, because deleting or modifying the wrong thing can really causde headaches. I'm not sure I want to see my road design go MBD, because I feel I have a ways to go before I am comfortable with it myself.

Think about this: who is the best CADD user in your office, your CADD Guru? Now, think of the vertical application (Revit, InRoads, Civil 3D, GEOPAK, AutoPLANT, take your pick...) your office uses most and who the "guru" at that application is. I'll bet you three things:

1. Unless the CADD Guru and Vertical Guru are the same person, the CADD guru doesn't use that vertical app - or only rarely so.

2. The Vertical Guru is probably one of the best (if not the best ) CADD users outside the CADD Guru.

3. If forced to part ways with one of those two people, the CADD guru would likely be seen gone first.

And this is not to say the CADD Guru has no worth. That person has an amazing knowledgebase - and I am far, far better off for that person's efforts in my office. But the CADD Guru would be almost always be let go first. Why? Because creating design models requires understanding and knowledge they most likely do not have. MBD demands that knowledge.

I hate to say it, but the days of the drafter are coming to an end. And that's sad, because drafters are by and large technical people.  They have skills a lot of people would probably benefit from having, or want themselves. But how come there are not many people coming straight out of school who draft, and not design, no matter how good they may be in CADD? Yes, a lot of new AEC professionals come to the field already knowing CADD at least, but most of them now arrive having had the chance to work in the verticals too - creating (drumroll please)... models!  They just do. They come having been taught MBD, or experience with it at least.  As a part of engineering and design, upcoming AEC folks are taught MBD.  So it becomes almost a certainty that a new designer/engineer knows model based design and already has a superset of the skills most drafters have.  And they have the how-and-why knowledge to create models, too.  Once they are created, though: nearly all of the drawing is done! 

Where does that leave a good drafter? 

  • I'd make the same assement regarding where we're headed, towards MBD, but I'd make a different assement of what that means to the drafter.  As with the invent of mechanical automation, where we see a decrease in man hours to do the work, we also see an increase in man hours to create and maintain the process.  For every man replaced by an automated process, there is a man (plus?) needed to dream-up, develop, check, or maintain that process.  In my estimation, we might be headed towards fewer super-CADD users needed to operate this MBD workflow, but we'll also need a broader pool of support users to keep it afloat.

  • I'd make the same assement regarding where we're headed, towards MBD, but I'd make a different assement of what that means to the drafter.  As with the invent of mechanical automation, where we see a decrease in man hours to do the work, we also see an increase in man hours to create and maintain the process.  For every man replaced by an automated process, there is a man (plus?) needed to dream-up, develop, check, or maintain that process.  In my estimation, we might be headed towards fewer super-CADD users needed to operate this MBD workflow, but we'll also need a broader pool of support users to keep it afloat.