You know, just last week the fellow in the cube next to me, who is an Electrical Engineer who knows nothing about CAD but has heard me rant over the last 4 years about the problems with non-history based systems, came up with a
great analogy that exactly describes the problem with non-history-based CAD packages.
He said:
"From your description of the problem, it sounds like someone took an Excel spreadsheet, with many cells driven by the contents of other cells, did a 'select-all', 'copy', and then did a 'paste special' with the 'values' option checked, and thus gutted all the relationships in the spreadsheet so that you had to manually do all the calculations by hand from then on."
From a non-mechanical guy, I was astounded at his grasp of the situation. This is
precisely the situation that a non-history based CAD package puts the designer in. In order to "save" the user from having to deal with imbedded logic errors, non-history based systems simply gut the logic altogether, leaving the designer, and every future designer who has to work with the model, to know, understand, and manually maintain all geometric relationships themselves. It's like taking the tires off of my car and then telling me how great it is since I will never have flat tires anymore.
Really, that's what this whole debate about the two approaches is all about. Do you want to be able to define relationships between geometric features or not? Do you want to be able to control things about the model that don't physically exist, like theoretical intersections, or basic diameters of tapered holes or bosses or not? Non-history based modelers today severely limit your ability to do these kinds of things. Yes, with history-based systems you sometimes have to think hard from time to time to debug these relationships. With non-history-based systems you have to think hard
all the time to maintain those relationships yourself. Eventually, most CAD users wake up one day and think, "Gee, wouldn't it be great if I could get the
computer to keep track of these design relationships?" That is how these systems came to exist
over a decade ago.
I guess there must be a market for this kind of CAD - CAD for people who can't understand the concept of imbedded geometric relationships, I guess. Just last week I heard of a new player in the non-history-based CAD market, "Spaceclaim" (
www.spaceclaim.com ) and watched their webcast demo. It, too, is a Boolean-style modeler, and it, too, seems to suffer from all the limitations that arise by not being able to define relationships in your designs. But they claim that out of some 5 million folks who could be taking advantage of 3D data, only 1 million or so actually make use of feature-based, parametric CAD tools, supposedly because they are too complicated for most users to understand. So they see those other 4 million folks as a good market to cater too.
Fortunately just this week I am starting a position at a new company that uses Solid Edge. I've never used it before, but after just 2 days I am already breathing intense sighs of relief at being able to, once again,
actually nail things down in my model and let the computer take over the drudgery of keeping those relationships straight! Here's to being back in the top one-fifth again!
Steve