Direct v Parametric Smackdown commentary

Cadalyst magazine put together a podcast of a panel discussion on the topic of direct modeling vs parametric modeling. If you want to live the experience on your own, here is the link. If you don’t think you can bear an hour of people trying to interrupt Mike Payne , then this blog post might be more your speed.

The members of the panel were all active CAD industry employees, except Ben Eadie. How Ben got elected to represent SolidWorks is a little unclear, but more power to him. SolidWorks has a tendency to not participate in this kind of thing, which I’ve got to say I don’t understand. It results generally in SolidWorks getting pretty well pounded.

Anyway, the panel was:

Mike Payne – founding member of 3 major CAD companies – PTC, SolidWorks and Spaceclaim
Dan Staples – director of development for Solid Edge
Carl White – Autodesk digital design product management director
John Buchowski – director of development for PTC
John McCullough – Kubotek product management
Ben Eadie – SolidWorks user, and lots of other stuff

The moderator was Bill Fane of Cadalyst.

If you want to follow the twitter stream, the tag was #directvsparametric. Of the twitter stuff out there, I think Evan Yares had the most interesting stuff to add.

Anyway. It’s rare enough seeing folks from CAD companies cooperating on an event like this, not to say it wasn’t without its sniping and cheap shots. In the end, there weren’t any really new ideas that came out of it, no profound insights. But there were a few little nuggets that I would like to reuse here. Mainly it was interesting to hear what folks who are intimately involved in the battle between the two technologies think. In particular, I sprayed my desk with ice tea when Mike Payne started on an impassioned plea for interoperability between the CAD packages, and followed that up with “we should all learn how to get on a bit better”.

There was a lot of talk about interoperability, but I think all of these guys are insincere when it comes to that topic. They all want to be able to read other people’s formats but no one wants to let anyone read their format. It’s the same old same old, and the user is the one that gets shat on. SolidWorks is probably the worst offender in this regard, with non-compatible SolidWorks versions and non-compatibility between SolidWorks and Catia.  Direct edit capabilities start to erase the difficulties of interoperability, and even version non-compatibility, which is why these folks are willing to talk about the topic at all. There is a point where competitiveness gets in the way of collaboration, and that point came in the CAD world as soon as the second CAD product was available. The people involved with these products are too self-absorbed to actually deliver something that users really need in terms of bidirectional interoperability.

First. The title. Direct v Parametric. Some of the panelists weighed in on this too. I think it would be better called Direct v History. Almost any name you could come up with has some flaw, but to me, the two competing interests here are Direct editing and History modeling.

Second. Why a bunch of industry wonks? Nothing was really accomplished because they could only really agree on obvious stuff (like “use the right tool for the job”), but no one was really in any danger of changing their mind or admitting another company had a good idea unless it helped validate their own idea. Still, it could have been worse. They could have had press types have a discussion. I’m not sure that users would have improved it, it would have been a different kind of slightly less polite train wreck.

Ok, here are the little tidbits I gathered:

Mike Payne: “Design intent” is a misused term

Bold, and I agree 100%. I replace the term “design intent” in my books with “design for change”. You hear a lot of long winded explanations of what exactly design intent is. I think the explanations are complex because the term itself doesn’t make sense as CAD people want to use it. Completely off-topic, but it was interesting to hear this come from a guy who was responsible for inventing the term in the first place.

Mike Payne: (paraphrase) computer hardware power has changed what we can do – we used to have to be “feature programmers”, but now we can change 3D like we used to change 2D Autocad, just move lines around

Well, you know, moving lines around in Autocad wasn’t all it was cracked up to be, and certainly shouldn’t be a goal for a 3D modeler. So I’m not sure I really like that analogy, although I do see the resemblance. Maybe the analogy is accurate, but it doesn’t cast direct editing in a very sophisticated light. I get the thing he says about “feature programmer”, though. Setting up features in a history tree is a bit like programming lines of code, and it matters what order the commands are given in.

Dan Staples: Ease of use vs hard core modeling

I think Dan Staples was the guy who came out of this sounding the best. Maybe that’s because he’s got the product that I think comes closest to the ideal.

Carl White: You have to be able to use both

Again, I agree completely with this.

Dan and Mike both talked about a scenario where you have a cell in Excel, and one time you change the cell with an equation, and the next time you change it by manually keying something in. How can you expect the equation to keep working after you’ve done that kind of edit? I think this was a shot at how SolidWorks currently handles direct editing. SolidWorks does have some limited capabilities in that area, but they certainly don’t match up to any of the other products represented by the panel. Putting a history based feature in the tree that represents a direct edit change is a bad idea, I think. SolidWorks probably didn’t show up to this panel because the current tools are weak, and direct edit is something that will be delivered with the V6 product whenever that hits, and the new administration is being smarter about keeping its mouth shut.

It would be interesting if you could detach each one of these guys from trying to sell their particular idea. Mike Payne seems to be playing both sides of any fence you might name, and the fellow from PTC seemed more interested in scoring points than in a real discussion. The only product name I didn’t hear mentioned repeatedly was Kubotek. Again, I think Dan Staples really believes in his product, and the Solid Edge product is essentially his vision, although Dan did get a little lost in parochial terminology. I understood what he was talking about with Synchronous changes and PMI, but I’m not sure that stuff is general knowledge.

What would I have said if I were on this panel? Nothing. I don’t think I would have participated. My interests in the topic are more practical, and more of “what do I need” rather than in trying to show all other ideas as somehow misguided.

What do I think is better? I think a tool that I would use should be history based with direct edit capability. Closer to what Solid Edge does with ST3 than what SW does with Move Face. If you could take the ST3 model, and allow the direct changes to happen to the history-based geometry, that would be what I think is ideal. I keep referring to the SolidThinking method, where you essentially combine SolidWorks and T-splines all in one package, and you can make both kinds of changes without unnatural reordering.