Is the Direct Modeling Honeymoon Over?
In a comment, Matt Lombard remined me what Dan Staples, Director of Solid Edge at Siemens PLM, said to him last October:
"Ten years from now history based modeling will not exist"
Today, Dora Smith, Director - Global Social Media at Siemens PLM told me something which appears to quite the opposite:
"We have no plans to deliver NX or Solid Edge without a history-based option."
I believe 10 years is a very long time in the world of technology and would prefer to go with Dan Staples. I also feel that Dora is merely addressing the immediate concerns of her customers.
I simply cannot bring myself to believe that the best way to create and modify a solid model is by cooking it up in a sequential manner all the time. Features are here to stay, at least for the length of my lifetime. Whether they need to be created and maintained in an orderly manner is something that I am not comfortable with.
But there is another solution, something which just about everyone I have spoken to, including myself, feels is next to impossible. That is, to have a history tree containing features and allow the user to make direct edits to the features, but not add these direct edits as new features at the bottom of the history tree. Rather directly edit the existing features in the history tree. If such a solution exists then probably I could live with the history tree because I would not need to figure it out before I make any change to the model. Current history based modelers do only a fraction of this. I say fraction, because you can push/pull only those faces which have a clear relation to an underlying feature. Case in point, Instant3D by SolidWorks.
However, Autodesk claims that Inventor Fusion will do exactly what I said above was next to impossible. If that is indeed the case, then this year's Nobel Prize (doesn't matter which category) should go to the person at Autodesk who figured this out. I can think of few cases when this solution may be made possible, but there would be severe limitations to how much you can mess with the solid model. But then maybe I am not Nobel Prize material.
So to answer the question, "Is the Direct Modeling Honeymoon Over?", I say "No". I would say that probably the marriage between History Based Modeling and Direct Modeling has only just begun and they are now fighting over which one of them is wearing the pants.
posted by Deelip Menezes at
9:29 AM
9 Comments:
Deelip,
In any relationship between the two, I think history has to wear the pants.
I would be interested in your take on what is happening in the software SolidThinking. It very prominently features the combination of history modeling and direct edit. They don't say it that way, but that's what it looks like to me. I see a history tree, and there are edits that don't show up in the tree.
I don't know anything about Fusion, so I can't say how SolidThinking relates to that.
Another observation. To the best of my very poor memory, all of the proponents of Direct Editing as the New World Order for CAD are non-users, and the proponents of history are users. Even the direct editing users admit they can't get everything done with a single tool. Any comments on that?
By
Matt, At
6:54 PM, June 12, 2009
Deelip,
To the Dan and Dora comments, it seems perfectly consistent to me (but I'm on the inside so maybe that's my problem). No one in software plans out 10 years in advance, just makes predictions. We are really lucky in software if we can plan 3 years in advance. I think it is safe to say no one in CAD is "planning" the end of history just yet (but ask again in a year or two ;-)
By
Burhop, At
9:39 PM, June 12, 2009
Matt,
"History has to wear the pants" - hmmmm... I'm going to run with this analogy as much as it pains me.
"History" and "Direct" are not two 30 year olds. Assuming you are right today, history is more like 60 and "Direct" is 22. I think the one wearing the pants will change.
By
Burhop, At
10:01 PM, June 12, 2009
I'm enjoying this debate about whether a all direct modelers should be able to do everything that the massive heavy-duty feature-based systems can. At SpaceClaim, our customers are begging us to keep our product simple and not turn it into a mega-CAD system. That's precisely our goal. Our customers love an environment where they can organize their thoughts and hand off design concepts to the CAD team to detail.
Or to flip it around, how would one create a 3D solid modeler than just about anyone could use? Not ever engineer wants debate different ways of specifying design intent and structuring parent-child relationships. That's why we made SpaceClaim.
-Blake (A founder of SpaceClaim)
By
Blake Courter, At
10:03 PM, June 12, 2009
Deelip,
The reason Autodesk’s plans may sound impossible could just be a matter of terminology. From the way I’ve heard the terms evolving during the ‘honeymoon’ direct editing (directly using objects in the graphical model) and feature editing (using parameterized text-based shape descriptors) are both possible with or without history.
Without history a list of features can become merely a categorization of the parts of a model. A pocket feature, for example, can be thought of as a group of faces. By default it makes sense to add such a feature to the tree in the order they were created. For simple models (and well-planned demos) it can look like history based modeling.
Your quotes of directors at Siemens on this topic raise several interesting points to consider. I think Dora is saying that users who are comfortable with and invested in history-based modeling have nothing to fear. I believe it will be a while before all of them change their thinking and habits and utilize the new history-free options they are being provided. That is simply human nature.
Dan’s claim is actually exciting. At CADKEY and Kubotek we have been on this bandwagon longer than almost anyone. Based on what we see possible we agree with Dan in spirit that the overwhelming majority of mechanical design will not benefit from history-based editing and model storage at some point in the not-too-distant future.
John (Kubotek USA)
By
John Wright McCullough, At
4:11 AM, June 13, 2009
Matt,
I think I did look at solidThinking sometime ago, but not with great detail. But now that you mention it, I will revisit it.
As regards your distinction of users and non-users, I happen to have a view on that, which I will hope to state in separate post on this blog.
By
Deelip Menezes, At
11:14 AM, June 13, 2009
Hi All
Just to let you know that i would like to add my 2cents to the conversaton.
reply is long for the comment so i made a post.
http://soliddna.wordpress.com/2009/06/13/linear-or-non-linear-modeling/
By
Anonymous, At
3:38 AM, June 16, 2009
Hello Deelip
"but not add these direct edits as new features at the bottom of the history tree"
Why are you afraid of that? In my years as a designer, I have been on the receiving end of many a late change - being able to go back through the history and "change a change" is a life-saver in the real world. It also serves as a record and reference of modifications, of great value when reviewing a work-in-progress file with a customer.
Chris Ward. VX Corp.
By
Chris Ward VX Corp, At
3:50 AM, June 18, 2009
Chris,
I agree. I does have a few advantages. But you would agree that this more of a hack and whack approach. The whole point of history based parametric modeling was to be able to go back in time (history) and edit a parameter in order to effect a change. The fact that people often opt to hack and whack (apart for the reasons you mentioned) instead of doing it the hard way says something.
In programming we face the same situation. It is better to understand the code and make a modification where it belongs, rather than add a patch later down in the code. It just makes messy and sloppy programming. History based parametric modeling should be no different.
By
Deelip Menezes, At
11:48 AM, June 18, 2009
Post a Comment
Subscribe to Post Comments [Atom]
<< Home