Monday, October 18, 2010

Solid Edge ST3, My First Impressions (part 2)

In Part 1 I described my observations regarding how Solid Edge ST3 lines up with some of the typical issues related to traditional parametric history-based modeling. In this Part 2 I will share some of the other crazy observations I have:

Here is a statement coming from “You no longer have to choose which modelling workflow to use up front, instead it is a simple case of switching between ordered (traditional) and synchronous whenever required.” And, by the way, if you “switch”, or move an ordered feature to Synchronous, there is no going back. It’s a one way trip. Siemens still provides no ability to rationalize the history-tree (ordered) from a direct model/feature (synchronous). Something that Autodesk is struggling to solve with their Change Manager. With SE ST3 it seems real simple to move ordered features to synchronous. Since it is a one-way trip, I wonder how long it will take before all your Solid Edge history-based (ordered) data is completely history-free (synchronous/unordered).

When reading about SE ST3 I started having flashbacks to the mid 90’s. This started sounding strangely familiar. At the time I was fairly close to the developers of TriSpectives at 3D/Eye, now known as IronCAD, (CAXA in China). Here is a good article with some of the IronCAD history. IronCAD has always been and is still a parametric history-based system at its core. Although it was only a few years ago that they exposed the history tree structure to the user. Due to the systems ability to automatically reorder the tree structure, it was infrequent that a user actually needed to manipulate the structure manually. IronCAD did several things different from traditional history-based modeling systems. I had included much more detail on IronCAD while writing this but as I was writing Deelip Menezes published an extensive review of IronCAD. So I will refer you to his review for more detail on the similarities of IronCAD and SE ST3. In summary there are many similarities, although IRONCAD has been providing this type of "mixed mode" capability for many years now.

It seems to me that the synchronous features in the ST3 tree simply represent a single B-Rep model organized somewhere in the tree. You might consider it what other parametric history-based systems refer to as the “base feature”, although with ST3 this B-Rep is ordered somewhere in the tree, and can contain UDF’s (user defined features). In ST3 these UDF’s are actually defined by a modeling operation. When you drag and drop a feature from ordered into synchronous, you are simply taking the geometrical results of an ordered feature and adding that geometry to this single B-Rep, and then removing the ordered feature definition. This single B-Rep becomes more complex, or detailed, with each addition. The resulting new faces in the B-Rep are tagged to keep them grouped as a UDF. Am I over simplifying this?

I am not saying that any of this is bad. On the contrary I like to see this kind of stuff pushing all of us in our thinking of CAD. But as you are considering this, just keep in mind what our ultimate objective really is; and that is - designing products (sometimes 3D models can help us do that).


Solid Edge ST3, My First Impressions (part 1)

It looks like Solid Edge is getting a lot of new functionality and modeling capability. I am not a user of Solid Edge and as such my “first impressions” have nothing to do with the new functionality. I will leave that up to the users and experts. What I am most interested in is what they have done with Synchronous Technology and the Solid Edge model structure.

When reviewing technology I try to see things from a process and workflow point of view. As you probably already know I am kind of partial to direct modeling. I certainly understand the value of the parent/child relationship that comes with the history tree, and the associativity that it can provide, but I see all too often that many designers are over-served with this technology. The challenges that come with history-based modeling sometimes overshadow the benefit. These challenges have driven and accelerated the development of direct modeling. The challenges I speak of include the following:

  1. Complexity of the tool and general training requirements
  2. The need to plan ahead before modeling
  3. Interoperability and data exchange
  4. Reusing history-trees (history-based models)
  5. Resolving conflicts in the history-tree
  6. The need for modeling standards and best practices

Again, I understand the benefits, but all of the above can keep you from getting your job done in a timely manner. And it is these issues that are driving the interest in direct modeling. There are still some yet unsolved challenges with direct modeling of course, but none of the above apply with mature robust direct modeling.

So my question is: How does Solid Edge ST3 impact the issues listed above? Is life improving for the typical Solid Edge user? Let’s take them one at a time.

  1. Complexity of the tool and general training requirements: In my opinion SE just became harder to learn. Even if you will never use Synchronous, or will never use Ordered, just having those buttons on the menu just made it more complicated. It is safe to assume that the training manuals just got thicker. It looks intuitive enough, but it goes deeper than that. How, when, why is it acceptable for a user to move a feature from Ordered to Synchronous?
  2. The need to plan ahead before modeling: It doesn’t matter if you are in a history-free mode (synchronous, direct, unordered, whatever) or in a history-based mode (ordered, parametric, whatever), if the system is recording information about the modeling process (modeling features) you will need to plan ahead and model carefully, if you intend to realize any future value from the record. In synchronous mode, the order may have no relevance to the future use of the model, but the collection of faces recorded in the feature certainly will. Will SE ST3 be a good concept design tool? The chaos of concept design often times renders a history tree completely unorganized and useless. Very often this will lead to a model rebuild when moving to detail design. Does SE ST3 resolve this?
  3. Interoperability and data exchange: We all know that history-trees are proprietary and that exchanging history-trees between different history-based systems is flaky at best. Geometry is common between all CAD systems, and with robust direct modeling we can make use of any geometry and edit it freely. My only problem with SE ST3 in this area is that I see no evidence that they have improved the robustness to handle the demands of complex direct modeling. I’ve had a few chances to try Synchronous Technology in NX and SE in the last year, but was never really impressed with its ability to resolve complex topology changes on imported models. Live rules or not, the system needs to be able to handle the demands of direct modeling. Like I said “I see no evidence” of improvement in this area so I can easily be proven wrong on this one. I just need to see it to believe it.
  4. Reusing history-trees: It has always been a bit of an issue when reusing a history-based model that you created many months/years ago or that someone else created. You need to spend some time studying the tree structure before you can really make use of it. It seems that with ST3 the tree structure just got more complicated. I know, you don’t have to use both modes in one model, but can you control the other users?
  5. Resolving conflicts in the history-tree: I guess if you do have a conflicted or corrupt history tree, with ST3 you can just take all those features and move them to Synchronous. That would fix it, right? What about all the other references and relationships tied between those ordered features and their children and/or other parts/features? How are those resolved in ST3?
  6. The need for modeling standards and best practices: I have seen some fairly thick 3-ring binders full of documentation on modeling best practices. These are usually created by larger organizations to train users and ensure that these users are creating history-trees that have some consistency in them in order to better ensure these models (history-trees) can be used in the future and by others. I think with ST3 these binders just got bigger. I can see some more best practices coming.

Solid Edge has certainly come a long way. The SE ST3 videos they have online look fantastic (as they should). I do commend the product managers and developers of Solid Edge for pushing the envelope on this. They have introduced some interesting concepts with ST3. I’m anxious to see how it continues to develop. I am still not convinced that they have the “best of both worlds”, as some fellow bloggers have already indicated. Sorry, I am just a bit skeptical. I hope it doesn’t turn out to be one of these:


Part 2