Tuesday, March 29, 2011

Creo, My Take on It - Part 2

Part 2: Creo and “AnyMode Modeling”

<< Part 1

There is a significant difference between using a parametric (history-based) system as your primary design tool and using a direct modeling (history-free) system as your primary design tool. These two technologies offer completely different methodologies and experiences for product development. Certainly a company that has standardized on parametric modeling may get value from the flexibility of direct editing in certain areas of the process, but if they have chosen the history-tree to be the engineering master (document of record), the history-tree needs to remain the master. If, on the other hand, a company chooses direct modeling as their primary design tool and understands geometry to be the engineering master, geometry needs to remain the master.

With Creo, I don't think PTC is somehow attempting to mix these two in some convoluted way, inevitably leading to a compromise in both methods and data. On the contrary, PTC is providing powerful purpose built solutions, tied together through the common data model.

For the company that has chosen to standardize on a parametric history-based solution, Creo provides a high-end purpose-built no compromise parametric modeling app AND a complimentary purpose built direct modeling app. This particular direct modeling app provides an option to track direct edits into the common data model. These records can then be accessed and consumed by a parametric app that can make sense of them as it relates to the ordered feature tree. The user of the parametric app is given a choice to accept or reject modifications. If accepted, the edits will be reflected appropriately in the tree. This direct app can be used in a parametric environment to allow infrequent users, or perhaps even non-CAD users of the parametric app to more easily participate in the design process. This capability makes it possible to take a fully parametric model into a much simpler direct environment, make edits as necessary, then bring it back into the parametric environment with NO data loss – maintaining the engineering master. I’ve heard some call it “round tripping” (P>D>P: Parametric>Direct>Parametric). PTC’s round trip is made possible via the common data model and is very unique. It is of course important to understand process to determine where this type of workflow makes sense, but when and where it does, Creo can support it.

With Creo, PTC is also providing the best most powerful direct modeling app on the market that is capable of supporting complete art-to-part. For companies that chose direct modeling as their primary design tool, and geometry as their master, Creo provides the best, no compromise solution. This app comes with a synchronous parametric solver when firm design intent is needed. These customers can also take advantage of the parametric app as needed if and when there are circumstances that are well served with the parent/child relationship of parametric modeling. Via the CDM, the resulting geometry can quickly and easily be incorporated into the direct design environment and carried forward as native data.

Part 3

Paul

24 comments:

Anonymous said...

Hi Paul,

Will you comment on the freeform feature of Creo in futur blog post, it look very interesting !

TIA

Anonymous said...

Very interesting. But you were "confused" with a similar workflow in Autodesk's Fusion saying that "My primary confusion with this concept is that if in fact the edits are possible in Inventor, then why do I need Fusion?" See 2009 - The Year of … Confusion?

Paul Hamilton said...

Anonymous #1 - Yes, I am looking forward to getting my hands on this capability. The demos I have seen are amazing to say the least. I will certainly write about it when I can.

Anonymous #2 - Good catch! I was wondering if anyone would call me out on this.

What Autodesk is doing though is different than what PTC is doing. Autodesk is attempting to edit existing features/parameters in the tree based on direct edits in Fusion. PTC chose not to edit any features in the tree, but rather add to the bottom of the tree features that represent the direct edits. Autodesk will do this as well, if it can't first find a feature edit that can support the given direct edit.

I do think that it's a bit risky to have non Inventor or Parametric users doing edits on geometry that will eventually (and potentially automatically) change the modeling intent. Changing the modeling intent is best left to the expert Inventor or Parametric user. Although I guess in both cases the option to reject is always there.

I do still question the overall workflow in general - in either case. In the end, the users, and process requirements, will decide.

Anonymous said...

I don't think that your description of Autodesk Fusion is quite correct. It's true that Fusion will offer a feature edit AS THE DEFAULT, whenever possible, to change the history-based model to match a modification done in Fusion. But even when a feature edit is possible it offers other "treatments" (like "Move Face", "Delete Faces" or "Extract Faces") which will add features to the bottom of the feature tree. In some cases Fusion even offers more than one such treatment for each Fusion change. So if a user prefers he can always select a treatment that does not edit any original feature. My understanding is that Autodesk chose to offer feature edit AS THE DEFAULT because it attempts to simulate what a human user would have done to accomplish the change made in Fusion. For example, consider a rectangular cutout produced using a extrusion cut, where the sketch rectangle used by the feature has two linear dimensions to define the size (width and height) and two dimensions to define the position (x and y location). Now let's assume that the four faces of the cutout were moved in Fusion while preserving the size of the cutout. When you pick the default feature edit option, Fusion changes only the position-defining dimensions. This is most likely how a human user would have edited the Inventor model to achieve the same change. So, what advantages do you see in adding features like "Move Face" to the bottom of the tree rather than editing features?

Since Creo appears to offer a subset of the "round trip" functionality that Inventor/Fusion offers could you please describe what is unique or better about it?

Kevin said...

I like having both parametric/history and direct modeling on the same part in the same app all the time! No tripping :)

www.ironcad.com

Anonymous said...

So if a Direct edit is made to a Parametric file, and the Parametric user accepts the edit and thus it's appended to the history tree, can Parametric edit that feature or does it have to be opened in Direct to edit that feature?

Paul Hamilton said...

Anonymous 1, I think it comes down to reliability and usability. I don't think this "round tripping" is in production use yet anywhere, so it's all theory so far. We can speculate here, but time will tell. There are several other differences in how these two companies are attempting to do this. We'll just have to wait and see.

Right Kevin, so do the Solid Edge guys :). The question here is about moving from ordered to non-ordered and back to order again, with no data lose. Order is propitiatory and complex and sometimes gets in the way - no mater how well the tool hides it. In my book direct editing of ordered geometry is completely different from direct modeling. Very different.

Anonymous 2, that depends on what solution you are talking about :) I'm sure we'll see a lot of changes here as this stuff matures.

Kevin said...

Here's what I was thinking of, with not feature data lost:

http://vimeo.com/21752377

Paul Hamilton said...

Kevin, We are getting a little off topic on this, but whatever...

That is an excellent example of direct feature modification - not "direct modeling" in my book.

Try this:
http://www.youtube.com/watch?v=JBNwrVSRw9M

An example I used with the TriSpective/IronCAD guys many years ago. Interested to see the results today.

In this example the intent was never to have the two holes aligned, it was just the fastest way to get the model I wanted.

Paul Hamilton said...

Kevin, I forgot one part. Do the edit, without losing feature info and order. That is what PTC and Autodesk is trying to provide.

Kevin said...

Now I understand, no, we aren't their yet either.

Kevin said...

We are close though:

http://www.youtube.com/watch?v=73IHE1ehul0

Dan said...

Paul,
We met in Peoria a few weeks ago. Can't wait to get my hands on Creo Direct ... I have a problem that I think we can address with this right now that is giving me fits with Pro/E alone.

Dan McCaherty

Anonymous said...

I quite understand your scepticism about "round tripping" between history-based and direct modeling modes since as I recall you always thought that "never the twain shall meet".

But then, if you discount that, what is "AnyMode Modeling"? You say that Creo will have a world-class history-based modeler and a world-class direct modeler. But PTC already has those in Pro/E and CoCreate. And, I assume that just like NX and SolidEdge (and other CAD vendors with multiple products) that Pro/E and CoCreate have decent interoperability -- that one can place and constrain a CoCreate part in a Pro/E assembly, that one can open a Pro/E part in CoCreate, etc.

One might be tempted to say that it's all about the "common data model". But I think that CDM is simply about software architecture, the stuff that is under the hood. Though I don't want to diminish the importance of good software design for a variety of reasons, as a user I care only about what the software does. So can you please describe in a few words the significance of "AnyMode Modeling" to users? What would users be able to do because of "AnyMode Modeling" that they cannot do today?

Anonymous said...

If you think you can do the round trip, you don't understand the true value of direct modeling.

Paul Hamilton said...

Dan, thanks for stopping by. Look forward to working with you on it.

Value of AnyMode Modeling to the User? Great question. Round tripping applies mostly to the company that has standardized on history-based modeling. As stated by the last anonymous comment, it offers little value to the company that has standardized on “professional grade” direct “modeling”. There are some use cases perhaps.

For companies that have standardized on history-based modeling, AnyModel Modeling and round tripping and the CDM may have little impact to the heavy CAD user, but to the process it can have huge impact. The full time CAD users will continue to use the professional grade history-based tool. With Creo this tool gets a full set of direct “editing” capabilities. No need to pop up the direct app and do the round trip.

AnyMode Modeling, Round Tripping and the CDM are all just “stuff” that makes it possible for the non-CAD experts and infrequent users to more easily participate in the engineering master. In PTC’s language they include 2D in ”AnyMode Modeling”. Consider a person that prefers to sketch concepts in 2D. For the most part that 2D data is completely disconnected from the engineering master, and if it is desired to use it in the engineering master, there is typically some duplication of effort. AnyMode Modeling, in my opinion, is about reducing duplication of effort in the process while allowing people to use the technology of choice - and just get their job done.

Anonymous said...

Kevin,

I just wanted to say thanks for participating. Would like to see you on Twitter as well.

Jon Banquer
San Diego, CA

Anonymous said...

The YouTube video for Creo AnyMode modeling shows two examples of geometry modification, where a "boss" and a "nozzle" are moved from one place to another, without losing their parametric definition. But both of these "features" do not seem to have any unchanged geometry that depend on them. The nozzle supposedly consists of 13 parametric features, but the direct modeling change has affected all of them. In such a case it's easy to see how a feature could be "cut and pasted". Are there examples that show how Creo handles directly edited features, that happen to have unaltered dependent features?

Paul Hamilton said...

With the concept that PTC is considering with Creo, the direct “geometry” edits do not affect existing feature definitions, sketches or parameters. The edits are represented in the tree as new “Flex Move” features and ordered at the end of the tree, as any new feature. A Flex Move feature will typically contain references to all the geometry that was involved in the edit, along with the transformation information of the edit. These Flex Move features are editable as any other parametric feature.

So to answer your question, direct geometry edits do not “affect” the existing parametric features. The existing parametric features are processed and generated as usual. Flex Moves are processed as they are ordered.

Anonymous said...

Thanks for the explanation about "Flex Move" features which was very helpful in understanding the Creo approach. Will there be a FlexMove feature for each "user action" in Creo/Direct? If you move two holes some distance first, and then move one of the two holes differently what sort of FlexMove features would you get? If you move a hole twice in Direct would you get a single FlexMove feature? Even if the user changed the model in some way between the "move hole" actions?

Paul Hamilton said...

Sorry, I don’t have that much detail yet. Hopefully I will soon. I can say that “created geometry” in Creo Direct will be represented as full parametric features in Creo Parametric. “Edited geometry” in Creo Direct will be represented as editable Flex Moves in Creo Parametric. I suspect that with each direct edit in Creo Direct you will get a Flex Move feature back in Creo Parametric. I’ll have to check on this.

Peace Opinion said...

where I will get more about parametric/history??

Paul Hamilton said...

Peace Opinion, I am not sure I understand your question. If you are looking for more information about Creo Parametric (a parametric history-based tool) please refer to http://www.ptc.com/product/creo. I do also have several other articles about parametric history-based modeling posted on this blog.

eagle said...
This comment has been removed by a blog administrator.