The interview is about 12 minutes long. About 2 minutes in, Bass makes possibly the most interesting comment in the interview:

“I’d say two to three years from now, every one of our products will be used online. The only way to use them will be online.”

If that happens, it will be a truly transformative and disruptive event.

But how can it possibly happen?

Autodesk has, over the years, built or bought a large stable of high-end graphics applications, including 3DS Max, Maya, AutoCAD, Inventor, Revit, and many others. The company continues to invest massive resources in all the tedious details of updating, fixing, and connecting these products. Customer wish lists still include items from years ago. Performance and bugs are still issues. Within this context, how can Autodesk transform the architecture of these major applications, used by on the order of 12 million people, so that they will run efficiently online?

Possibly the answer will reveal itself. Two or three years from now. If they manage to pull it off, it’ll be impressive.

3D CAD Tips

A couple of years ago, Autodesk Labs previewed a product, Project Krypton, which ran inside of 3D CAD programs (including Autodesk Inventor, DS SolidWorks, and PTC Pro/E), and gave real-time feedback on manufacturability, cost, and sustainability of plastic injection molded parts.

Project Krypton has now reappeared, in commercial form, as Autodesk Simulation DFM (Design For Manufacturing.) It works as a plug-in, running in a number of versions of Inventor, Inventor LT, Wildfire, Creo, and SolidWorks. It is available as a subscription benefit for Autodesk Simulation Moldflow Adviser 2013 subscribers, or as a stand-alone product, at US$2,000 for a license to run on any of the supported CAD platforms.

It’s reasonable to argue that engineers who are designing plastic parts should know enough to be able to recognize manufacturability, cost, or sustainability problems. And, if they don’t, they should take the time to learn (for example, by taking a few hours to read any of the many freely available books on the subject, such as General Design Principles for DuPont Engineering Polymers.) Even though that argument is reasonable, it doesn’t recognize human nature. People, even engineers who should know better, don’t always take the time to “read the manual.” Often, it makes sense to build the “manual” into the tools that engineers use every day. Simulation DFM does that, and quite a bit more.

For inexperienced designers, Simulation DFM provides quick feedback to help them avoid rookie mistakes. It’s sort of like an “idiot light” on a car’s dash, that warns you when something is wrong. And while old-hands might say they prefer gauges to idiot lights, experience has shown that idiot lights are useful to experts (even F1 drivers and fighter pilots) for catching their attention, and getting them to actually look at the gauges.

Simulation DFM doesn’t require that users have any background in molding simulation. It uses “green is good, yellow is not so good, and red is bad” indicators to identify potential manufacturing, cost and sustainability issues, showing the source and location of the problem. Any issues that pop-up can be expanded upon, to provide more detail on the exact source of the problem, even showing, for example, mold filling analyses.  The software requires no additional training, and doesn’t require much user input.

The open question with Simulation DFM is “how good is it?” Since it’s based on the Autodesk Moldflow simulation engine, it should be quite good, even for relatively complex parts (though it doesn’t support multi-body parts.) Yet, even if its capabilities were modest, it would still be of value, in either helping beginning designers to learn good design practice, or helping old-hands catch mistakes they might have otherwise missed.

As an engineer, I’ve long had the habit of using the “anything I can see” test to evaluate the usefulness of software. I look around the room, looking at anything I see, and ask myself “would this software have helped the engineers who designed these things?” In this case, as I sit in my office, I can see at least 20 items (without even turning to look behind me), each with multiple injection molded parts, that would have been quicker, easier, and less-expensive to design, had their engineers had access to up-front DFM software, such as Autodesk Simulation DFM.

The most significant benefit of Autodesk Simulation DFM comes not from its detailed capabilities, but rather from its clean integration into the design workflow. A user need not press a button, or take any specific action when designing a plastic part to benefit from it. All they need to do is notice, as they design, whether the software has picked up any obvious red-flags.

That Autodesk decided to make Simulation DFM available for Pro/E, Creo, and SolidWorks (as well as Inventor) shows that rational minds sometimes do prevail: There are untold thousands of PTC and SolidWorks customers who design plastic injection molded parts, and who are unlikely to switch primary CAD tools any time soon. The challenge Autodesk is going to face is in getting Simulation DFM in front of those users (since PTC and SolidWorks sales reps and dealers are not likely to recommend it.) Maybe not so much of a challenge: Many of Autodesk’s existing Moldflow customers are Pro/E and SolidWorks users.

There’s a certain charm to software that does something of great value, but does not impose any extra demands on its users. Autodesk Simulation DFM looks like it may be that kind of product.

Autodesk www.autodesk.com

Autodesk SimSquad simsquad@autodesk.com

3D CAD Tips

The new browser-based access is powered by the Autodesk Inventor Engineer-to-Order Server, which includes the ETO (Intent) Rules Engine and the Inventor Server (for model and drawing generation), as well as web services and server farm management software. Graphic display is via the Autodesk DWF format, for browsers with the Autodesk’s DWF Viewer browser add-on, and via raster graphics otherwise.

The Inventor ETO Server is licensed on a per-server basis, supporting 10 concurrent users. The included server farm management software supports load balancing and scaling. Because the server framework is session-based (i.e., not stateless), system requirements are about the same as for Autodesk Inventor. With big and complex models, you’re going to need to have pretty stout servers.

Autodesk is not currently offering Inventor ETO Server with software-as-a-service (SaaS) licensing, though, from a technical perspective, there doesn’t seem to be anything to prevent this. The software can be run in a virtual machine (VM), and hosted on a cloud service. The applications at http://etosamples.autodesk.com, for example, are running on Amazon EC2 instances.

Applications to be deployed on the Inventor ETO Server are created with the Inventor ETO Series product, using a Visual Studio-based development environment, supporting the Intent language and .NET languages like VB.NET and C#.

While the Intent language has evolved and been modernized for .NET compatibility, and based on feedback from users, its heritage traces back to the mid-1980s, to ICAD, one of the pioneering products in Knowledge Based Engineering (KBE.)

The Intent Rules Engine used by the Inventor ETO Server is powerful enough to implement nearly any sort of engineer-to-order application you could envision. It can be used to capture geometric and configuration knowledge, as well as business rules. Because the Intent Rules Engine provides the capability to create dependencies between designs (objects), it effectively allows the creation of workflows.

Out of the box, Inventor ETO Server has no ready-made integrations with enterprise systems, such as ERP, SCM, CRM, PDM, or even, for that matter, Autodesk’s new PLM 360 product series. This is not to say such integrations are not possible or practical. Autodesk has done integrations, for example, with ERP and CRM systems, either by direct access to the ERP/CRM database (support for Oracle, SQLServer, Access and IBM DB2 is included), by reading a database extract file from the ERP/CRM system, or by reading an XML based export file from the ERP/CRM system.

Autodesk has a number of large implementations of Inventor ETO, and has apparently had some solid successes with the product. Swedish hydraulic press manufacturer, AP&T, for example, notes that Inventor Engineer-to-Order has helped it reduce cost estimate errors on key components from 10% to 1%. Hytrol Conveyor currently uses 800 seats of Inventor ETO (and is likely a good candidate for the new web-deployed version.)

Interestingly, all of the companies referenced in Autodesk’s customer showcase worked with Autodesk Consulting to develop their Inventor ETO applications. This is not a big surprise. The Intent rules engine and language are definitely powerful, but they’re not for dilettantes (or average Inventor users with no programming skills, for that matter.) You can get a sense of this for yourself, by looking at the source code for Autodesk’s Inventor ETO samples.

Web deployment greatly changes the reach, and the economics, of Inventor ETO. Rather than deploying their Inventor ETO apps on notebook computers carried by salespeople, companies can deploy those same apps on the web, and make them available to their customers, worldwide, 24 hours a day. On a per-licensed-user basis, Inventor ETO is more expensive to deploy over the web than on notebook computers—but, when you account for actual utilization of the software by concurrent users, it’s likely far less expensive.

The actual licensing cost of Inventor ETO is probably only a minor part of the total cost of an implementation, when the cost to develop and deploy applications is factored in. The decision of whether to acquire Inventor ETO probably requires some careful analysis. If you are already an Inventor shop, you have a significant sales volume in configure-to-order or engineer-to-order products, and you have a commitment for enough budget, resources, and time to do the implementation right, you’re probably on the right track.

Autodesk Inventor Engineer-to-Order

3D CAD Tips

Why so slow?  Because Creo was recalculating the blend on the fly, as I was dragging.  Since the blend was in the middle of the feature tree, with a bunch of other dependent geometry, each time I moved the cursor, it had to do a very time consuming recalculation.  The right way of resizing the blend would have been to click on the blend radius text, and type in a new value.  Presto… the blend is resized exactly as I wished.  Alternatively, I could have used Creo’s flexible modeling tools, which would have recalculated the blend more quickly (though, when I tried to interactively drag the blend radius using flexible modeling, I still found the dynamic response to be unsatisfying.)

Is there a way to fix the problem I’ve described?  Yes.  Were I moving a boss across a surface that forced topological changes, Creo would have switched to a simplified representation of the boss, rather than recalculating it accurately on the fly.  So, Creo does know how to adjust it’s dynamic response depending on the needs of the situation.  It just doesn’t do it with blends.

The point of talking about this isn’t to beat up PTC. Other CAD vendors have similar problems. Creo is a very powerful product, and expert users can make it sing.  But put a user with a black thumb in front of it, and they’ll often do things that will bring it to its knees.

If PTC were building Creo expressly to satisfy expert users, the kind of experience I’ve just described wouldn’t be an issue. Expert users are smart enough not to do dumb things.  Yet, PTC is going to increasingly find that its products are used by normal folks (not Pro/E gurus), who are less likely to understand the program’s nuances, and who are less likely to be accepting of its quirkiness.  It’s important to get the little details right, so that all users, and not just the experts, can get the most out of their CAD programs.

3D CAD Tips

What is Creo?

Creo was rolled out in the Fall of 2010. It was, at the highest level of abstraction, a bet-the-company rethink of PTC’s CAD strategy, based on a recognition that not all users (or enterprises) have the same needs.

While I can’t say, with certainty, what brought on this revelation, I can speculate that PTC’s 2007 acquisition of CoCreate was a big eye opener. It doesn’t take a genius to figure out that PTC’s legion of salespeople would have swarmed into CoCreate’s 5,000 customer accounts, seeking to convert those people to Pro/E users. Like that was ever going to work.

The reason people (and companies) used CoCreate software was precisely because it was not Pro/E (or any other history-based solid modeler.) CoCreate’s dynamic modeling (more commonly called direct modeling these days) was comparatively easy to use, especially for people who weren’t CAD gurus. Companies with product lines that fit within the capabilities that CoCreate offered had no motivation to change at all. CoCreate customers that needed history and parametrics in their CAD software had long since moved on to other tools (including Pro/E).

I imagine that a number of CoCreate customers took the time to explain to PTC management why dynamic modeling fit their needs so well. At the same time, PTC management was likely watching the former CoCreate marketing people (who now worked for them) telling a compelling story that didn’t exactly jibe with the PTC’s historical “parametrics will solve all your problems” message.

What to do? How to rationalize these seemingly irreconcilable things? The only reasonable answer is to offer customers what they want. For PTC, this required a new strategy: Offer a range of products sharing a common data model and a common user interface design, and allow users to choose whether they want to use history-based, direct, or any other form of modeling that might come along in the future.

It’s a good vision. But getting there is the challenge.

The first phase in the Creo strategy was launched, with quite a bit of fanfare, in the Fall of 2010. The next phase was originally due to launch in the Fall of 2011. It’ll launch next month.

I can’t give PTC any grief about being a few months late in shipping, given the immensity of the task before them. Taking several very disparate and complex products and merging them into a family of interoperable apps is not easy. Truth is, CAD is hard. Developing a professional CAD system is about an order of magnitude harder than, for example, developing a product such as Microsoft Office.

It’s going to take some time for PTC to fully deliver on the Creo vision. No problem with that, if they really deliver what users need.

Creo 2.0

I’m not going to do a review of Creo 2.0, or even tell you in any detail about what’s new in it. What I will say is that it looks like PTC is making solid progress, and is delivering useful capabilities that will help their users to get their jobs done better.

In the pantheon of Creo products, Parametric (formerly known as Pro/E), and Elements/Direct (formerly known as CoCreate) are maturing nicely. Creo Direct—the new product that essentially merges Pro/E and CoCreate capabilities—is taking some time, if only because it’s a much bigger job. Users with a long history (so to speak) with CoCreate should look carefully, to see whether Creo Direct has reached their particular threshold of “good enough.”

The big news in Creo 2.0, beyond maturation, is PTC’s delivery on their AnyBOM strategy. They are shipping the Creo Options Modeler, which works with Windchill to support assemble-to-order processes in a way that PTC’s largest customers (e.g., Caterpillar) will certainly appreciate.

PTC will be talking a lot more about Creo 2.0 in the very near future. For now, what I can say is this: It looks good so far.

PTC www.ptc.com

3D CAD Tips

Here’s an amateurish video I shot at SolidWorks World with Mark Driscoll, of 3Dconnexion, showing their new SpaceMouse Pro.

3Dconnexion

3D CAD Tips

SolidWorks (the company) got major flak from bloggers concerned that SolidWorks V6 (the software) would replace SolidWorks (the software.)

This year, at SolidWorks World (the show), Dassault Systèmes SolidWorks (as the company is now known) didn’t talk much about SolidWorks V6 (the product), other than to say that they’d talk about it in 2013.

Now that we’ve got that all clear, let’s talk about what matters: SolidWorks V6 is confusing branding. It confuses not only users, but even pundits who write about CAD software.

My understanding, after talking to company representatives at SolidWorks World, is that SolidWorks V6 is the name for not just one product, but a future series of products. Those products may incorporate some existing SolidWorks technology, but they’ll be based largely upon CATIA and ENOVIA V6 technology. Because they’ll use the CGM modeling kernel (which was originally written for CATIA V5), they’ll likely be more compatible with CATIA than with today’s SolidWorks.

It makes sense that Dassault Systèmes would want to leverage the strength of the SolidWorks brand for this upcoming series of products. The SolidWorks brand is one of the strongest in the MCAD world. If SolidWorks V6 were actually based on, and entirely compatible with, SolidWorks—the name might fit. But it’s not, and it doesn’t.

The SolidWorks V6 name creates unnecessary fear, uncertainty, and doubt among SolidWorks users who are concerned that they’ll be forced to transition from a CAD program they know and (sometimes) love to this new technology, whether they want to or not.

What’s particularly unfortunate is that, if Dassault Systèmes had originally used a code name for the technology instead of calling it SolidWorks V6, they never would have created this whirlwind of FUD among their users. People might have seen it as just what it is: A really interesting future product, that they might want to add to their portfolio of CAD tools some day (when it’s ready.)

The bottom line is that the new technology called SolidWorks V6 isn’t SolidWorks, and won’t replace SolidWorks. According to Fielder Hiss, SolidWorks VP of Product Management, the development team working on SolidWorks 2013 is even larger than the teams that worked on previous versions.

The real SolidWorks—the CAD program now used by about 1.7 million people—is going to be around for a long time.

3D CAD Tips

During the same period, other companies also sold fast Camaros. Though GM’s official policy in the late 60′s and early 70′s was that they didn’t support drag racing, there was a way to get nearly drag-ready cars, if you knew the trick. A few dealers, notably Yenko Chevrolet, managed to get Chevrolet to install Corvette 427ci L-72 engines in Camaros, through the “central office purchase order” process. These factory hot rod COPO Camaros came with a full factory warranty. Nearly perfect examples have sold for over $2.2 million USD at auction.

Muscle cars have little to do with CAD, but I was reminded of these cars, at least by analogy, when I was at the SolidWorks World 2012 show, in San Diego, last week.

While there, I attended a press conference announcing HP’s new Z1 engineering workstation. This machine is sort of analogous to a factory hot rod. It comes with a stunning 27” built in display, a quad-core Intel Xeon processor, NVIDIA Quadro graphics, and uses ECC (error correction code) memory—which is particularly desirable for critical engineering software applications (See Wikipedia’s entry on ECC memory for background on this.)

There’s no doubt that the Z1 costs more than a typical commodity PC. But, for people doing serious CAD, CAE, or CAM work, the performance and reliability the system offers is worth the premium.

While at SolidWorks World, I also had a chance to chat with Rick Krause, CEO of BOXX Technologies. BOXX makes what could be considered the equivalent of a tuner-built hot rod. Their 3DBOXX 3970 XTREME workstation is designed to provide the best performance possible for serious 3D CAD work. That is, it’s performance isn’t tuned for doing spreadsheets and web browsing (which benefit from multiple core processors), it’s tuned for doing serious CAD work (which requires fewer, but faster cores.)

Let’s go back to the car analogy: Yenko Chevrolet sold stock Camaros, with the biggest and best engines GM offered. Baldwin Chevrolet sold hot-rodded Camaros, also with the biggest and best engines GM offered, but tuned to put out over 500+ horsepower (while still being streetable.)

The HP Z1 engineering workstations use Intel’s biggest and best processors. The BOXX XTREME workstations also use Intel’s biggest and best processors – tuned (overclocked) for the most horsepower.

BOXX doesn’t really like to use the work “overclock,” because it implies that they’re pushing the processor past it’s design spec. BOXX works closely with Intel, to make sure they stay within the processor design specs. Since they use liquid-cooling, they can push the processor faster, without reliability problems. Their workstations are backed-up by a 3 year warranty, and, in their history of selling overclocked systems, they’ve never experienced a processor failure.

If you’re a serious CAD, CAE, or CAM user, and you can out-run your current computer, you need to take a serious look at getting a factory-built or tuner-built hot rod computer.

HP Z1 Workstation

BOXX Technologies 3DBOXX 3970 XTREME Workstation

 Photo courtesy Baldwin-Motion

3D CAD Tips

What’s little known in the CAD industry, and rarely mentioned today, is that AutoCAD had an ancestor that predated the founding of Autodesk. A product called INTERACT.

This is a photo of the INTERACT CAD system, circa 1978. The hardware is an S-100 computer with dual 8″ floppy drives, and a 640×480 pixel graphics board.  Input is through a Houston Instruments digitizer and a Televideo terminal.

INTERACT was the first CAD system to run on mainstream microcomputer hardware. (Other contemporary systems ran on mainframes or minicomputers.) Its first commercial customer was Atlantic Richfield, which used the system to plan deep dives for offshore oil rigs.

INTERACT was written by Mike Riddle. He had previously worked on a ComputerVision CADDS3 system, which was used by his employer, Marathon Steel, to detail the structural steel used in the Palo Verde nuclear power plant, west of Phoenix. With the self assurance that many hackers have, Riddle figured he could do better than CADDS3.

He wrote INTERACT in his spare time, starting in 1977. He was slowed down by the state of hardware at the time — he had to write the program in pieces, and assemble it as larger memory boards became available. Ultimately, he decided he needed a processor that could support hardware multiply. Marinchip Systems, owned by John Walker and Dan Drake, made an S-100 main board with a TI TMS-9900 processor that fit the bill.

When Walker saw INTERACT running on the Marinchip Systems computer, he was impressed enough to become a dealer for the software.

In late 1981,Walker, Drake, Riddle, and about a dozen other people, came together to co-found what, in January 1982, would become Marinchip Software Partners, and shortly thereafter, Autodesk. INTERACT was rewritten in the C language, to run on the new IBM PC, and was rechristened–first as MicroCAD, and then (when the MicroCAD name was sniped by another company) as AutoCAD.

Today’s AutoCAD bears little resemblance to INTERACT.  Yet most of the original INTERACT commands still work in AutoCAD 2012.

3D CAD Tips

With around 1.7 million people using SolidWorks, it would be surprising if SolidWorks World wasn’t a large conference. Yet, the secret to its success is not just mere user count. It’s community building.

SolidWorks, the company, has for many years focused on building communities of interest among its customers. It does this more effectively, and more thoughtfully, than any of its competitors.

Consider user groups: SolidWorks provides model bylaws, funding, gifts, guest speakers, and just about any useful thing you could think of to help people start and run user groups. One need only visit www.swugn.org to see the level of support provided by the company. Richard Doyle, who runs SWUGN, even won the 2009 CAD Society Joe Greco Community Award.

Here’s a hint of how much value SolidWorks management places in their user groups: At the SolidWorks World SWUGN meeting, I watched both Jeff Ray and Bertrand Sicot (former and current CEOs of SolidWorks) taking audience questions, and answering them with complete candor. No filtering or blustering.

SolidWorks also spends a lot of effort on building its community of certified users. At SolidWorks world, you can tell the serious users by the CSWP (Certified SolidWorks Professional) ribbons they were wearing. And you can tell the really elite users by their CSWE (Certified SolidWorks Expert) ribbons. At this conference, SolidWorks hosted a reception for over 500 CSWP/CSWE users on the USS Midway aircraft carrier museum.

At SolidWorks World, there were also focused programs and events for resellers, educators, members of the press, tweeters, and even users from different geographic areas.

My sense is that one of SolidWorks’ biggest competitive advantages has been its focus on building and supporting communities of interest.

Dassault Systemes SolidWorks

http://www.solidworks.com

SolidWorks User Group Network

http://www.swugn.org

Photo courtesy Oleg Shilovitsky

3D CAD Tips