Day 2: Tuesday, June 23

Last year's newsletters probably had more Pro/E technical substance and content, more Pro/E beef. The theme this year is probably more PTC futures and data management, less of the Pro/E technical beef. Why?

Seems that times are changing, at user conferences as well as at PTC. Pro/E after 20 versions has probably become a mature software package, much like other mature software packages---changes are less frequent, witness the gap between R18 and R20 (R19 just a sort of R18 point release). The major PTC emphasis is more on improving reliability and ease of use, rather than adding more functionality. And PTC's interests as a company are extending beyond product design to data management.

Interesting enough, seems PTC users are going through the same transition simultaneously. 5 of the top 10 user enhancements compiled by Pro/User immediately before the conference were in the area of Data Management (all Pro/Intralink, who says nobody's using Intralink). And only 2 of the top 10 were in Modeling. That's probably the reverse of the ratio a few years ago, when Modeling would have dominated over Data Mgmt.

So PTC, the company, and its customers do seem in some accord, that's good. In a few years, PTC projects to get most of its revenues from data management products, and perhaps by then data mgmt. will dominate the user group agendas (not yet now). But the change away from a Pro/E focus to a general data management focus will probably be stressful at times, for everyone, and we'll probably all look back to the good old days when things were simpler, when the conference menu was just that good old familiar Pro/E beef.

What's Windchill? Sun just bought $1 million worth of Windchill, and as is generally known Sun has about 50 Pro/E licenses, so perhaps you too should budget for $20k of Windchill for each of your Pro/E licenses? We may have heard that Windchill is based on Java, but there's so much hype about Java, humph, there's Sun again. What are we supposed to think?

But, perhaps the basic question is, what problem is Windchill supposed to solve? There's plenty of words, but I haven't seen a description of a real world situation that is the basis of Windchill. If we can just get the problem clear in our minds, then we might know how relevant Windchill is to us.

How about this as a problem description. Stab in the dark, take shots at it if you like, but hopefully better than nothing. This description might apply to you, whether you're a one-man outfit or a multi-billion dollar company:

• your Industrial Design is contracted out to a group in San Francisco. You want to get their every idea, every whim and flight of fancy, but you don't want them to see all your work, because some of it is already going in a different (secret) direction.
• your Finite Element Analysis consultant in located in the Cayman Islands. You want him to have access to some geometry, but only certain specific geometry in a particular format.
• your sheetmetal shop is an outfit in Singapore that only accepts 2D CAD geometry (you put up with that, because this shop is used by all the major computer manufacturers, so they must be good). So you're stuck with that 2D format requirement, it had better be easy.
• your favorite PCB board designer moved recently to Topeka and he took his workstation and his Allegro license with him. You want to automate processing and tracking the transfer of information between your Pro/E layouts and his Allegro board designs, in both directions.
• your board fab house is right off the freeway in an industrial area of Santa Clara, Calif. You want to intercept the photoplot files between your designer and the fab house, because (very reasonably) you do want to do a final check against your design before you commit to making boards.
• your plastics house is in England, and they have one license of Pro/E, but the tool makers who actually will make the molds are scattered around Scotland, and they all use Unigraphics. Some areas of your design are highly sculptured surfaces, and you want to check them against the Unigraphics model before they cut metal.

• everybody uses a different computer and different OS: MacOS (that's the industrial designers), NT (that's the Singapore sheet metal people), Linux (the FEA consultant, he writes all his own software, that's why he's so good), Solaris (the board designer), HP/UX (the mold makers, they are partial to UG on HP), and so forth.

Now this is a kind of virtual enterprise. Not the kind of traditional enterprise where everyone is under one roof or at least gets paychecks with one company name. But a virtual enterprise, scattered around the world, and where the people have entirely different software and hardware and loyalties. That was a key phrase in Jim Heppelmann's talk Monday morning, the virtual enterprise, his description of the reason for Windchill---and he should know, since he founded the company.

Just getting the physical network links between all parts of your virtual enterprise is manageable, not impossible. You'd expect your partners would all have some network access these days. Some links would be faster, some slower, but that's reasonable.

But the really huge big problem is managing all your product data as it travels everywhere. Who has access to what data? Where is it stored at each location? At what revision level? How does what someone is using correspond to what they got from you? What do you have now? When did you get it? Did you check for latest changes? When are you going to be finished? What's the cost look like now?

If you tried to do all that data management yourself throughout a virtual enterprise, you might go crazy, even if you're a multi-billion dollar company. That's a problem for Windchill, product data management in a virtual enterprise, managing the flow of product data between partners and suppliers with every kind of computer and OS and application software. That's the real world, and in the real world not everyone runs Pro/E software on Sun hardware.

Why Windchill as a solution to managing product data in a virtual enterprise? Why not Sherpa, or Metaphase, or PTC's own Optegra product? Well, these older PDM packages were designed for the more traditional kind of enterprise, where some organisation (call it MIS) had some kind of control over all the software and hardware. So that, for example, you could hope to upgrade all your PDM users to a new version over a weekend.

But in a virtual enterprise nobody has that absolute control over the entire enterprise, because the enterprise is composed of totally different companies. Just the idea of upgrading everyone to a new rev. of a shared application is difficult to imagine.

But with Java technology you can actually upgrade every user of some application of yours, because they can download the code when they start up the application. And you don't have to think what O/S and what version they're running, as long as they can use Java.

Down at the bottom of Windchill there will still be some database servers, like Oracle. Windchill doesn't attempt to replace Oracle, Oracle or its equivalent will still store the actual data. But using Java and Web technology, Windchill should be able to supply an entire array of tools for controlling and tracking product data throughout a scattered virtual enterprise (Dick Harrison mentioned that he expected to see 30-40 Windchill modules eventually).

So who would buy Windchill now? Should you budget $20k for each Pro/E user in your company? But all you can buy right now is a Windchill foundation and toolkit, the basis for developing further Windchill applications. How long until you can buy Windchill applications for a virtual enterprise? PTC sales people say 18 mos. to 3 years, while an international CAD/CAM consultant with 30 years of experience says 4 years. Applications are notoriously more difficult to develop than foundation software or toolkits, because applications are where the real world intrudes.

Is Windchill going to absorb Pro/Intralink? Jim Heppelmann, the Windchill founder, certainly said so Monday morning---he showed a slide where Windchill and Intralink gradually converged, and then after Intralink 2.0 and Windchill 3.0 Intralink simply disappeared, and only Windchill remained.

Dick Harrison had a different view Monday afternoon, he said Windchill would call Intralink, and that Windchill was complementary to Intralink, but Intralink would continue to be separate, as a data management system specific to Pro/E parts and drawings and assemblies.

Other PTC staff tended to follow Dick's lead, although at least one person simply refused to answer the question when asked in a session.

Sure seems Jim must have just been yielding to some of that Windchill enthusiasm, because over time Windchill might interface with literally dozens of application specific data management systems similar to Pro/Intralink. Systems for mechanical data, electrical data, perhaps other data too. If Windchill incorporates those application specific systems, or even if just some of them, it'd probably become more difficult to develop and maintain. So Intralink looks safe. If Intralink satisfies all your needs, you should not need to budget for Windchill.

John Stuart, PTC marketing VP, said that Computervision didn't have any Pro/E links in Windchill, "because the last thing CV wanted was to support Pro/E".

Now in fact CV was careful to set up Windchill to be entirely application neutral, no more supportive of Pro/E than of CADDS, CV's own product. And CV was glad to offer a Pro/E module for Optegra, CV's own PDM system.

CV's attitude to Windchill and to Optegra was that of a PDM vendor. Companies know that PDM systems may need to accomodate a variety of CAD systems, product data can come from any direction. So companies buying PDM systems will probably be more comfortable with a vendor which doesn't favor one CAD tool over another.

Now imagine this interview with a PTC senior executive in June, 2003, just five years into the future from from now:

"You're asking about Pro/Engineer? Yes, we have a wide selection of mechanical design software for our different customers, CADDS5 and PT/Modeler and DesignWave and Medusa and also Pro/Engineer. But our key focus at PTC now and the source of our amazing growth recently is a different kind of software, product data management software for the virtual enterprise. As a PDM supplier, we are proud to be associated with such other leading mechanical design software suppliers as SDRC, Unigraphics, and (cough, cough) CATIA. That is why the name Parametric Technology Corp. has for some time failed to describe our main products and our mission, and why we have now changed our name to PDM Technology Corp. Still PTC, but a new PTC, no longer limited by our origins in mechanical design."

Pure science fiction, you say? But unless PTC manages some such cultural change over the next 5 years, it may lose the opportunities it has now with Windchill, someone else will fill the gap. It took Computervision about 10 years to make the basic change from a CAD vendor to a PDM vendor, gradually, from about 1983 to about 1993. But change is accelerating, PTC probably doesn't have 10 years to make the same change, 5 years may be more like it.

• safe to use the Pre-Production R20, it's supported. Production releases will normally follow 4 to 6 months after the Pre-Production release, length of time variable depending on customer feedback.
• PTC has a direct translator, geometry only (no parameters) from CADDS5 R7 and R8 to Pro/E. Further improvements will depend on development of an in-house "topology bus", that will allow exchange of (at first) geometry and surfaces and (later) parameters between PTC applications.
• once again the question was asked in the data mgt. steering committee whether there were any production Pro/Intralink users present, and once again there were none
• there were major problems at PTC last fall, when the business computer systems converted over to an Oracle database. There were problems getting licenses issued and problems getting R19 out the door. About 1600 customer requests for R19 were lost for several months, and the entire R19 release process took an unprecedented 6 months. Normally in the future 85% of worldwide customers should get a new release within 2 months of the release date.
• Windchill won't kill Pro/Intralink, instead Intralink will live on, but there will be a merging of architectures, Intralink will acquire features from Windchill (Jim Vaughan).
• the $500 special for Intralink upgrades has been extended another time until the end of the quarter. Intralink will eventually be part of Windchill modules (John Stuart).
• PTC is already distributing a Computer Based Training CD to everyone who registers for a Basic class, to prepare them in advance for working with Pro/E. Next there will be a CBT CD on R20, in August. More CDs for training are in development.
• PTC has established a software usability lab in Waltham, managed by a usability group, under Technical Marketing. There will be more usability improvements in R21 and R22.
• Pro/PDM "will live on indefinitely", but with no new functionality and no new platforms supported (you'll never be able to run PDM Server on NT)
• the PTC Curriculum Dept. has developed a Professional Evaluator self-test of Pro/E proficiency, which will be released soon, and available to anyone
• there's a new supplementary grading system for bugs. Absolute highest priority is TOP. If your bug is graded TOP by Services, PTC Development has to answer within 2 days, have a software solution ready for testing within 5 days, and you should get a new release with the fix within 2 to 3 weeks after you reported the bug. Next level of priority is HIGH, but then Development has 20 days to answer (10x longer), and they have 40 days to provide a fix (8x longer). Other priority levels are MEDIUM and LOW (no times for those levels given). PTC Services engineers are agreeing they have never seen Development so responsive.
• DesignWave price will increase from $1995 to $3495 on July 6.
• Work is being done to make shading multi-threaded on NT, but it has been multi-threaded on Sun ever since R17.
• PTC has software development centers in Salt Lake City (old CDRS), San Jose (old Rasna), Waltham, Bedford (old CV), UK (user interface toolkit), Israel (2 centers, one did Pro/FlyThrough), and India (about 200 programmers there, CV site). PTC is committed to support and enhance CV CADDS5. PTC is also supporting Medusa through two outside contractors, no PTC developers involved. PTC has used CADDS5 programmers on Pro/E, and Optegra programmers on Intralink, and Pro/E programmers on CADDS (for ISD, Interactive Surface Design, where CV had an on-going project).
• all training is now being given on R20, even if it is Pre-Production software. The training group was re-writing the Basic and Advanced classes anyway, to emphasize an overall design approach instead of just menu picks, so they did it for R20.
• PTC has been working for 2 years systematically to improve handling of customer hotline calls, including a few million $ on the phone system. 80% of all calls from anywhere in the world should be answered by a live engineer (how long you wait not clear). In Feb. 1999 Technical Support should be ISO 9000 certified.
• PTC is now tracking software problems internally with a MTBF (Mean Time Between Failures) metric.

There was a meeting at the conference of ex-CV users, people once using CV but now on Pro/E. The main subject of the meeting was how CV user meetings might be scheduled with Pro/User, no decision there yet. But when asked what were particular CV strengths in comparison to Pro/E, this is what these former CV users said:

• geometry creation: fellow using Pro/E for some years still misses ease of using solid booleans for certain operations
• Design View: this was the PC parametric sketcher that CV bought and later incorporated into CADDS
• Medusa sheet metal: still the best sheet metal package around, was the observation
• detailing: the CADDS detailing package was developed under the direction of a former drafting room manager, and showed attention to details of drawing production
• CVNC 3 axis programming
• nurbs surfacing: a complex package, but substantially more capable than Pro/Surface, more like Pro/Designer
• large scale assembly management: CV had parametric solid geometry, but an explicit assembly model. Seems Pro/E has had more problems with large assemblies just because it uses a parametric assembly model.
• rounds: the former CV benchmark team was usually happy to see a benchmark that included complex rounds, because they found that was a CV strength against Pro/E

Probably unrealistic to expect to see booleans in Pro/Engineer, but otherwise each item on this list could be worth some attention. Especially large assembly management, since that's probably the main reason why some big companies with big assemblies chose CV over Pro/E in benchmark tests.

John Seeley Brown, director of Xerox's Palo Alto Research Center (PARC, where a whole lot of current computer interface technology was developed over 20 years ago), has said that complexity is easy, it's always easy to add products or to add features. Then he said that simplicity is much more difficult, and that creating simplicity has usually much more value than adding complexity.

Seems all companies go through cycles of adding complexity and then pulling back to simplify things. Every company eventually reaches a limit in the complexity of its product line, and has to pull back, even if it's just so management and employees can understand the product line. With 4 PDM products and 5 CAD products in the current product line, might be time for PTC to simplify. Even not counting the 80 modules in Pro/E.

Another company that went through a change from complexity back to simplicity was Sun Microsystems, back in 1989. Sun launched products on 3 different platforms simultaneously (Intel, Motorola, and Sparc), and soon after announced the first (and only) quarterly loss in its history. And then Sun abandoned the Intel and Motorola product lines, to simplify and concentrate on Sparc, and that decision certainly worked out pretty well for Sun.