 |
 |
| Atlanta, Day Two: Tuesday, June 11, 2002
brought to you by PTC/USER and PTC with Sun Microsystems Today's topics:
The hotel said I could keep the room, so I'm still writing from the suite living room, 20x30 feet in size, outside corner room with full height glass windows on all sides (50 linear feet of glass windows). Plus a solid wood carved gilded oriental style peacock on top of the granite sideboard. Only element out of place is my rollaway bed in the middle. And the hotel is charging me $50 a night for the accommodations, perhaps mostly to make up for having sold my original reservation. So I have met the challenge people gave me of finding a $50 hotel room in Atlanta, wasn't that hard. Gillian Karamanlis, the Product Life Cycle solution business mgr. at HP, was back this year to report on HP's implementation of Windchill. But what makes them an apparent model of PDM progress isn't actually Windchill itself. Instead the team with which Gillian worked dedicated themselves to understanding their current process and the key elements in that process, before they brought in Windchill. That has to be the right way to advance in any product data mgmt. effort: understand your process and the key elements first. Undoubtedly PTC has customers asking to buy Windchill who may be problem cases, companies looking for Windchill to solve internal process issues just by itself. HP seems to have had good support from PTC, and you can figure PTC must have been glad to have a customer who did not see Windchill as some magic cure. A good example of the thoroughness of the HP approach is that they don't have production metrics yet for the Windchill implementation, which only started a few months ago. But they already have large cost savings just from the process work, the initial foundation for Windchill. They used the CMII process methodology, the Configuration Management Institute process framework (which Windchill also supports). But all this is general observations, and a key to the HP approach was to narrow down the scope. Not all of HP, but the Imaging and Printing Group. And not all of that group, but the Inkjet Printers division. And not all of that division, but the inkjet cartridges part. And not all of inkjet cartridges, but the Tooling BOMs---that's the BOMs and related specs and documentation for the assembly lines producing inkjet cartridges. HP's key interest here was making the same cartridge interchangeable worldwide, so that cartridges produced in Singapore and Ireland would fit the same. And there are very tight tolerances on the spray holes. Also, a new assembly line is being introduced every month, since HP is producing 120 million cartridges a year, and demand is increasing fast. HP could have concentrated on inspection of finished cartridges. But instead they went after the assembly lines, to get control over the BOMs and specs and other documentation that define the assembly lines. That's where Windchill and this project came in. Perhaps HP could have lost $750,000,000 on a failed over-ambitious enterprise wide PDM project, as other companies have done, according to newstories. But by starting small, but still in a strategic area (inkjet cartridge revenues are about $4 billion a year), they have a good foundation for more growth. After the Tooling BOMs, then the inkjet cartridge Product BOMs are next. They began with routine steps, such as capturing business requirements and developing a common process for all assembly sites. But just the common process took everyone involved 3 days of hard work, some compromises inevitable. They needed a formal issue resolution process throughout, so that everyone could feel they were heard, fairly, but also so the work could continue, with timely resolution of concerns. Just because of the CMII process work, they've seen a reduction in change order cycle times from 34 days to 18 days, using the same metrics. Which Gillian estimates is worth at least $10 million annually. Key to developing a Windchill tool to use with this process was a "conference room pilot", where key users were brought in for a 1/2 day each to try it out. There were a good number of comments, and adjustments to make, in process, management of change, training, and security, all gained from this pilot. Training in particular turned out to need more attention than the team had expected, and some security issues users found are still being worked on. About 300 people at HP are now using Windchill for inkjet cartridge assembly line Tooling BOMs, in the initial release, since Feb. of this year. In another few months, there should be a total of 1500 using it. As mentioned, inkjet cartridge Product BOMs will be next, with their related documentation too. Seems every PTC user conference is not complete without presentations on resolve problems, parts and assemblies. That's got to come with using parametric technology. But what's the future here, endless resolve mode failures, forever? Clearly they are an on-going problem. You can imagine that while still keeping the parametric database, PTC could deliver one or more versions of Pro/E that reduce resolve mode problems by reducing the chance to create them. Such a version of Pro/E wouldn't have all the capabilities that the full product has, but companies that see time being consumed by resolve mode problems might accept some less capabilities in return for more consistent results. Jeff Filo of PTC Global Services gave this presentation on resolve mode, filling an honorable place in the long series of user conference sessions on the same subject. He had a good point to make at the beginning: the more experienced you are with resolve mode, the less likely you are to need it. The more you see of the possible problems, the more aware you'll be of them, and of avoiding them.
This PTC session, by Natesh Gohil and Chad Jackson, covered new features in the Wildfire version of Pro/E. I've already mentioned some of the user interface changes, won't repeat those. But although I've seen some more remnants of the old menus here and there, it does seem now we have a consistent user interface, you can see a complete design intent there. Which wasn't so clear in the years it took to get here, from rev. 20.
Suppose you're getting a cup of coffee, and one of the guys turns to you and says that he thinks Pro/E is "the Queen of CAD Systems". You would probably wonder if he was off his rocker, to put it mildly. Even PTC would probably be sceptical of such an ornate compliment ("Queen of CAD Systems??! Who is this guy? What does he want?!"). But Queen of CAD Systems is a title given to Medusa, a 2D drafting system, at BMW in Germany, and they certainly know their CAD systems at BMW---auto manufacturers seem to accumulate one of every kind. And Medusa was selected by PTC as the basis, the software kernel, for a new product to enhance Pro/Detail. PTC was the very latest owner of Medusa, which was originally developed in the UK over 20 years ago. Then it was owned by Prime, then it was owned by Computervision who bought it from Prime, then it was owned by Prime again, when Prime bought CV, then it was owned by CV again, when Prime collapsed leaving CV behind, and then owned by PTC, when PTC bought CV. All these changes of fortunes have done nothing to affect the dedication and loyalty of Medusa users, who have to be about the most dedicated and loyal CAD users of any product. Hence the title, "Queen of CAD Systems", that seems to be just how they feel. Some part of the dedication and loyalty seems to be due to the initial concentration of Medusa developers on 2D drafting. AutoCAD for example didn't at all benefit from such attention to drafting users at the beginning: AutoCAD was just one of a planned series of office software products, and the developers only gave it more attention when they found it was more successful. So Medusa had the 2D drafting orientation right at the beginning, and for people doing 2D drafting (about 90% of the world's CAD users do 2D drafting, by most estimates) it seems to fall easily to hand, as a comfortable tool. Over a year ago, early 2001, PTC planned to use Medusa to fill a couple of gaps: (1) some PTC customers have up to 2,000,000 legacy drawings, which they want to maintain, and (2) customers were looking for a tool just to to finish work on Pro/E drawings, like adding textual information or adding symbols or adding related views. Then PTC decided that such a tool would be a distraction from 3D CAD, and sold Medusa (the last sale of Medusa) to CAD schroer, a German engineering and software company (mostly mech. engineers doing mech. eng. and also writing software). That seems to continue the Medusa tradition, it's software solidly based in mech. engineering. This product, based on Medusa, was originally called Pegasus, who sprang from Medusa's body when Perseus cut off her head, but on second thoughts it's been renamed Stheno, one of Medusa's sisters. Although right now the files still have a .peg suffix, that's going to be changed to .tsh (you don't get any choice of suffix you want). When you have Stheno installed, from Pro/E you do Applications > Stheno, while in a drawing, to get an exact copy of that drawing in Stheno, another application in another window. You can lock the Pro/E component of the Stheno drawing, so no changes can be made to it. Or you can allow standalone Stheno use of the drawing. If there are changes on the Pro/E side, you can startup Stheno again and update it from Pro/E. The Stheno user interface is supposed to be similar to Wildfire, but actually it's more iconic, double columns of icons on the left side. It doesn't seem to have the special graphics window features of the Wildfire interface. Stheno beta testing begins in 2 weeks, end of June, and initial release is planned for Aug. 28, on Windows and Solaris (thanks) and HP. The German developers adopted Java as a user programming language, just because they're concerned to make it really platform independent (so they didn't use Visual Basic). Since PTC did sell Medusa to CAD Schroer, if you're interested (perhaps you have 2,000,000 drawings yourself?) you'd probably want to contact them for any more info., or to participate in beta testing. You can email Mark Simpson, msimpson@cad-schroer.com. All credit to Ron Harkins at Enser, who persevered to give this presentation after two graphics board swaps and then a hard disk transplant into a second CPU, all done live on stage before a sympathetic audience. Drawing templates are a good example of user group and PTC cooperation. Many different enhancement requests were combined into a 1998 Drafting Tech. Committee white paper, to which PTC responded with a software mockup. And drawing templates were then introduced on 2000i2. With drawing templates you can have predefined views, preformatted text in different syles, all editable, also tables for special formats. The value is particularly great for the end-users reading these drawings, because now they get more consistency for standard drawings, the same information in the same places. To create a template, either you start with an existing drawing, or you create a new drawing from scratch. Most people will just start with an existing drawing, the Applications > Template. First view you define in the template will be one in the drawing. You do have to specify Model Display and Tanget Display, out of the list of options. Pro/E defaults to a drawing symbol, to represent a view in the template (no model displayed yet, just creating a template). But you can create view symbols of your own, identifying the type of view visually, and its relation to other views (2D symbols of a block with a chamfer and a hole, or a similar set of view symbols). After specifying the first view in the template, then hit New in the Template View window to creat a projected view off that view. You can use 4 different types of text in a drawing template to advantage:
With the drawing template, you can choose the primary view for dimensions, when a dimension could appear equally in 2 or more views. Ron has a simple program that concatenates part no. and both lines of the drawing title into one single longer parameter, and saves that in the drawing for Windchill searches, a Windchill search word. Controlling External Refs in a Top Down Design Scott Schultz from Portland gave this presentation. External references are just passing information from one model to another. That's what they do, that's their value, even if you don't want them in a particular case. Scott suggested that in early conceptual design, ext. refs. should be allowed freely. But he sees a major transition to a product database, where ext. refs. need to be cleaned up, and only a few perhaps retained. When passing info. with a Merge Feature, making changes in the target model can affect the origin model too, watch out. An Extended Copy Geom (new with 2001) lets you get information right out of one part into another part, direct---only dependency is on the source part. But a Standard Copy Geom includes in the dependecies all the links in the assembly structure between the source part and the current part---everything in that path, subassemblies and top level assembly, everything. Definitely avoid that Copy Geom if you want to manage dependencies. Tools for managing ext. refs:
Inheritance Features (new on 2001) are useful to get skeleton or other part geometry, and deal with it in the process, suppressing what you don't want for example, or resuming features, or changing dimensions. When you have hardware (like handles or hinges) or other separated components that have a similar function, Scott recommends putting them in a subassembly of their own, located off the basic assembly skeleton (some people call that an "installation"). He uses the same approach for cables and pipes and their associated hardware. Then he can make changes to a pipe, such as inserting or removing a valve, and the support hardware isn't affected. Great Smokies and Blue Ridge Parkway The day before the conference I stopped by Great Smoky Mountain National Park, on the border between N. Carolina and Tennessee. Since the highest point is less than 7000 ft., people from Colorado and California and Switzerland might wonder what's the big deal. But the Great Smokies were originally formed about 1 billion years ago. The Rockies and the Sierras and the Alps all under 50 million years, probably. So the real test of a mountain range is not just how it looks today, but how it looks after a billion years. Those modern mountains may not amount to much at all in 1,000,002,002, perhaps not even a bump. So the Great Smokies deserve some respect. Plus the Great Smokies are a very distinct separate terrain and environment. They are steep (even after a billion years), so mostly undeveloped and uninhabited, even outside the national park. They are the southern end of the Appalachian mountains, like an island to themselves. Rising out of the surrounding terrain, special flora and fauna everywhere. And the Blue Ridge Parkway starts from the Great Smokies and runs north to the Shenandoah valley in Virginia, over 400 miles. In the Chamber of Commerce tourist office in Asheville they have a 3D relief map of the area with the Blue Ridge Pkwy traced on it, which shows why this road is very unusual. Most roads encountering mountains go over a pass, or tunnel through, or just go up to the top and then back down. But the Blue Ridge Pkwy follows the actual ridge lines, for all that distance. This isn't an economical way to build any road, on top of the ridges between the valleys on either side. And the parkway was begun in the 1930's as a public works project. It's also unusual because it really is a parkway, no buildings or development or farms along the way, at least as far as I drove in N. Carolina. Worth a visit if you're ever in the area. Peter Nurkse |