"I have some different perspectives to offer on the issue of poor sales for AutoCAD 2000. Blaming the lack of add-on availability is a plausible but insufficient excuse.
"While it is true that in Europe most seats of AutoCAD are sold with a third-party application, in America -- which still represents a substantial portion of the market -- it is extremely common to purchase AutoCAD without any add-ons, to be used straight out of the box with no customization whatsoever. So, if it's not the availability of third party applications, then what's the problem?
"Suppose you had a two-year-old car that performed well, and that got
you where you needed to go in a fairly cost-effective way. Why would you
go buy a new one? That's the current situation for the installed R14 base.
Many work for Dilbert's boss, and are lucky even to be using R14 instead
of R13, R12, or a T-square! Thus we have millions of users who are completely
satisfied to continue with the status quo."
-- Steve Wells
"You will notice that there is no Building Services (formally of Softdesk). The electrical, mechanical, and plumbing engineers have no software that will move to AutoCAD 2000. What are they supposed to do?
"Autodesk is incredibly stupid for not syncing up releases of LT and
the main product. By not keeping LT at the same rev, it causes people to
not upgrade the main installation of AutoCAD for the drafters when all
of the minor users are using LT."
-- John Rutkowski
"Upgrading can be more of a hassle than a blessing. AutoCAD R14, combined with AutoCAD LT 98, IntelliCAD, FelixCAD, and Bellingham CAD (I don't know the current name [Vdraft - Ed.]), brought some good stability to most AutoCAD users. Third-party software (Autodesk's and others') is running pretty well.
"The apple cart doesn't need artificial upsetting at this point when the economy is pretty good, the computers are pretty good, and R14 is getting the job done. The number of bugs in R14 is small, and many of us have vivid memories of the R13 disaster.
"AutoCAD 2000 offers some interesting features, but none that the market
can't live without for a while. Autodesk will eventually get their upgrade
revenue, but not quickly enough for the impatient Wall Street crowd."
-- Richard Price
"Don't forget the impact of so many users going with Autodesk's VIP
program. That pretty much negates any new revenues for AutoCAD 2000."
-- Dave Stein
"Your sentiments as to why Autodesk is experiencing their current situation are exactly what I've been telling our people. If it weren't for the fact that we are on the VIP program, I wouldn't be upgrading anytime soon. And just because I've received it doesn't mean we are installing it. If there is one A2K add-on that doesn't work correctly, we won't use any A2K add-on -- or A2K at all.
"We can't afford to. The translation between versions, especially now
with objects and other technology, is more difficult than in times past,
and is not for the inexperienced CAD operator to handle."
-- Jeff Foster
"Upgrade to AutoCAD 2000 when we're just getting to know R14? Not likely.
Autodesk should not try to milk the cow so often. We may just get 'udderly'
fed up."
-- Brian Grishaber
"Decisions such as where to get technology have more to do with each company's requirements and capabilities than with their skill with a crystal ball. There is no right answer that fits all, even for those people who like to bash Autodesk.
"But there are two things that our beloved computing industry has proven over and over: (1) everything will change, given enough time; and (2) any solution based on an open systems architecture (freedom of choice for applications and utilities, plug and play, own your own data, faster availability of new technology from third party vendors clearly, blah, blah, blah) will always be the better choice in the end.
"For example: Apollo vs. Sun in workstations; Wang vs. Microsoft in
word processing; Microsoft vs. anyone in PC operating systems; MAC vs.
PCs in general; etc., etc. You can find negative cases to support any argument,
but market dynamics themselves have identified the open systems approach
as the correct strategy, at least as far as I can tell."
-- David Prawel
Since it was close to noon, I asked if he'd like to go for lunch. "Coffee is what I need." He had gotten up at 4 a.m. in San Raphael CA to catch the plane north from San Francisco airport.
We drive to Club Villa for lunch, the closest my town of Abbotsford offers for a California-style dining experience. John gets his coffee -- three times over -- to go with a turkey club sandwich; I have the chicken penne.
Over lunch, we discuss the state of the CAD industry. Both of us have watched AutoCAD flourish on the desktop computer, him for 15 years, me for 13. He began with AutoCOGO, one of the earliest add-ons to AutoCAD to use AutoLISP. More recently, he was the chief technical officer at Autodesk. A couple of years ago, John formed his own company, Design Variations.
Both of us have watched CAD change from community to big business, from pleasing the user to pleasing the shareholder, from companies run by programmers to companies run by, well, people who run companies. He watched from the inside; I watched from the outside.
I think that AutoCAD's openness (customization and programmability) made it a success; John insists the trivial ability to create a block (symbol) make AutoCAD successful. He is unconvinced that object-oriented CAD is going to work (I, too, think objects have many hurdles to overcome). "For example," he asks, "Did you know that before a user can create his own custom object, he has to know how to program?" Today's object-oriented CAD packages come with a limited number of objects. One of them, he says, has a single fireplace; another comes with just 38 doors -- not a lot, when you start to add up all the different kinds of doors. "Once you consider that there are a nearly infinite number of object types in such a system and therefore an infinite number of possible design relationships, you start to understand the intractable problems of the object-based approach."
He picks my Palm III: to design this with object-oriented mech CAD software, you might need several hundred objects; to design a skyscraper takes tens of thousands of different objects -- and nobody is ever going to create all those objects. Traditional CAD is too hard for architects; object CAD isn't suitable for AEC. He calls his solution "feature-based CAD." He is taking the best aspects of three software packages: Pro/E's feature modeling, dimension-driven editing, and associativity; Visio's ease of use, generic approach to diagramming, and simple user model; and AutoCAD's blocks, as well as its general purpose, multi discipline approach.
Back at my office, we step carefully through the plaster and sawdust littering the floor from the renovations. John sets up his notebook computer and hopes the pre-alpha software won't crash. "I recompiled it on the airplane this morning." It does crash -- three times. He shows me how his software works with just six modeling commands: Profile, Extrude, Cut, Offset, Taper, and Place. Typically, you draw a 2D component, then extrude it to make it 3D, then apply other editing commands. John says this mimics Pro/E's feature-based modeling. All drawing is done in flat-shaded mode.
Components in the drawing have traditional grips (called "positioners") and "line grips" that orient the component. Components typically snap into place; when not snapped, a dimension line appears to let you type a dimension. To draw in 3D space, touch a wall or a roof, and the "UCS" is automatically oriented. Changes are automatically tracked and propagated; 2D sections and views are automatically generated.
A few days later, John emails me: "It is quite clear from our conversation that the term 'feature-based modeling' will not mean much in the AEC market. This term was borrowed from the MCAD market where it is a well-established paradigm. We probably need to spend more time emphasizing our product's ability to create and capture user-defined design relationships, rather than feature modeling. This came from your example of the light switch relationship with the door. This is not supported in an object-based system unless the programmer thought about it for you, and designed a light switch object that knows about doors."
John sees two groups of people that'll be interested in his product: (1) those who are looking for "something better than paper; and (2) those who are thinking "The world is going 3D, but I'm going to wait for the herd to decide." This product is not meant to replace AutoCAD, nor run on top of AutoCAD. Instead, it is meant to be used along side AutoCAD. Beta versions of the software are expected to be available later this year.
Info on the upcoming I-DEAS Customer CoOperative Network (ICCON) conference is at http://www.iccon.org/events/iccon99/
Framework Technologies
ActiveProject 5.0 is the latest release of Web-based project communication
software for facilities,construction, and product development teams. Enhancements
include live Web conferencing; publish a multi-source, project newspage;
customized tracked communications; and new Web site templates. ActiveProject
5.0 is due to ship May 24; 15-seat bundle is US$16,495.
Generic CADD v6.1.5
"I have NEW shrink-wrapped (never opened!) retail copies of Generic
Cadd 6.1.5 (last version!) for sale for US$125 (incl. priority mail shipping
inside USA); US$100 ea. additional copies. Package includes 3.5" disks,
all manuals including reference guide, user's guide, and programming reference,
and command quick reference card. Send a check or money order in US funds
made out to Robert Cowan to PO Box 991, Gonzalez FL 2560-0991.
Intel
It used to be an increase of 50MHz in CPU speed would double the speed
of our computer. Now, it represents a mere 10% increase. For that reason,
headlines like "Intel's speed demon chip keeps roaring" become a yawner.
BTW, Intel is now shipping the 550MHz version of its Pentium III CPU.
Silicon Graphics
Linux will be added to SGI list of supported operating system, including
its own Irix, Unicos supercomputer OS, and Windows NT. "No single operating
system can meet the needs of all customers," said John Vrolyk of SGI.
Unigraphics Solutions
Parasolid Version 11, the geometric
modeling toolkit, is now shipping. New features include: new parallel algorithms
for symmetric multi-processor (SMP) workstations; advances in core geometric
modeling operations; and a new Parasolid Web Update Service program.
ZoomOn
The Swedish Web CAD company ZoomOn AB became Sweden's most promising
IT-startup company in a competition called "e-challenge".
ZoomOn won because of its strong business plan and most interesting future.
http://www.paug.org
PAUG
The Web site for Professional AutoCAD Users Group, located in Denver
CO USA.
http://www.softcover.com
scan2cad
It converts bitmaps to raster in DXF format.
-- Enrique Juaristi