return to q&a page

Interview first posted
8 January 2002

return to upFront.eZIne

q & a

five minutes with 
Applied 3D Science

 3D Applied Science
32 Daniel Webster Highway
Suite 14
Merrimack, NH 03054 USA

General Information:

info@a3ds.com

 www.a3ds.com  

Tel: 603-577-9660

Toll Free: 877-900-A3DS (2337)

Fax: 603-577-9669

Steve Mastrangelo <steve@a3ds.com > is the president and ceo of Applied 3D Science. In addition, his roles include running development and doing much of the product management for ParaLogix. [The company was upFront.eZine's first advertiser.]

upFront.eZine: Where did Applied 3D Science come from? What's the history of you guys?

Steve Mastrangelo: Applied 3D Science was founded by Jeff Hall and myself in late 1996 with an investment from Andor, a Japanese CAD company in dire need of their next generation product. I was the original architect of CADKEY, starting in 1985. Jeff Hall was vp of international sales at Cadkey in the late 80s.
     Our original goal was to market in Japan only through Andor, but we had to change direction when the Japanese economic situation affected our ability to get funding. We opened up an OEM business here, and have just begun selling the product direct.
     Andor reduced its share in A3DS due to its own change of direction, and I am now the majority stockholder. We may decide to establish a relationship in the future with another company, if it makes sense. As for our developers, many were with me right from the early days of Cadkey, and are some of the best in the business.

upFront.eZine: Jeff Hall says, "This product is light years ahead of both Inventor and SolidWorks in its architecture." Tell me why that is.

Steve Mastrangelo: There is a reason that "solid" modelers built on the same kernel tend to look-and-feel similar. We believe they took the path of least resistance provided by the modeling kernel. This is unfortunate for the user because while the developers of these kernels are bang-up mathematicians, they clearly are not CAD users.
     Autodesk took an interesting approach with Inventor and its "adaptive" stuff. Applied 3D Science took a fresh approach as well. ParaLogix and its procedural modeling methodology allows modeling relationships and dependencies to be established without explicit relationship definitions or 3D constraints.
     For example, to extrude a rung between the legs of a ladder, we just tell it to go to the next face, or to a face or to a plane when we made it. No switches required; no need for 3D constraints. 3D constraints are only necessary when bringing in reusable stock parts or subassemblies into the assembly model.
     In short, our procedural modeling takes feature-based modeling a step further because our procedures have _all_ of the following characteristics, whereas other modelers typically have one or two:

a. Create one or more features. For example, the New Component Extrude feature extrudes in either one or both directions with optional shells, drafts and fillets on the start and end.

b. Time sequence. You control the ordering of the procedures used to construct the model, resulting in different output geometry when the order is changed.

c. Operate across multiple components (parts). You can do assembly modeling with interdependencies and rarely, if at all, use a 3D constraint unless you are bringing in stock components. True top down assembly modeling, in other words.

I appreciate Jeff's comments, but we are a bit more modest. Our product has the potential to be light years ahead because of its core architecture -- it just isn't all exposed to the user yet. We want all of our customers to feel that they got 100% of what we advertise at the time they get the product, so we are intentionally low key in our claims at this time. There is one differentiating feature that we haven't been able to bring to production readiness just yet.

upFront.eZine: Once SolidWorks, EDS (Solid Edge), and Autodesk programmers see what you've done, are you worried they'll simply copy your concepts?

Steve Mastrangelo: Not at all. Their architecture limits their possibilities. If they did rewrite their architecture to incorporate our modeling concepts, so be it -- their move lends credibility to our ideas. It is true we are immediately compared to the products you mention -- what features they possess and we don't. We understand it is inevitable. Those three share the same niche; our niche is different: folks who are miles away from justifying a $10k investment for every seat.
     One concept they will not be able to touch is our ability to provide raw value and consumer-level prices. The CAD world needs a fresh business model. In other words, we won't engage in a feature war. We have a lean/mean company, stressing quality, not quantity, at a much lower total cost of ownership for the user.

upFront.eZine: Did you use competent architecture (D-cubed, ACIS, ParaSolid), or did you write everything from scratch?

Steve Mastrangelo: ParaLogix is build using the ParaSolid solids modeling kernel and the D-Cubed constraint manager. I am humored by the perception that building a good CAD system is a matter of simply packaging these; I wish it was that simple.
     Our implementation of ParaSolid is, however, somewhat unique. We defined what we wanted to do without the restriction of knowing how challenging it would be to accomplish.

upFront.eZine: In this era where file formats lock customers to a single vendor, why should anyone want to consider your product?

Steve Mastrangelo: This is a trap in the CAD world. Incremental increases in productivity do not sway these consumers; heck, significant increases often don't because real- world implementation is hard. It is time consuming. It takes good sound management practices that do not come bundled in a CAD product. And worse, very few CAD sales channels actually deliver much added value. The truth is that file formats don't lock customers to a single vendor as much as the perception that a change will not have a great enough ROI to justify the pain. Let me change your questions: "how does Applied 3D Science demonstrate enough value to justify the pain?"
     The best technology doesn't always win. For a little upstart like us to compete head-to-head with the name brands is wishful thinking. Our game plan is simple: we intend to provide unprecedented value within a profitable business model.

 Return to www.upfrontezine.com.

Entire contents copyright ©2002 by upFront.eZine Publishing, Ltd. All rights reserved worldwide. Article reprint fee $500. All trademarks belong to their respective holders. "upFront.eZine," "Talking About CAD," and "On your desktop every Tuesday morning" are trademarks of upFront.eZinePublishing, Ltd. Letters to the editor may be reproduced in an edited form for clarity and brevity. Opinions expressed in letters are not necessarily shared by upFront.eZine Publishing, Ltd.