Providing geometric solvers (be it Siemens PLM or LEDAS) is cool but became not so profitable

23 years after D-Cubed has released the first ever commercial geometric constraint solver (2D DCM), its current owner Siemens PLM Software published, practically for the first time, more or less full statistics on this part of its business (see a link below).

D-Cubed is a company founded by an outstanding mathematician John Owen. It seems natural that this happened  in Cambridge - one of the recognized world centers of math foundations for CAD: note that ACIS and Parasolid have their roots also at Cambridge, and D-Cubed was initially somewhat invested by Spatial. In 2004 D-Cubed was acquired by UGS, and in 2007 – already as a part of UGS – sank into the great arms of Siemens PLM.

A key point of the above mentioned statistics is 200 - a total number of licenses sold since 1988. It includes five different and important technological components: 2D DCM, 3D DCM, PGM, and other. 99 from that 200 are the 2D DCM solvers.

LGS 2D and 3D from LEDAS appeared in the market in 2004, fifteen years after DCM. In total, during 7 years, LEDAS sold 30 licenses of its LGS. My main conclusion is that independently on any classification approach, such comparison of LGS with a child of John Owen and with a great Siemens is honorable for LEDAS:). 

Some of those who sell or resell products by millions copies would probably be laughing at 30 and 200. Yes, you may laugh but note that these 200 and 30 are the programs that to be designed and developed need extreme math qualification, hundreds of man-years to make them industrial, and they are sold not to end-users but mainly to sufficiently serious CAD/CAE/CAM vendors… BTW, the post with D-Cubed statistics mentions very many (millions) user seats for their 200 copies, it’s not clear how many of them are actually survived but historically it’s doubtless impressive: it’s enough to note that until recently these millions included millions of SolidWorks seats. Well anyway, I want to quote a nice sentence from the post about the milestone statistics of D-Cubed “Not bad for a range of products that many end-users will never have heard of”.     

    

One may reasonably ask: in the long run, income is much more important indicator than number of licenses. Maybe. And in some cases – definitely yes. I have no right to open some numbers but try to imagine how much Dassault Systemes, fifteen years ago, was paying to D-Cubed and then to UGS… Here is the key point: compared with 90^th when (1) D-Cubed was a monopolist and (2) big customers-vendors had no their own components such as geometric solvers,  today this has drastically changed, in the first place – in prices. Without opening some confidential data, I can just say that LEDAS has substantially contributed to this changes – both in (1) and (2). (Who knows, if around 2002, D-Cubed and LEDAS combined their efforts, they both could earn more:).       

What is my conclusion? Those who are able to develop an industrial technologic component of a DCM/LGS type are very probably smart guys. And yes, a vendor of such components justly gets some good additional points to the market rating of its company. Also, if such a vendor itself intensively applies its own technological components in its own mass products, this may become an important competitive advantage. My opinion is that distribution of such components today can hardly be a proportionally serious part of a successful business. However, in the long run, much depends on the sales skills.

On the other hand, those who succeeded in development of advanced components can likely be efficient providers of program development services – be they related to the maintenance for such components or to the development of whatever advanced programs in engineering software or beyond. But this is another story…  

Thanks to Ralph Grabowski, you can take a look at two much explaining publication:

AllAbout D-Cubed's 2D DCM and

LEDAS Responds to D-Cubed.

And of course look at “Major Milestones for the D-Cubed Components” which deservedly celebrates the above-mentioned 200-milestone and actually motivated me for this blog post. My respect and congratulations to D-Cubed!

Discussing LEDAS’ heavy technological karma...

Cadovod is one of the most noticeable and bright bloggers writing in Russian about CAD. (Cadovod in Russian sounds like a person who is cultivating CAD - similar to cultivating vegetables or something…).

Recently Cadovod published an interview with me (in Russian), which might be useful for those English readers who are interested in history, current positioning, and development trends of the LEDAS Company.

Among the questions, there are:

Why did you decide to found LEDAS?

How did cooperation with DS start?

Should CAD vendors build their own geometric solvers or should they license them from Siemens or LEDAS?

Does LEDAS plan to produce its own CAD?

Why did LEDAS start building CAD-oriented applications?

and other...

We at isicad decided to translate that interview and now you can read it in English at isicad.net.

LEDAS Responds to D-Cubed / Siemens PLM Software

For obvious reasons, I reproduce a guest editorial (by Dmitry Ushakov, director of product management at LEDAS, Novosibirsk Russia) just published by Ralph Grabowski in his u p F r o n t . e Z i n e, Issue #646. The only additional remark is that the technologies in question will be openly and in details discussed within one of the working groups in September at COFES-Russia seminar in Moscow.

"All About D-Cubed's 2D DCM" (in #645) was especially interesting, since the name of the company I work for (LEDAS) was mentioned. In answering upFront.eZine's question about competitors, D-Cubed's Dr Howarth said that "For the DCM geometric constraint solving components, LEDAS, based in Russia, is a relatively recent entrant". May I make a correction?

LEDAS Ltd. was established eleven years ago -- in 1999. Seven years ago, we began selling our technology component LGS 2D, a direct competitor of D-Cubed's 2D DCM. Our component is designed to support parametric sketching/drawing in 2D/3D CAD/CAM/CAE applications. Since then, a dozen software companies have licensed LGS 2D, as well as our 3D version, LGS 3D (our component that is a competitor of D-Cube's 3D DCM), and embedded them successfully into their applications. So, while we are ten years younger than D-Cubed, we are not new.

From our point of view, the primary difference between us is that D-Cubed is a part of a big vendor, Siemens PLM Systems; we are independent. Siemens PLM Systems competes in the CAD/CAM applications market against other companies who license D-Cubed's components. Do these companies receive the same level of technical support and functionality as do divisions inside Siemens? For example, are any competitors able to license Siemens' synchronous technology? This is not a question that our customers need to ask, for all of our technologies are completely available to them, because we do not compete with our customers.

We are aware of Dr. Owen (founder of D-Cubed) and his remarkable contribution to the field of geometric constraint solving. At the same time, it is worthwhile mentioning the contributions of other researchers. Among them are Prof. Hoffmann from Purdue University (USA), Prof. Michelucci from Université de Bourgogne (France), and Prof. Clement from Dassault Systemes (France), as well as other pioneer researchers whose work on solving geometric constraints go as far back as 1975.

Constraints Solved to the Users' Satisfaction
Whereas algorithmic issues have been elaborated enough during the last thirty years, there are just a few commercial geometric constraint solvers on the market. In my opinion, it is not enough to only solve the constraint satisfaction problem. What is also needed is a way to find a solution that best corresponds to the expectations of end-users -- this is called natural behavior, and sometimes depends on the subjective opinions of users. The maturity of a particular commercial technology for constraint solving is based on continuous improvement in the quality of its solving procedure. This process is driven by requests from customers who use the solver in different contexts.

LEDAS also does research in constraint solving. It seems that our team is somewhat similar in its scientific and technical background to the one Dr Howarth works with; in particular, about half of our developers have their Ph.D.s and many others have a good chance to receiving it. This brain power is necessary, because the development of CAD components combines strong mathematical, computer science, and software engineering know-how. We actively collaborate with universities, such as the ongoing research at Purdue University (USA) in the field of CAD user interfaces. Together, we are improving constraint-driven freehand sketching, which is the key to building robust sketch-aware systems and sketch-based interfaces for future CAD systems.

Some customers who license our technology are not only traditional CAD vendors (Tecnos G.A. and CD-adapco), but also firms who embed our LGS 3D component into custom CAD systems. Joe Gibbs Racing of NASCAR, for example, is using our software to assemble suspension parts onto the chassis. Other customers prefer to exploit our mathematically-skilled team to develop proprietary key components, such as for CATIA V5, which has been on the market now for several years.

Last year, we collaborated with Open Design Alliance, a non-profit consortium of over 2,000 software developers, to integrate our LGS 2D geometric constraint solver into their Teigha platform, and is fully compatible with the DWG 2010 file format. Now our solver is available to ODA members who require constraint support for their applications.

Constraint Technology for the Future
One of Dr. Howarth's interesting points is regarding the future direction for geometric constraint solvers. We share his vision of increasing the use of 3D solvers as the engine for a new generation of direct modelers. For us, this is an important field of research. As direct modelers become more popular, the question becomes how to keep design intent in a history-free environment. To answer the question, LEDAS is now developing our "variational direct modeling" (VDM) technology.

(VDM allows users to intelligently modify any parameter in a direct modeling system, while design intent is automatically recognized by our engine and expressed as a set of geometric and engineering constraints. We have developed a plug-in for McNeel & Associates' Rhinoceros direct modeling software, which allows anyone to evaluate our VDM technology.)

I trust that this addition to the Dr. Howarth interview is useful for readers of upFront.eZine, and I am available to answer further questions at ushakov@ledas.com."

[Dmitry Ushakov is director of product management at LEDAS, Novosibirsk Russia.]
http://www.ledas.com/

LEDAS Reproduces AutoCAD 2010 Parametric Drawing with ODA and Moves towards Variational Direct Modelling with Rhino

1. In June 2009, the Open Design Alliance (ODA) and LEDAS announced their cooperation to develop constraint support for the DWG 2010 format in DWGdirect™. In that announcement, ODA President, Arnold van der Weide, commented, “… We are excited to give our members the opportunity to integrate DWGdirect and LGS 2D (one of the LEDAS Geometric Solvers) together with minimal member development effort, and we are convinced that this work will benefit existing as well as potential ODA members.”

Now, in March 2010, ODA and LEDAS announced that the intended constraint support for the DWG 2010 has been implemented on the basis of the latest version 3.0 of the LEDAS LGS 2D. Dmitry Ushakov, Director of Product Management of LEDAS will present this result at the second ODA World Conference in Orlando (USA) , May 4-5. By this integration, ODA and LEDAS enable 2000+ companies (members of the Alliance) worldwide to implement parametric drawing functions in their DWG-compatible applications.

The ODA – LEDAS cooperation was initiated just after the first ODA World Conference in Leiden, Holland last year, where Dmitry Ushakov gave a talk entitled “Using LGS Geometric Solver to Create a DRX-module that Supports Geometric and Dimensional Constraints”, which caused significant interest of the conference attendees. This interest was heated by release of – just one month before the conference – AutoCAD 2010, a new generation of popular CAD software that finally supported parametric drawing functions (widely available before in almost all many MCAD software packages). LEDAS presented a module that implements similar functions on top of LGS 2D, a geometric constraint solver developed by LEDAS since 2004 and embedded in several industrial CAD packages. (See Dmitry’s detailed report on the first ODA World Conference).

At the second conference in Orlando Dmitry Ushakov will also present a free demo application that everyone can use to evaluate these parametric design tools. The application allows users to open and visualize parametric drawing files created in AutoCAD 2010, to edit their dimensional and geometric constraints (they are solved immediately after editing), and to save the files in DWG 2010 format enabling further editing these files in AutoCAD 2010. The attendees of the dedicated ODA Technology Exchange workshop will learn how they can embed the parametric drawing technology into their applications for different platforms.

2. In Orlando, Dmitry Ushakov will additionally present the current state of the LEDAS project related to variational direct modeling technology (VDM), which allows CAD users to edit solid models parametrically in history-free environment where design intent is expressed by explicit and implicit geometric constraints and driving dimensions.

VDM goes far beyond traditional direct modeling approach: this technology allows parametric modification of 3D geometric shapes in history-free environment that preserves design intent expressed with explicit and implicit geometric constraints. LEDAS considers VDM as a technology ideologically similar (but different from a technical point of view) to what is being implemented in Siemens PLM Software as synchronous technology. In contrast to synchronous technology (and similar things from other Big Vendors like Autodesk Inventor Fusion and PTC CoCreate), VDM is proposed for licensing to CAD developers. Based on company's resources and objectives, LEDAS is implementing its project in stages, each of which results in releasing an application – a plug-in for a popular market product. Every subsequent plug-in (or its version) accumulates a new portion of the technology, improves (CAD-independent) architecture, attracts potential customers of the LEDAS direct modeling, implements a market-demanded parametrization for the systems that lack this functionality, and of course brings LEDAS an increasing income.

The first plug-in (made for Google SketchUp which enabled LEDAS to implement and debug an initial architecture and a comparatively simple functionality) has been by today downloaded about 20 000 times.

A more serious system was in 2009 implemented for Rhinoceros, see LEDAS Adds Assemblies and Kinematics to Rhino (LEDAS moves McNeel & Associates’ Rhinoceros NURBS modeler closer to the mainstream MCAD with $395 Rhino Assembly 1.0 plug-in for assembling complex mechanisms and simulating kinematics).

http://www.drivingdimensions.com/ is a special LEDAS website devoted to present a series of the above mentioned plug-ins.

A couple of days ago, LEDAS released a new beta version of its RhinoDirect plug-in . Here is a video which demonstrates a full cycle of parametric hammer modeling in Rhinoceros – combining Rhino history with constraints.





3.The two projects mentioned above in this article are fundamentally based on the LEDAS Geometric Solvers, LGS 2D and LGS 3D , -- the market products that quite successfully compete (as you can see from CD-adapco licenses LEDAS variational geometric solver LGS 2D or Joe Gibbs Racing to Use LEDAS LGS 3D Geometric Solver) with D-Cubed componentsfrom Siemens PLM SOftware.

Some (or many) people might be surprised by such ambitious projects and plans of LEDAS, a company with 30 employees. This feeling may vary in the range from admiration to mistrust – especially if one could get information about some big outsourcing LEDAS projects that are naturally closed by strict NDAs. Well, in the Web one can even find some publications related to CAD technologies that compare LEDAS with SiemensJ.

No, LEDAS can in no means be compared with Siemens. But LEDAS can reasonably be compared with D-Cubed. D-Cubed is a remarkable UK company (founded in 1989 by an outstanding expert John Owen), which was first to release industrial geometric solvers and for a long time had the market monopoly. In 2004 D-Cubed was acquired by UGS, and in 2007 became wrapped even deeper when Siemens acquired UGS and transformed it into Siemens LM Software.

LEDAS and D-Cubed seem to have comparable academic history and background. See some information about LEDAS competence and history. LEDAS is operating in a world-famous research and university center, thus having inexhaustible opportunities to spoon out very high quality human resources in math and software development. Since the company foundation, these potential is effectively being implemented: today about a half of the LEDAS R&D staff are PhDs with average age around 28.

Of course LEDAS has no budget-like financing (as it happens inside of Very Big Vendors) but on 1st of April, the company is going to celebrate its 11th anniversary independent, flexible, implementing a number of high-competence-required internal and external projects, opened to multiple business contacts, and full of ideas.
.