Early in October a trial jury ended in the ongoing saga of Spatial versus Autodesk. This was a fiercely contested case that Autodesk ultimately won. MCAD industry analyst, writer, and trainer, Joe Greco, was on the scene for almost the entire trial and provides us with the blow by blow account of what transpired from beginning to end.
The ongoing lawsuit filed by Spatial Corporation against Autodesk came to an end on October 2, 2003 in a Marin County courtroom when a 12-member jury voted in favor of Autodesk, saying that it did not breach its 1991 contract with Spatial. I had the opportunity to sit in on the final eight days of the nine-day trail and here are some of my thoughts.
Spatial claimed that Autodesk breached the contract by disseminating the ACIS source code to parties that it was not allowed to and that it had also violated the contract terms that restricted the number of locations the code could be sent. Spatial was most disturbed by the fact that Autodesk, beginning sometime in the fall of 2001, had allegedly shared the ACIS source code with D-Cubed for the purpose of using it to as a foundation to create a new modeling kernel -- what would eventually become ShapeManager.
Spatial presented testimony and video depositions from many people, and perhaps the four key witnesses were Peter Charrot; a senior programmer for D-Cubed who did contract work for Spatial; David Kershaw another high-level programmer at D-Cubed contracted by Autodesk; Mike Hansen a former VP at Spatial; and Kathy Cunningham who was part of Spatial legal team that worked on the Autodesk/Spatial contract of 1991. Charrot admitted to having the ACIS source code on many of his computers at D-Cubed's headquarters in Cambridge, England, and admitted to making changes to the code -- changes that were supposed to be cleared with Spatial, but at least one time, he did so only after the fact.
Spatial had okayed D-Cubed and Kershaw to have access to the source code in order to work on a hidden line project called AHL for Autodesk, beginning in 1996. However, in his deposition, Kershaw admits to being taken off the AHL project in November of 2001 to begin work on ShapeManager, saying that the two projects "sort of overlapped." This timeframe is key because about the same the time it was beginning to be clear that Spatial and Autodesk couldn't work together anymore -- in fact Autodesk had just exercised its right to exercise a one-time buyout of ACIS. This relieved Autodesk of their yearly one million dollar license payout to Spatial, but not from its contractual obligations which included the non-dissemination of the source code.
I felt Mike Hansen was a strong Spatial witness who said that he would have been upset if he had known what Autodesk was doing with the source code. He knew and approved of D-Cubed using the ACIS source code to work on the AHL project, but when asked "did anyone say that D-Cubed would be working other stuff," his reply was "no". Kathy Cunningham, who left Spatial a long time ago, gave testimony that backed up what Spatial was declaring was the proper intent of the contract. In addition, the point was also made by several Spatial witnesses that in none of its other licensing agreements with companies such as HP, Bentley, CADAM, etc. did non-employees have access to source code.
Autodesk's best defense to this was a video deposition from Spatial co-founder Dick Sowar, who testified that his main goal was to make sure that Autodesk was able to use the ACIS source code to get their software products to market, which would result in both companies making money. While he was concerned about too many people getting code, as long as the proper agreements were in place, he felt that it was okay. Sowar also testified that he knew about the fact that Autodesk had the code in many locations and he didn't object. He also said Spatial also didn't object when Autodesk gave the source code to a company called GSSL to develop translation software.
Spatial's counter to this testimony was that while Sowar may have certain "feelings" to how things were done, he wasn't familiar with the actual details of the contract, and that the person who was, Kathy Cunningham, testified that the contract didn't permit this type of action.
Another strong witness for Autodesk was Roger Mollon, a product manager, who used to work for Spatial. When asked if it was Spatial's goal to restrict source code license he essentially concurred with Sowar he said, and I paraphrase, "no -- it was Spatial's goal to get ACIS in as many products as possible." Mollon also testified that it was Spatial that sent the source code to different Autodesk locations.
So here's what the trial essentially amounted to: was D-Cubed, a third-party contractor for Autodesk and, therefore, within its rights to work on Spatial's source code? That issue brought up numerous days of testimony and counter-testimony, all stemming around the wording of the 1991 contract between Autodesk and Spatial. This was the contract that replaced the original 1989 agreement and one of the key edits was the changing of a single word from "employee" to "individual". Autodesk felt that this change was needed because at the time it was difficult to hire programmers to work on code if they were considered employees, as this restricted their freedom and mobility. Autodesk feels that this change allowed them to hire D-Cubed to work on the ACIS source code, and points to the aforementioned AHL project, as proof that Spatial did not object. However, as mentioned earlier, the testimony from Spatial executives such as Mike Hansen stated that Spatial gave Autodesk special permission to share source code with D-Cubed only for the AHL project. Mollon, on the other hand, said this permission wasn't only for the AHL project.
In addition, there is one section in this 1991 contract that essentially says "in no event shall Autodesk have the right to distribute, sublicense or otherwise transfer to any third party the source code to the ACIS Modeler, unless the appropriate agreements are in place" and Spatial felt they weren't. Spatial points to the fact that no one at D-Cubed ever signed the standard Autodesk Employee Confidentially Agreement until early December 2001. This after all the trouble had started -- and after an internal emailed noted that Autodesk had put its defense team on "defcon 3" -- essentially Spatial was suggesting that Autodesk was expecting a lawsuit for its actions.
Autodesk insisted this "defcon 3" was a defensive position (and technically it is, as the US military definition means "defense condition") because of Spatial's aggressive actions during those months, discussed a few paragraphs below. This December 2001 date is also after Autodesk issued a press release issued on November 27, 2001 where Autodesk announced its intention to develop a new kernel with D-Cubed, built on top of ACIS. So if Autodesk felt it was within its rights to work with D-Cubed under the 1991 contract, then why did it have D-Cubed sign an Autodesk Employee Confidentially Agreement at all? And if it felt that maybe it was stretching the meaning of the 1991 contract and tried to cover itself with the signing of the confidentially agreement, it was done too late.
Several Autodesk executives testified that they were frequently not happy with the product or the support they were getting for their annual million dollar licensing fee. (Personally I found this odd - if Autodesk was so unhappy with ACIS, why was it was still used in all their 3D products, including non-CAD software like 3D Studio Max and also Inventor, which was billed as completely new technology when it was introduced in 1999.) Anyway, while Spatial wasn't happy with the deal either, feeling that a million a year wasn't enough considering all the products ACIS was used in, it admits that its product and services were not up to par at that time. It says that this was the "old Spatial," -- or the time before Dassault Systemes bought Spatial's component business, including the ACIS technology, in November 2000. Autodesk admits the product and support got better after the acquisition, but says they felt worried that a competitor was now in control of the kernel its software was based on.
Assault Systemes (stock symbol: NASTY)
Autodesk, via the testimony of CEO Carol Bartz claimed its fears were justified, as perhaps they were. She recounted an October 2001 meeting between themselves and Spatial, which also featured Spatial's CEO Mike Payne, where a representative from SolidWorks phoned into the meeting to suggest that Autodesk drop Inventor and focus on reselling SolidWorks. Autodesk also said that Spatial threatened not to share CATIA technology that was being planned for incorporation into ACIS with Autodesk and made other threats regarding how poorly it could treat Autodesk. According to Carl Bass, a VP at Autodesk, the combinations of these situations caused them to start working with D-Cubed around November of 2001, whereas before they were only thinking about it.