At DEVELOP3D LIVE in March Dr Jonathan Ingram gave a fascinating talk on his work in the development of early CAD systems which were the forerunners to the BIM systems of today
At this year’s excellent BIM Show Live event in Newcastle, AEC Magazine bumped into someone we hadn’t seen for about 20 years – Dr Jonathan Ingram, the creator of the first BIM tools (Sonata and Reflex (and Pro Reflex), had come to his first BIM event in roughly the same length of time.
Dr Ingram is writing a book on modelling development and usage from the early 3D systems and gave a fascinating talk on the evolution and application of intelligent 3D systems for the built environment in the 70s, 80s, 90s and 2000s.
We were so taken by this talk that we asked him to speak at DEVELOP3D LIVE, a conference organised by our sister publication He agreed and now his presentation is available to view online here.
Intelligent AEC modelling really started with BDS – the Building Design System. While this was a 2D system, it featured multi-representational objects, where each ‘object’ would display a different graphic if it were in plan, section or elevation.
From this experience in the early 80s, Dr Ingram set about to create a true 3D modelling environment, which culminated in T squared’s Sonata application which ran on Apollo workstations and is now recognised as being the first true BIM system. Despite the incredible cost of both hardware and software (over £100,000), many global practices took on seats and it really started to change the way large practices thought about modelling and Computer Aided Design. Sonata was eventually sold to Alias.
Dr Ingram’s next application was called Reflex and took the concepts defined in Sonata and delivered them on the then new range of Silicon Graphics IRIX Workstations, the Pro Reflex running on desktop PCs. Eventually this was sold to mechanical CAD specialist PTC, which failed to enter the AEC market. Two of PTC’s programmers, Leonid Raiz and Dr. Irwin Jungreis left the company and took with them a development licence of Pro Reflex. Pro Reflex plus others eventually inspired them to developed Revit, which was eventually sold to Autodesk. Global domination ensued.
In the summer of 2016, Dr Jonathan Ingram was awarded the Royal Academy of Engineering Prince Philip Medal for his work developing systems which led to BIM, by HRH The Duke of Kent at the Royal Academy of Engineering’s Awards Dinner. The award is given out biannually to those who have given exceptional contributions to engineering through practice, management or education.
In response to this article, Dr. Irwin Jungreis, one of the founders of Revit, submitted a letter to AEC Magazine. It is printed in full below.
Dear Editor: In your announcement of Dr Jonathan Ingram’s March 28 presentation at DEVELOP3D LIVE you referred to him as the man behind the foundations of Revit, a viewpoint expanded upon by Dr. Ingram in his presentation (tinyurl.com/ History-BIM). I thought your readers might be interested in the real story connecting Revit and Dr. Ingram’s AEC design product, Reflex.
In 1997, Leonid Raiz and I founded Revit after having been key developers of several Mechanical CAD systems, most notably PTC’s Pro/Engineer. Our goal was to bring many of the innovations that PTC had spearheaded in the MCAD market to the AEC market.
Soon after leaving, we agreed to provide PTC consulting services in exchange for a non-exclusive development license to Reflex, which PTC had recently acquired. Although we were planning to develop our software from the ground up, we thought this license would be useful for three reasons. First, it might help us in negotiating with venture capitalists; second, it would protect us if PTC asserted we were violating our non-compete agreements; and finally, we thought there might be some low level utilities we could extract from the code.
After receiving several hours of instruction in the software architecture of Reflex from Reflex developers, we decided not to use it as our starting point because of several important differences at the very foundations of the software. At that point, we put it aside and never looked at it again.
These foundational differences had their roots in our product vision. Prior to writing a single line of code we spent considerable time interviewing architects and studying their workflow. We wanted our users to express their design intent graphically, because that was their natural language and because that allowed for a much greater variety of designs than any dialogue boxes could. We did not want our users to have to do any programming, so they would need a graphical way to specify not just a static design, but also the relationships between elements, so the software could keep the design refcoordinated through continuing design modifications.
Reflex components are created using a specialised programming language (VEL), whereas Revit components are created in a graphical “family editor”. What’s more, some of the relationships between Reflex components are programmed by the user into the component definition, whereas no such coding would be possible in the Revit family editor. Instead, Revit handles relationships through a centralised change propagation engine that gathers all the elements to be changed, sorts them in a context- dependent way, and then executes all the changes.
Our decision not to use Reflex as a foundation was probably influenced by our earlier experience at PTC. Like Revit, Pro/Engineer had also been built from the ground up. We had found that maintaining bi-directional associativity between components, views, and annotations required all code to be tightly interconnected, so we thought that starting from a system with a different foundation would be more difficult than starting from a clean sheet. It would have been like building the Freedom Tower on the foundation for the Empire State Building.
During Dr. Ingram’s presentation, he speculated that Revit was based on Reflex. As the dozens of programmers and architects who created Revit could testify, that is not the case: Revit was not based on Reflex. No code from Reflex was used, and we were less familiar with Reflex than with many other architectural design packages, such as ArchiCAD. In fact, we found the most inspiration from a $50 product called 3D Home Architect (a home version of Chief Architect), which had a clean, simple user interface. No design decisions in Revit were influenced by our modest familiarity with Reflex, and any similarities between the two were either well known ideas in the industry or purely coincidental.
Dr. Ingram provided various lines of evidence to support his hypothesis that Revit was based on Reflex, but that evidence does not stand up to scrutiny.
First, he asserted that several bugs in Revit were identical to bugs in Reflex. He gave the example of a complex wall joint. However, the Revit and Reflex behaviours are actually quite different in this case: in Reflex no wall component layer extends the full length of the vertical wall, and many of the Reflex layer boundaries end in space with no connecting layer boundary. Furthermore, the behaviour in Revit is not in fact a bug — it is one of several possible ways this wall joint can be built. Revit has an “Edit Wall Joins” command that allows users to specify which of these valid constructions they intend. So, it turns oseut this is an example of a difference between the two systems, not a similarity.
Dr. Ingram’s second line of evidence was a list of features common to both systems, such as “Instance based building model providing coordinated drawings, model, and information”.
All of the features he listed were in Pro/ Engineer, or were construction-industry analogs of Pro/Engineer capabilities, and Leonid and I had helped implement them a decade before we had ever heard of Reflex. Many of those features also existed in other AEC systems, including 3D Home Architect. To assert these commonalities as evidence that one system was based on the other would be like asserting that one car design is based on another because both have an engine and doors.
Finally, Dr. Ingram states that 100 manyears of effort went into Reflex, and he doubted Revit could have been built from scratch as quickly as it was. However, by the time Revit was acquired by Autodesk in 2002 it had a comparable amount of development effort. Revit raised more than $50 million of venture capital, more than enough to fund such development. We had an elite team of talented developers who had experience developing parametric systems and knew what they were doing.
It turns out, though, that Reflex did have a role in the history of Revit, but it is one that Dr. Ingram was not aware of. The team that helped PTC evaluate whether to acquire Reflex included both Leonid Raiz and me, as well as several consultants from the AEC industry.
Those consultants were very impressed with Reflex, and considered it to be the state of the art for that industry. Leonid and I were surprised to hear that, because the software lacked many of the capabilities that we had come to expect in mechanical design systems. For example, at the time, if you moved a wall in Reflex, all of its doors and windows stayed where they had been, hanging in space, instead of moving with the wall. Pro/Engineer had been able to correctly handle that kind of thing for more than a decade. That helped us recognise that there was an opportunity to provide better software for the AEC industry.
Dr. Irwin Jungreis Revit Founder
Dr Jonathan Ingram has the right to respond.
If you enjoyed this article, subscribe to AEC Magazine for FREE