Originally Posted by
Roger Evans
So how did it all start? What was the spark?
Question for those in the know ..
What I am wondering is what sequence of events / circumstances / coincidences / meetings / technology / thoughts / motivations / collaborations etc etc that coincided to give rise to the birth of Revit?
I know the light bulb needed changing .. but how did you find the light switch in the first place?
In partial answer, here are the remarks I made at the Revit launch event, April 2000.
Irwin Jungreis, Revit launch, April 2000:
I’ve been looking forward to tonight since two and a half years ago. At that time, I decided to investigate whether there was an opportunity for a new company in the AEC market. It turned out that Leonid had independently decided to start a company to work on the same problem, so we decided to get together.
I have a friend who is an architect and former CAD manager. The first thing I did was spend a week with her looking at the software that architects used and getting an initial understanding of what they needed. I couldn’t believe how far the software was from what it could be.
I’m someone who likes to fix things, and when I saw the state of software for the AEC industry, I knew I had a mission, to build the software that would let architects do their jobs more easily and productively. To let them focus on design, while the software took care of the rest.
The first thing that Leonid and I did was talk to dozens of building design professionals to learn more about their needs and ideas. Some of those people are in this room now, and I can’t thank you enough for the help and encouragement you gave us. I hope that what we are building can live up to the vision that you articulated.
We realized that the fundamental problem was change propagation. A large building and its documentation have hundreds of thousands or even millions of components and annotations, with many complex interrelationships between them. We knew that we would need an engine for propagating changes between elements that could handle that level of complexity.
We looked first to existing solutions. In the Mechanical CAD industry, people had found two ways of dealing with constraints, Variational and history-based parametric. Variational solvers solved all constraints in the whole model simultaneously. That could never work for a building of any size. History-based parametric solvers remember the order in which things were created, and replay the whole sequence with new parameters any time a change is made. That would mean that even the smallest change would require regenerating the whole building. In the AEC industry, for the most part the software had very little change propagation at all; certainly nothing general enough to handle all of the changes that come up in real architectural work.
What we came up with was a new approach that we call Context-driven Parametrics. The idea is that, unlike history-based solvers, the way things solve does not depend on the order in which they were created. Instead depending on what change you are making, the solver dynamically constructs the minimal sequence of steps needed to propagate that change. The time it takes to make a change depends on the size of the change, not on the size of the model. Furthermore, the effects of the change don’t depend on how the model was created, so they are easier to understand by other members of the design team.
Once we figured out the context-driven parametric solver, we knew that the technical challenges of building Revit were manageable and we turned our attention to team building. We have been extremely fortunate in the quality of people who has joined our team. You’ve gotten to meet Dave and Alex. I feel very lucky that we were able to attract such talented people to our effort. The development team that has been bringing this effort to fruition is, I am certain, the most talented group of developers the CAD industry has ever seen. I would really like to thank everyone in the company for an almost miraculous achievement of building this company and product.