Ever since Autodesk announced its intention to purchase Revit back in February this year, Autodesk's AEC strategy has been, to say the least, a little unclear. Autodesk already had a fairly successful AEC [architecture, engineering, construction] vertical product in Architectural Desktop. One wonders if the company bought Revit to take the product out of the game, or to port the technology to AutoCAD, or to replace Architectural Desktop? One could infer a lot from the deal.

Autodesk paid US$133 million for a company with little global reach, no developers, and only a handful of customers. From these facts alone, it seemed Revit was more of a strategic purchase for which many in the industry felt that Autodesk paid too much. Revit's aggressive marketing generated the perception that is was as much an up-start company, as it was a start-up!

The previous time Autodesk bought AEC technology, it purchased SoftDesk (formerly DCA Engineering) and its AutoArchitect product. Initially, the deal ran into some trouble after Autodesk announced its intended purchase price of around US$70 million. PTC, an MCAD company with AEC intentions at the time, joined the bidding process, pushed up the price, and forced Autodesk to pay substantially more -- around US$90 million. For that price, Autodesk got a large, mainly US-based customer list, a mature product, and new technology based on Autodesk's own ARX development language (which promised to deliver a CAD system based on intelligent modeling). AutoArchitect and the new object technology became Architectural Desktop, which is sold worldwide as Autodesk's AEC solution based on the industry de facto standard AutoCAD.      

With the complex new concept of the Single Building Model, Autodesk had the choice to:

(1) create a completely new product; or

(2) develop on its hugely popular AutoCAD foundation.

The SoftDesk purchase meant Autodesk was pursuing the latter option, even though using existing technology could create limitations in the future. (Autodesk's Mechanical team originally took a similar route, creating a solid modeling application, Mechanical Desktop, on top of AutoCAD.

This proved too limiting, and they found themselves in an uneven fight against new 3D modeling tools like SolidWorks. This lead to the development of Inventor from the ground-up.) The big question was: would Autodesk's AEC division develop its own Inventor-style product, and replace ADT -- or buy something?

At the time, Phil Bernstein (vp of Building at Autodesk) told me, "It's hard to predict, but we have always thought of ADT as a bridging mechanism, in much the same way that MDT was a bridge to Inventor.

It's a way of starting to change the way people think about things, but whether it's going to last three years, five years, or ten years is something we are trying to figure out at this very minute." Little did I know that at that very minute Autodesk was opening channels with Revit concerning a possible buyout.

- - -

So, into this setup comes Revit and its Waltham (near Boston MA USA) - based development team. The Revit developers, after being purchased for $133 million, consider themselves the future, with backing from the very top tiers of Autodesk. The ADT team, while being successful in shipping the numbers, and working to fix shortcomings of the product, now find out they have internal competition.

Autodesk's competition has been quick to jump on the incompatibilities between ADT, Revit, and AutoCAD. Because Revit was not developed by Autodesk:

• Its DWG is reliant on the OpenDWG interpretation of the format.

• Its user interface is nothing like Autodesk's products.

• It requires a single building outlook and works in 3D, not 2D.

• It has no third-party applications or API [application programming interface].

On the face of it, Revit requires architects to jump from the relative safety of the AutoCAD platform to a new-new thing. Switching mission-critical applications is risky business, and I'm afraid to say that many architectural practices are institutionally risk-adverse.

So Autodesk has had all summer to come up with a cohesive story for its Building Design Solutions division and various applications. Carl Bass, Autodesk executive vp of Design Solutions Division, recently visited England to launch the UK version of Buzzsaw, where I got the chance to ask him about Autodesk's elusive AEC strategy.  

2

Prior to getting into the details of the products, Carl Bass put forward Autodesk's view on the building industry and its processes. The first task appears to have been to give this view a title, and here Autodesk has jumped on the latest popular three-letter acronym -- Product Lifecycle Management -- but contorted it to fit its purpose. Be prepared to hear a lot about "Building Lifecycle Management" (BLM) and "Infrastructure Lifecycle Management" (ILM).

Mr Bass explained that the building industry has to change. At the moment, the building industry is heavily fragmented and inefficient, a business in need of change. He drew parallels to the manufacturing industry of the 1970s: "Things were easier in the manufacturing industry as the value chain wasn't broken. Unfortunately, the people we are asking to make the change are the ones who aren't going to get the benefit. Many of your readers are the ones that are going to have to change the way they work, and they are not going to wake up tomorrow and get paid 10% more."

I asked if his opinion was 'a bit rich' coming from a software company, to which Bass responded, "It's not up to us [Autodesk] to change the industry. We are tool providers, and we think we enable the people who want to make changes to their business. It's not our job to change it. We are in the business of providing software and solutions.

"Fragmented teams lead to the disintegration of the digital data. We are not trying to change the industry. The industry is trying to innovate in different ways, and it's much more about business innovation than it is about technology innovation.

"We are already seeing this with the growth of design-build and other vertical integration, together with other mergers acquisitions. We are seeing many architects and engineers combining; architects, engineers, and general contractors combining. The framework agreements are changing. too; it's about sharing risk and reward. It takes different forms and tries to get rid of some of the byproducts, like the litigation, which has become an integral part of driving behavior out of the system. At the end of the day, you need recourse in case somebody doesn't perform, but it shouldn't drive the business.

"It's going to be particularly difficult for the architects. The contractors have lived with risk all along. Architects are not familiar with that risk. Going from time and materials to all of a sudden being able to make $10 million or lose $10 million, is almost something that puts them out of business.

"So, it's not just a technology change. It's a cultural-sociological change that needs to go on for any of this stuff to take place. We are trying to get some parallelism in the system, to compress the process to deliver projects on time. The first step is to keep the information digital throughout the process. We also want to get the data digital earlier in the process and stay digital throughout. Every time it goes analog -- printed out to paper -- people lose information. Something bad happens every time data leaves the digital realm. Finding errors at each subsequent stage of the process is about an order of magnitude expensive to fix."

- - -

Preaching about how architects have to change is nothing new, and it's a fact that the design part of the building process doesn't spend a lot on its design tools. I suggested, however, to Mr Bass that the reason was that architects didn't see any immediate benefit of moving to a 3D single building model, when their old technology still 'works'.   

Mr Bass replied, "The reasons why people are not willing to invest is back to that problem of the value chain being broken. The people who need to change are the architects, and that's why all those other parties -- the owners, developers, and contractors -- are getting more involved, as well as architects: to take on some of the risks in return for greater reward. It's absolutely not a technology provider that changes this. If an architect is now in a design-build, joint-venture, or framework agreement, all of a sudden it's a fascinating idea about how they work more quickly.

An RFI [request for information] or change order that used to mean more money is now a cost. It's an entirely different incentive system that's in place, in the way the architect will see changes. Is it worth it, at that point, authoring more intelligent information? Absolutely. The architects that choose to be in design-build have an entirely different relationship.

"I think we have made a mistake just thinking about the architects and engineers. If instead we just think about construction, then we have an entirely different view of the market. Yes, we still have to produce products for people who are happy to work the way they have for the last ten to 20 years. That's absolutely fine, however we'd be shooting for the tail if we didn't try and do more.

"Owners and developers have a real need to do things differently. A good example in the USA is Home Depot. They open a store every 40 hours. Other companies like Starbucks, and McDonalds, are taking a much greater interest from design through to maintenance. They are interested in the building lifecycle. If I can get some of the precision from manufacturing into the construction process, we'd all be better off."  

3

As we know with Autodesk's Architectural Desktop and now Revit, the concept of the "single building model" [SBM] is an essential part of Autodesk's digital construction strategy. To date, the building industry has not made major steps toward the single model, because, as I have pointed out in the past, the single database doesn't map well to the fragmented construction process. It's worth noting that the companies Bass mentioned previously -- Home Depot, Starbucks, McDonald's -- were very typical design/-build, where a centralized model database may well be possible. So, I asked Car Bass, how long before the rest of the market gets it?   

"It takes about ten years for the introduction or widespread adoption of new technology, certainly in architecture," Bass responded. "Some say MCAD [mechanical CAD] is a lot faster in technology adoption but PTC's Pro/Engineer was introduced in 1987. At best, you could say it went mainstream when SolidWorks appeared in 1997. So it was ten years before it became mainstream and today you'd say it's more of a commodity, because there are several vendors providing it. That's 15 years and, as every one of us would tell you, there are still a huge number of people still converting from 2D to 3D.

"Is 2D still important? Absolutely, yes. Will it ever be replaced? No. But more and more 2D is generated automatically [from 3D], more design intent is captured, greater complexity can be more easily expressed and more 'bookkeeping' is done by the computer instead of the architect/engineer."

It's now clear that there is a lot of technology that can help wring out inefficiency from the process, but what will be the catalyst that helps make companies do it? Fortunately or unfortunately, most AEC [architecture, engineering, construction] business is local thereby insulating it from the forces that caused/enabled manufacturers to change. But I think we're beginning to see new technology, new project delivery mechanisms, new economic reasons to drive costs out of the system and a reworking of the economic value chain that will be the real story for the next decade."

- - -

To deliver on this vision of the future, Autodesk's product strategy attacks two areas: collaboration, and what Bass termed "Building Information Modeling" (BIM, oh goodie, another three-letter acronym!), which seems to be Autodesk's new wording for Single Building Model. (Graphisoft calls this the "Virtual Building Model," and it has been known as "modeling with objects"). When it comes to BIM, Buzzsaw appears to be the collaboration and management element, with Revit being the weapon of choice for model creation.

At this point, looking at the terminology, it's worth going over some old ground. Many architects talk about modeling with 'objects,' while  I use the term SBM. Here, Bass wanted to clear the air: "As for 'objects' and SBM, I think both terms are regrettable choices by the industry: 'Objects' is the popularization of a programming concept; it is as interesting to the user as the fact that their cell phone uses CDMA or TDMA. 'SBM' is equally dysfunctional.

"I see ADT and Revit as tools that enable people to more accurately (not geometrically but from a design intent standpoint) model buildings. In addition, as I said before, I think Revit's strength lies in being built upon a particularly strong technology, its parametric change engine."

4

In light of the Revit acquisition, what's the story with ADT [Architectural Desktop], its customers, and the translation plans?

I guess this is the hundred-million-dollar question -- $133 million to be exact. The arrival of Revit at Autodesk caused some internal rumblings between development teams. From the outside, the acquisition appeared to seal the fate of ADT. With 250,000 customers, however, one would expect to hear a certain amount of soothing and reassurance from Autodesk. Here, Carl Bass obliged me:

"When we introduced ADT, MDT [Mechanical Desktop], LDT [Land Desktop], and Map, people thought that they would mean the end of AutoCAD. Obviously, that didn't happen. We have customers using ADT on three-to-five year projects; they are not going to change in the middle of their projects. They are not going to introduce new design software at the tail end of a ten-year project. We will continue to develop AutoCAD and ADT as long as it serves a customer need; I suspect this will be for many, many years. I would not suggest that anyone should blindly believe the superiority of products like Revit.

"People who benefit from the change, need to drive the change. What you are seeing are people who understand that the industry needs to and will change. Fifteen years from now, I think you'll look back and see this was the beginning."

- - -

I guess that could also be interpreted as the beginning of the end of ADT. Autodesk has openly committed to carry on developing ADT for the foreseeable future, but now states that Revit is the premier authoring product that will assist the constructing industry in reworking its process. So where did ADT fail? As some of my readers know, I have many issues with ADT.

Bass highlighted the database as his fundamental reason: "I'd say, from a technology point of view, AutoCAD was not designed to manage relationships. It was meant to draw, keep things on layers, along with some organizing principals. It was not the idea that if I change one line, 500 other lines or text strings also have to change. Revit is built with that fundamental idea.

"I think what's interesting from the user's point of view is that when I change something in Revit or Inventor, my drawing updates. I'm working in 3D, and I move a wall three inches, and somehow everything changes in the plan view. So one of the main goals with Revit is to get people spending 95% of the time worrying about the design problem, not how the software works. That's the essence of it.

"If you can spend most of your time authoring a more intelligent model, you are sharing information that's more relevant, and not worrying about layer standards, etc. Communicate real design intent, and downstream there's an upside for cost estimating, quantities, scheduling, analysis, facilities management, etc. All of which are benefits you get by creating a real building model."

- - -

So, if you were considering an Autodesk AEC solution, which one would you go for: ADT or Revit? Bass replied, "I'd look at Revit and ADT, and figure out which best fits your need." While on the face of it, that's a sensible route to take, a move to Revit would mean a major upheaval for an AutoCAD customer:

• Revit doesn't have a similar interface to AutoCAD, and would require retraining.

• It doesn't work the same as AutoCAD, and would mean examining the processes.

• It uses the Open DWG engine to produce DWG files, an engine that in the past Autodesk has marketed against by saying its own products were 100% DWG compatible (vs. the competition).

I put the DWG problem to Bass, and he replied, "We think the DWGs are 100% compatible."

I pressed him for a guarantee, and he responded, "For DWGs created in AutoCAD -- absolutely, we will guarantee that. With ADT, we will guarantee that the graphics come across. It comes down to what you actually want us to translate. We are working on the problem of what you want an ADT door to become in Revit. You obviously want it to become a door. Which family of doors, how do you control that data migration thing, is absolutely an issue. People in MCAD wrestled with data translation for 20 years. It is an understood, but not solved problem."

Herein lies the rub. You always hurt the ones you love the most. At the moment, the early adopters of ADT, who wrestled and tamed the product, have created intelligent ADT data to produce dumb graphics in Revit. Bass did add, "We will work on the translation absolutely.

The two important messages in translation from ADT to Revit is we will protect the customers investment with Autodesk, and second, is that we will do our best to make sure you can migrate your data to the best level possible.

"Why can I sit here and be confident about 100% DWG compatibility? There is no way I can assure you today that every single DWG you can find, would work. What I can assure you is that I know there's a guy sitting in Waltham [location of Revit's office] who if I sent him one tonight that didn't work, he'd send me software tomorrow that would make it work."

5

Conclusion by Martyn Day

Autodesk feels clearly that Revit offers new possibilities to architects -- should architects want to change their processes, take on more responsibility within the construction industry, and reap greater rewards. Unfortunately, on analyzing Autodesk's current AEC product offerings, the range now looks as fractured as the industry it's trying to sell to:

• Architectural Studio is a great new conceptual tool on a new code base that doesn't support DWG. Its user interface is differs greatly from any other Autodesk product
• AutoCAD and LT are the mainstays of the profession, setting the de facto standard DWG and DXF file standards. Although 3D-capable, the two are used predominantly for 2D drafting. I understand that the AutoCAD, due out next year, feature a new DWG file format, which may create problems for users, developers, and competitors.
• Architectural Desktop was Autodesk's first global attempt at an AEC-specific products (previous AEC products were region-specific). The technology broke DWG, however, by creating elements that could be read by ADT only, creating problems for non-ADT users, and requiring Autodesk to develop the Object Enabler plug-in. While a quickly maturing product, the purchase of Revit, and Autodesk's message (that Revit is the future), raises the specter of ADT's creation of legacy data in my mind.

• Revit has a brand new foundation and powerful parametric engine. Unfortunately (or fortunately depending on your viewpoint), it doesn't have AutoCAD's interface, doesn't have an API [application programming interface] for third parties, is essentially 3D, has a single unified database (not DWG files), and is at the moment an incomplete solution.
• Buzzsaw is an online project portal that offers some basic drawing management features with interesting collaboration potential. Unfortunately, it isn't available behind the firewall unless you have hundreds of thousands to spend. So, Buzzsaw is good for external collaboration, but not for internal file management.   

I am sure all this will change over the coming years, and from the time scales mentioned by Autodesk vp Carl Bass, perhaps over a decade. There's no massive urgency to fit all the products together seamlessly. However, I am also sure that this news will unnerve some users that have invested in ADT.

Along with Inventor, Revit is now an important product for Autodesk. Both represent movement away from Autodesk's flagship AutoCAD product. The MCAD side of things demonstrates this can happen extremely quickly. I get the impression Autodesk realizes that this is going to be a much more drawn out process in the AEC space.

Martyn Day is group editor of CAD magazines in the UK. http://www.cadserver.co.uk

Copyright © 2003 by TenLinks, Inc.