CADENCE MAGAZINE Contributing Editor and monthly columnist Jerry Laiserin, FAIA, provides strategic consulting services to AEC businesses and their technology providers, including some companies mentioned. He also is a contributing editor to and monthly columnist for Architectural Record and a frequent contributor to many other AEC and technology publications. Reach him at jerry@laiserin.com.

Jerry Laiserin, FAIA
Collaborative AEC Design


Figure 1. Rendered image of Rose Center for Earth and Space, from Polshek Partnership in NYC (courtesy of Polshek Partnership).




Figure 2. Vite Project is a "project strategy optimization" tool that can be used to model overall project workflows [a] and study detailed interaction among collaborative design tasks [b]. Through multiple simulations, Vite Project can compare project outcomes [c] for alternative scenarios [d]. One analysis predicts that collaborative design with integrated 3D models may yield 20 percent reductions in project time and cost (courtesy of Ken Stowe, George B.H. Macomber Company).

Figure 3. Graphisoft's ArchiCAD has for years set the benchmarks for AEC collaborative design, with world-class change management, extensibility and interoperability (courtesy of Graphisoft).

Figure 4. As an overlay to AutoCAD or IntelliCAD, Bricsnet Architecturals supports collaborative design by combining the best of integrated architectural modeling with native support for industry standard DWG files (courtesy of Bricsnet).

Figure 5. Bentley Systems' ProjectBank supports design collaboration through a finely granular, element-by-element level of change management that is capable of supporting multiple file types (courtesy of Bentley Systems).


Figure 6. ActiveProject, from Framework Technologies, supports interactive, Web-based collaborative markup in real time for design models [a] and project schedules [b] (courtesy of Framework Technologies).


Figure 7. Autodesk's OnSite View extends design collaboration to handheld devices that can be used on the job site as well as in the office (courtesy of Autodesk).

Figure 8. VisionPlanner, a Web-based collaborative service shown here in a late Beta version, captures and organizes project information to help make future design collaborations more effective (courtesy of VisionPlanner.com).
Is today's favorite buzzword-collaboration-fact or fiction in the AEC world? Perhaps a little of both, but the tools exist and the path broached for more to follow and sooner rather than later.
Building buildings is one of humankind's most collaborative activities. Yet, all of our most modern computer and communication technologies have done little to support and advance the collaborative nature of the processes for designing and constructing the built environment. To be sure, there are software tools in abundance to automate individual tasks, such as drawing, specifying, estimating and scheduling. And, each of these tools or sets of tools have been adapted to the specific needs of the disparate participants in the project team, from designers and design consultants through constructors and trade subcontractors to building owners and operators. In practical usage, these tools have done little more than to separately automate previously manual tasks, without altering their distinct-and essentially discontinuous-nature, a process that business guru Michael Hammer (Reengineering the Corporation) called "paving the cowpaths." Even the Internet, the much-heralded "mother of all collaborative environments," has been deployed in AEC to little more effect than "automating" FedEx out of the document-distribution chain.

Are AEC businesses so hopelessly backward, so mired in their heritage of mud and sticks that they will never adopt the kind of digitally-enabled collaboration that is sweeping through every other global industry? Is the AEC world so fragmented numerically and so parochial geographically that the Internet-driven collapse of time and distance that is imploding other industries will simply pass us by? This does not have to be the case, as occasional glimmers of optimism shine through the gloom, lighting the way to a brighter future for AEC. For some 20 years, pundits of AEC technology have rushed toward the light at the end of the tunnel, only to be flattened by the onrushing train of hidebound traditionalism and timeworn procedure. However, the relentless advance of Moore's Law (computing power doubles every 18 months) and of Metcalfe's Law (the value of a network increases as the square of the number of connections) may finally have brought us to the dawn of a new day in AEC collaboration.

The Way We Were

Ever since the Renaissance in European culture, when architecture (design) emerged as a discipline separate from building (construction), the inherently integrated collaboration of the master builder has been replaced with a highly stylized form of design communication. Architects describe, in words and pictures, what a proposed building will look like and what it shall be made of. Depending on the society and its customs, different methods may be followed for communicating that design intent and converting it into built form. In Japan, for example, the architect is in charge of all aspects of a building project, right down to construction staging, in ways that exceed the ambitions of even the most rabid design-build proponents in the Americas. German architects are responsible for all quantities in their bid documents, with builders bidding only on labor and coordination. In the UK, a separately licensed profession of chartered surveyors exists to provide material quantity take-offs independent of both the architect and the contractor. In the "traditional" US method of design-bid-build, all determinations of quantities and methods of execution are left to the contractor. In fact, until well into the twentieth century, US contractors also assumed responsibility for the design and execution of most major building systems, from structural and ventilation to plumbing and conveyances (the evolution of specialty engineering consultants, accountable directly to the architect for HVAC, MEP, vertical transportation and so forth, is a relatively recent phenomenon).

The conclusion to be drawn from this diversity of cultural and historical arrangements for building collaboration is that the present North American system of AEC communication and collaboration is neither divinely inspired nor the derivative of some natural law. Rather, it is simply a collection of customs, procedures and legal precedents that survives and thrives only because of its mutual acceptance by a majority of project team participants. The bottom line is that there is no reason inherent in the AEC process itself that dictates doing it this way and only this way. When American designers or constructors work in other countries, they readily adapt to local custom and protocol for project collaboration. It is reasonable to hope (if not expect) that some day soon we will demonstrate similar flexibility and enlightened self-interest in adapting the way we work on projects closer to home.

A case in point is the new Rose Center for Earth and Space at the American Museum of Natural History in New York City, as shown in Figure 1. The Rose, as it is called, replaced the beloved but sadly antiquated Hayden Planetarium on the same site. A tour de force of architectural design by Polshek Partnership of New York, the Rose already is a serious contender as one of the great buildings of a new century that is not yet one year old. However, for all the twenty-first century design heroics, the project also represents a heroic culmination of the collaborative methods of the twentieth century. Nearly a dozen distinct and mutually non-interoperable computer models were deployed by various participants on the Rose project team. Rather than indicating a breakdown in communication, this orgy of modeling resulted from each project participant's wish to use the best-suited tool for each task at hand, combined with a willingness to incur the inefficiencies of file translation and data conversion as a fair price for achieving such an extraordinary level of design excellence.

The architects designed the building and its major elements in MicroStation from Bentley Systems (www.bentley.com) and also used MicroStation Modeler for the numerous cast and cut, custom machined parts that comprise many key components, such as the steel fittings for the glass curtain wall. Most of the architect's consultants worked in AutoCAD from Autodesk (www.autodesk.com), while modeling specific aspects of design performance with separate tools for finite element analysis (FEA), computational fluid dynamics (CFD), natural lighting (DOE-Radiance) and artificial lighting (Lightscape from Discreet, www.discreet.com). Even the rendered images for fund-raising and publicity required their own separate models (albeit using data imported from MicroStation), crafted in formlZ (www.formz.com) and rendered in Softimage (www.avid.com).

Meanwhile, the contractors and fabricators were equally busy, modeling all major structural elements in a 3D AutoCAD model in lieu of conventional shop drawings. Foundries rebuilt the architect's computer models of castings in SolidWorks (www.solidworks.com) to better control the order of operations, while parts to be cut by computer-numerical controlled (CNC) machines were redrawn with an accuracy of eight decimal places (CNC equipment demands such accuracy, but it would be impractical for the architect to carry that level of precision throughout a building model where many construction components are built on-site to field tolerances of an eighth of an inch or more). Every trade had to perform quantity take-offs and material estimates (essentially, cost models), while the construction manager, Morse Diesel of New York, and various prime contractors modeled the sequencing and schedules.

Very few projects have the kind of budget and level of design intent to sustain and justify such extraordinary collaborative efforts, and it is unlikely that the process that went into the Rose will be repeated again. Surely, the built result is well worth the effort, but that degree of effort may no longer be necessary, at least not if visionaries such as Ken Stowe have their way.

All Together Now

Serving as Director of Project Services at the old-line Boston firm of George B.H. Macomber, builders and construction managers, Kenneth H. Stowe, P.E., is at the forefront of transforming not just the Macomber firm but the entire AEC industry. With long-term industry experience, including many projects for Walt Disney, Stowe has an intuitive grasp of the interplay among quality, cost and speed in design collaboration. In private conversation, Stowe is fond of the epithet "pitiful," applied as a blanket condemnation of current and historical AEC efforts at project collaboration, his own included. Fortunately, Stowe is one of those intrepid souls who would rather light a single candle than curse the darkness.

Seizing the promise of a new generation of integrated CAD tools and an emerging prospect of interoperability via the Internet and aecXML, Stowe embarked on an experiment to model the existing AEC design collaboration process and to quantify the potential impact on cost and schedule that might accrue from various alternative scenarios of improved collaboration. Working with a "project-strategy optimization" tool called Vite Project (www.vite.com) developed by engineers associated with the Stanford University Center for Integrated Facility Engineering (CIFE), Stowe says that he "started out by creating a baseline model, reflecting the way design usually goes, using the old process." Unlike the more familiar GANTT, PERT and CPM diagrams produced with project management tools such as Microsoft Project (www. microsoft.com) or Primavera Project Planner (www.primavera.com), Stowe's Vite model, as shown in Figures 2a and 2b, contains more than 70 activities that are linked by a complex network of "actor assignments," dependencies, "failure dependencies" (feedback requiring rework of prior activities) and information exchanges between activities.

With the help of architects Jill Rothenberg of Cambridge-based ADD, Inc., and Larry Rocha of WAT&G in Newport Beach, CA, Stowe developed and refined estimates of the productivity improvements that would accrue at different steps of the process and to different project participants as a consequence of introducing an integrated 3D building modeler as the primary tool for design communication and collaboration. In addition to the architects, Stowe also interviewed estimators to determine the potential impact of dynamically integrating the design model and the cost model. (In other words, on a typical four-week estimate and value-engineering cycle, how much effort could be saved through model-based automatic take-offs and the ability to digitally communicate the estimate and document its assumptions?) Vite then runs a project simulation on a per-activity basis that yields as its output a projected probability that any activity will consume the time and cost budgeted for it. Whereas a CPM (critical-path method) analysis tells project participants how the project will turn out if everything goes according to plan, a Vite simulation provides a more realistic simulation of the likely outcome. In one sense, this is a kind of risk analysis (what are the chances any activity or the entire project will run over schedule or over budget?), but it also can be inverted, as Stowe did, to look for opportunities (which are the activities and actors whose enhancement will contribute to the most likely improvement in schedule or budget?).

Stowe's baseline model assumed conventional tools and no direct collaboration between architect and builder. Alternative scenarios (Figures 2c and 2d) explored cost and schedule implications for architect and builder, sharing a high-level collaborative building model in 3D over the Internet; extending that collaboration to include the architect's consultants and the builder's subcontractors and various alternative workday schedules. Although the results to date represent only one pilot project, these simulations indicate that some 20 percent of project time and project cost could be saved by Web-based collaboration via an integrated 3D model shared by all project participants from the earliest stages of the project. There also are hints in Stowe's data that projects not only will become faster and less expensive, but less risky as well, because the new style of Web-model-based collaboration will reduce project risk (that is, narrow the variance between the ideal CPM duration and cost versus the likely simulated duration and cost).

Model Behavior

Ken Stowe's brilliant process simulations would amount to nothing more than an interesting academic exercise if there weren't tools on or near the market that could empower real-world architects and builders to act on the promise of Stowe's Vite simulations. This is where Moore's and Metcalfe's laws come in. As computing technology has become ever more powerful, in accordance with Moore's Law, affordable CAD software becomes capable of data-handling feats unimagined even in relatively recent times. Power doubling every 18 months means an order of magnitude shift every five years, or a ten-thousand-fold improvement in the 20 years since the likes of GDS (www.informatix.co.uk), Intergraph (www.intergraph.com) and the original ARRIS (www.arriscad.com) ruled the AEC CAD world. Ken Stowe's process modeling assumed that the 3D Web-enabled CAD model with integrated costing capabilities would be Revit, from Revit Technology Corp. (www.revit.com). However, there is an entire generation of new or updated CAD tools striving for the same niche. Much of the market appeal for Web-based collaboration derives from the consequences of Metcalfe's Law, which makes a network (or networked application) exponentially more attractive the more people use it.

What kind of model software will support this new level of collaborative design, and what are the key characteristics that will determine a successful level of collaborative support in any CAD package? It is increasingly clear that CAD users will value three major characteristics in their modeling tools: change management (see "Client-Driven AEC Design," CADENCE, June 2000, pp. 16-26); extensibility, which means the ability to interoperate with cost, scheduling and other tools that depend on attributes other than geometry; and compatibility with de facto CAD standards (for example, DWG). A quick review of some leading contenders clearly shows how major AEC CAD vendors are addressing these critical factors for collaborative design. Revit (see "Trial Runs," CADENCE, May 2000, pp. 35-36) relies on a "parametric change engine" to manage and reflect all changes anywhere in the model and to resolve conflicts among posted changes by multiple team participants. Revit has announced a partnership with data publisher R.S. Means (www.rsmeans.com) that links Revit models with Means' CostWorks to yield parametric costing, a prime example of extensibility. And Revit relies on AutoCAD DWG for detailing (in fact, Revit may be a better tool for managing and integrating AutoCAD xrefs than AutoCAD itself).

ArchiCAD, from Graphisoft (www.graphisoft.com, see Figure 3 and also "CAD Options," CADENCE, July 2000), with its "3D Virtual building" and "TeamWork" functionality, long has had world-class change management capabilities, along with extensibility for costing, bill-of-materials and so forth (via its built-in GDL language, now available in an Internet-enabled object format). ArchiCAD's recently added "hotlinking" capability not only creates a reference-file-like capability for change- management purposes, but also increases the product's DWG compatibility by allowing hotlinking of AutoCAD xrefs. Bricsnet Architecturals (www.bricsnet.com, see Figure 4 and a forthcoming "CAD Options" review in CADENCE, December 2000) combines the 3D change management functionality and data extensibility that its developers also provided to MicroStation TriForma, and overlays that on top of the user's choice of IntelliCAD (also from Bricsnet, although licensed from Microsoft/Visio) or AutoCAD itself. Needless to say, AutoCAD has the DWG compatibility base covered, and Autodesk is a pioneer in data extensibility (think of the original ADE, AutoCAD Data Extension). Just as the AutoCAD 2000i lineup extends AutoCAD's collaborative abilities across the Internet, it is reasonable to anticipate that coming upgrade releases of Architectural Desktop will enhance its change-management capability.

Bentley Systems offers yet another example of a vendor steering product strategy toward the new collaborative imperatives. Although it is an open secret that the next version of Bentley's DGN file format may sacrifice some backward compatibility in order to work more smoothly with DWG, Bentley already has invested considerable effort in its ProjectBank technology, as shown in Figure 5. ProjectBank provides the ultimate granularity of change management, with an entire project journaled in a way that allows team participants to roll any change to any entity forward or backward through the project history (this technique also allows the editing tool to be independent of the file format and the underlying data, a feature touted for the ability to operate equally on DWG and DGN, but surely a handy thing to have around for those operating on a future "DGN2" alongside an existing DGN). MicroStation TriForma, like Revit, has a partnership link to R.S.Means' CostWorks, although Bentley accomplishes its extensibility through aecXML (appropriately so, as the "X" in aecXML stands for "eXtensible").

Look for Bentley's ProjectBank change management and TriForma cost integration to migrate to Viecon (www.viecon.com), Bentley's project collaboration/portal hybrid, part of a burgeoning trend among AEC vendors-besides Architecturals, Bricsnet offers project collaboration, information and procurement and Autodesk has a close partnership with its 40 percent-owned buzzsaw.com (www.buzzsaw.com, also see "Understanding the AEC Project Management Dot Coms" in the special supplement to CADENCE, AEC/MCAD: State of the Art in Collaborative Design, June 2000). The ability to do live, real-time, collaborative markup of CAD models, project schedules and similar information already is available through collaborative tools such as ActiveProject from Framework Technologies (www.frametech.com, see Figures 6a and 6b). Other collaborative design vendors are sure to follow.

Closing the Loop

AEC design collaboration involves numerous participants who are either or both temporally displaced and geographically dispersed-in other words, out of synch both in time and place. Linking the extended project team-folks more likely to be out in the field than in place behind an office PC-is one necessary technological component for completing the circuit of information collaboration on AEC projects. The other necessary piece would be some means of capturing the experience gained on prior projects and recirculating that collaborative wisdom over time to members of current and future project teams.

Delivering collaborative information to members of the extended project team, wherever they may be, and whatever form-factor device they may have access to, is the province of a new class of collaborative applications exemplified by Autodesk's OnSite View, as shown in Figures 7a and 7b. Essentially, this extends the viewing and markup capabilities of Autodesk's Volo View out to affordable handheld devices running the Windows CE (2.11 or higher) operating system. Available as both a personal productivity application via desktop synchronization, or as an enterprise application, OnSite View represents a significant opening wedge of CAD collaboration tools into the emerging "space" of wireless communications to and from handheld devices. Not only can architects, for example, use OnSite View to explore design information collaboratively in the field, but they also can issue design sketches from the field and beam them directly to other collaborators.

Finally, all the process improvements and enhanced collaborative benefits potentially gained through simulation, modeling and communication tools all will go for naught unless project participants enjoy some means for capturing, analyzing and reusing the lessons learned from past projects to improve team performance on new projects. That, precisely, is the goal of VisionPlanner (www.visionplanner.com), an Internet startup company that had the insight or foresight to target a project-collaboration niche thus far untouched by all the hundreds of other would-be Internet project-collaboration services: project initiation and closeout. By systematically capturing project-closeout information, slicing and dicing the data and then recycling the lessons learned into future project initiation, as shown in Figure 8, VisionPlanner will bring the world of AEC collaborative design full circle. Designers, constructors, owners and all the subconsultants thereto will at last be able to move forward by learning from and extending the best practices of what does and does not work on real-world design projects.

© 1997–2001 CMP. All rights reserved.