Transportation Infrastructure and "String Theory"
Jerry Laiserin

Theoretical physicists generally follow Einstein's dictum that "everything should be made a simple as possible, but no simpler." The result often is multi-dimensional reasoning, such as string theory, which seems brain-numbingly complex to laypersons, but harbors an elegant simplicity. Down in the well-grounded realm of applied physics in transportation, the folks at Infrasoft developed a modeling concept called "string objects" that radically simplifies the multi-dimensional problem of collaborative design for road and rail infrastructure engineering and construction.

Unlike many alternative road/rail design software solutions, which are based on interconnected collections of templates and profiles (sections) extruded along a path, Infrasoft’s MX series employs 3D string objects to represent curbs, drainage, crowns and so forth. Because these objects are parametric and contain attributes that define their relationship to one another, changing an MX object—and the entire model of which it is a part—requires nothing more than entering a single change to a single parameter. For example, widening a shoulder simply requires entry of the new shoulder width; the software ripples the change through all impacted objects that comprise the model (whereas in drawing-based civil CAD, every road profile drawing that included the widened shoulder would have to be individually recalculated, redrawn, checked and coordinated).

Because changes to Infrasoft’s underlying object-based models can be initiated so efficiently, the company is able to deploy a distributed processing style of collaborative design, which it calls Arenium. Every Arenium user, regardless of location or platform, runs a local copy of the editing/annotating software. Models are stored centrally on a server (on an enterprise or ASP basis), from which each authorized user can check out a copy. Each user’s copy of the model then resides locally, along with the editing software. The server keeps track of: which users have checked out copies of which models; who is authorized to view only, to mark up or to make changes to the model; and which users have requested notification of changes to the models they have checked out.

Under this regime, if users A, B and C each check out copies of a model, and user A makes a change to any element of that model, what happens is as follows: user A edits her instance of the model locally, with only the changed attributes of the edited objects transmitted to the server; because the change information transmitted is so small, communication is nearly instantaneous—even over a slow dial-up connection; as the server-based instance of the model updates, notification of the change is automatically sent to users B and C; as soon as users B or C acknowledge receipt of the change notification, the server sends the change information to update their respective local copies of the model; again, the transmission and update process is quick and painless because only tiny amounts of changed attribute data are transmitted; all three users now have up-to-date local copies of the shared model, all synchronized with the server.

If user B were to query the change via a markup or annotation, only B’s redline information would be transmitted and synchronized, yet it would be permanently linked to the model and to all previous redlines. User C might offer a comment or question via the system’s built-in instant messaging capability, and both the question and any subsequent dialog would also be recorded on the server, linked to the model. The server logs all edits, markups and comments, and the project can be rolled backward or forward through that log; for example, to replay a specific decision process or to review the collaborative history of various design alternatives. This system not only supports truly collaborative design, but concurrent design as well, with multiple teammates working on separate instances of the model simultaneously yet independently (merging and reconciliation of changes occurs at the server).

With transportation clientele all over the world, often in locations where high-speed Internet access is problematic, Infrasoft empowers multi-national civil engineering project teams with a new level of collaborative design functionality. At "press time" for this issue, Bentley Systems announced it is acquiring Infrasoft. As it has with other recent acquisitions (such as Rebis, in the plant space) Bentley likely will use its "V8 Generation" technology to continue supporting MX/Arenium for DWG as well as for its own DGN format. By adding Bentley's civil design market strength to Infrasoft's world-class technology, this acquisition bodes well for the future of true collaborative design of transportation infrastructure.
JL

NOTE: portions of this article were previously published in the August 2001 issue of CADENCE magazine.



< back