Will Autodesk's intention to acquire Revit Technology, pioneers of parametric building technology, bring clarity to the relative merits of two different technologies?
It has been said that the AEC CAD industry has been lacking in innovation. That successive improvement in this technology has only been incremental and the AEC industry has been missing out on digital modelling technology benefits that other design and fabrication industries have been enjoying for the last ten years.
Harsh words. Even harsher is the accusation that vendors, more intent on extracting software upgrade revenue than pursuing product innovation and banishing entrenched working methods, have dominated the AEC CAD industry.
And we're not finished yet! A provocative war cry I heard recently heralded that "CAD is dead," a sign indeed that vendor emotions run high in this particular market.
Is CAD dead? And if it is, what's the replacement? And does the market place care?
What we are discussing here is the performance and uptake of legacy geometry-based CAD systems versus change-based parametric systems in what is now known as the Single Building Model marketplace.
Origins of geometric & parametric
Fifteen to twenty years ago, when the legacy CAD vendors set out to make computers useful for basic drafting tasks, it was all about describing geometry digitally in a way that could be stored, presented and manipulated on a computer or plotter. The 'geometry engines' resulting from these efforts were and still are the core technology for CAD products today but are arguably a decade out of date as well as notoriously difficult to use.
More recently, many AEC professionals became interested in using computers for something beyond geometry. They wanted to use computers for collaboration, for sharing data with consultants and other construction team managers, for recording all information about the built project for facility management and operations. This is now referred to as 'model-based working methods'.
A parametric building modeller is the first product to deliver truly integrated, digital models of a project that function as easy-to-change building databases. Parametric modelling means that relationships are built between the design elements (walls, doors, windows etc.), which are managed and maintained by the system itself. When editing the model, all the elements that are related to the edited component update in accordance to whatever relationship you allocated. For example, moving a wall in the first floor plan will cause all the related elements to adjust appropriately. The roof will move with the wall preserving any overhang relationship, the other exterior walls will extend to remain connected to the moved wall, the interior partitions will extend to remain connected to the moved wall, and the room areas will update on a room schedule to reflect the new size of the spaces. Instantly all views - 3D, plan, section, elevation, sheets and schedules - are updated with 100% accuracy guaranteed.
Telling the difference
What makes a straight comparison between parametric and geometric CAD products a tough one is that the lines between these two fundamentally different technologies are seemingly blurred by legacy geometry-engine based products, making a point of being "object-oriented," or having "smart" or "intelligent" objects that display parametric-like behaviours. This and much more accompanying techno vendor speak', gives the impression that the systems are somehow interchangeable, which does little to truly enlighten the user enabling them to make a rational choice for their particular needs.
Revit Technology, developers of Revit - short for revise instantly - and the first parametric building modeller, offer five basic user-based differences to distinguish between a parametric change-engine and a geometry-engine. (see adjecent panel)
However, just because a newer product by comparison appears to emerge as the younger and sexier "Brad Pitt" of the marketplace over the more mature and rugged "Robert Redford" (Jennifer Lopez over Raquel Welch for you boys out there) need not undermine the key fact that both are still beautiful and still working. They are part of a portfolio of leading players catering to different user tastes and job circumstances.
Amongst other things, this would seem to have provided strong motivation behind Autodesk's February announcement of its intention to acquire Revit Technology Corporation and the Revit product. Autodesk claim to have always supported the model-based approach to building design and management. with their Architectural Desktop (ADT) software which incorporates design objects within a familiar AutoCAD environment. With Revit, Autodesk acknowledges that their customers will be able to take "the next step toward a fully integrated model-based approach".
Geometry vs. Change Engine? How can you tell?
1) Legacy geometry-engine based products require that any significant editing be accomplished geometrically rather than through embedded relationships.
The user is expected to identify all of the geometry that is affected by a change, by selecting it with a "stretch box" or similar command, even geometry that is "not visible" or is "turned off".
2) Legacy geometry-engine based products, "extract" or "generate" plottable drawings and schedules from the model, breaking the connection from that model.
Their modelling products serve only as drafting productivity aids. Sections, interior elevations, schedules are at some point broken away from the model. The user remains responsible for making sure everything has been re-updated and coordinated, rather than the computer, if the model is changed.
3) Legacy geometry-engine based products make a point of being "object-oriented," or having "smart" or "intelligent" objects that display parametric-like behaviours.This is true to the extent that 2D or 3D geometry symbols can have data associated with them which could effect the geometry of the symbol making that data a "parameter" and thus the symbol "parametric." Also, relationships such as "hosting" were introduced between symbols, allowing a window to remain attached to a wall if the wall was moved.
What's missing is the entire network of relationships that is the core strength of a parametric change engine: recording, presenting, and managing relationships, no matter where they occur in the building.
4) Legacy geometry-engine based products usually include a "programme language" of some sort for describing objects and their behaviours and relationships.
This is because unlike a true parametric building modeller, the concept of change management and embedded relationships is not a fundamental property of the data structure and core technology of a geometry-based engine.
5) Legacy geometry-engine based products do not integrate graphic annotations into the building model.
Instead a dimension string will simply be text or it may update if you change the underlying geometry. In a change-based product, editing the dimension text will change the underlying geometry in a corresponding way. Similarly, a section key or detail callout in a geometry-engine based product will be a simple symbol with some text. In a parametric building modeller, the drawing number and sheet number will be maintained by the software as the referenced drawing is placed on a sheet or relocated during the project.
