OLE AS A DE FACTO 3-D STANDARD

Intergraph and Microsoft on March 6 announced Object Linking and Embedding extensions for Design and Modeling (OLE D&M). The specification, originally developed by Intergraph, has been placed in the public domain for other vendors to adopt. While the first release concentrates on 3-D compound document functionality, it supports a framework that potentially addresses at least some of the end-to-end enterprise requirements for product data. The blessings of the PC desktop leaders, Autodesk and Microsoft, confer instant status as a de facto standard. Other companies that have endorsed the strategy include ANSYS, Inc., Cadence, SDRC, and Spatial Technology Corporation. Given such heavyweight backing, this initiative may irreversibly and radically change the landscape in the MCAE and CAD/CAM industries over the long term.

Hitherto, NT has represented a slowly maturing alternative to UNIX for engineering workstations, without any dramatic functional advantages to tip the scales.

OLE D&M significantly tips the balance in favor of NT for 3- D on the technical desktop over the long term.

In contrast to standards such as IGES and STEP, OLE D&M focuses on data sharing instead of data interchange. The traditional strategy of data interchange involves translation of foreign CAD data into native CAD formats so that all CAD information within a single model has a homogeneous representation.

In contrast, OLE D&M involves "re-embedding" foreign CAD data as software "objects" in a "container," which can be any CAD database that supports the standard. Unlike translated CAD data, embedded objects retain their native CAD representations. OLE D&M acts as a "wrapper" that allows a CAD system to manipulate the CAD objects without knowing their internal structures. Today these extensions support translation, rotation, and the ability to reference the surfaces of objects from foreign CAD systems. Additionally, since OLE already works with office productivity applications such as Microsoft Word, Powerpoint, and Excel, these desktop applications may also inter-operate with CAD objects. For example, a CAD model can be embedded, accessed, and modified in a Word, Powerpoint, or Excel document. Alternatively, Excel spreadsheets or Word documents can be embedded in CAD files. Changes in the spreadsheet or the document will automatically update the CAD file, and vice versa.

At first glance, this strategy seems extremely appealing because of the potential benefits of object oriented programming, particularly the reuse of component software. Ideally, users will be able to pick and choose among software "components" that best fit their mechanical application needs without writing additional software for application integration and without performing data translations. Research by D.H. Brown Associates over the past five years continuously reinforces the market need for best- of-class software in MCAE, CAD/CAM that addresses a breadth and depth of functionality that no single vendor can provide. A successful OLE D&M strategy would allow users to select and operate among applications using OLE in the same way that OLE supported office productivity tools work.

The strategy also leverages the familiarity of Windows-based applications for the engineering desktop. Training requirements, a traditional headache in the technical environment, can be alleviated considerably. Users understand Windows; its look-and-feel, commands, and paradigms for use. Additionally, OLE D&M leverages software sanctioned by Microsoft such as learning aids Q Cue Cards, Tutorials, and Wizards.

Despite the tremendous potential, alas, the MCAE, CAD/CAM community must wait at least three to five years for the full development of the potential. Major concerns relate to the formative stage of development, performance and scalability.

The current extensions allow users to select, reference, translate, and rotate objects. CAD companies can easily agree on such extensions since they involve a relatively uniform and obvious set of technical requirements. For specific applications such as technical documentation, OLE D&M provides an immediate benefit. Changes in CAD designs can be automatically captured in associated documents. Without OLE D&M, tracking changes must be performed manually.

While this level of functionality supports the ability to insert objects into documents or even build assemblies with constraints, it falls dreadfully short of the functionality required to address the mechanical design arena. For example, users cannot change the geometry of parts without accessing functions of the CAD system that created them. Considering an AutoCAD Designer solid object that has been inserted into a Microstation Modeler database, a "double click" on the AutoCAD Designer solid accesses AutoCAD commands for editing. Microstation users who do not know AutoCAD will find this inconvenient. Ideally, CAD users could modify geometry without knowing the core CAD system that created it. However, definition of extensions that make sense and do not have diminishing returns will be challenging. Data interchange standards such as STEP and IGES avoid this problem by converting foreign CAD data into strictly defined formats.

From the perspective of systems architecture, OLE D&M must be extended significantly further to address features and constraints, also a deficiency currently shared by STEP. OLE D&M requires such feature representations to take full advantage of manufacturing and engineering analysis applications. For example, in numerically-controlled milling, the generation of efficient cutting paths depends on an ability to reference the collection of surfaces that comprise geometric features such as pockets or islands within pockets. In finite element analysis, OLE D&M must accommodate features such as fillets for accurate assessment of stress concentrations and midsurfaces of shells for accurate modeling of thin-walled structures.

Thus, the hard work now begins since enthusiasm and cooperation among the CAD vendors could wane as OLE D&M extensions touch the product core and the principal source of competitive differentiation. In the introduction to the OLE D&M design specification, Microsoft encourages the MCAE, CAD/CAM industry to take "ownership of the specification and move it forward." Indeed, Microsoft, combined with a small, focused alliance of strategic and powerful CAD vendors, could drive the specification and "force" its acceptance well ahead of the pace of de jure standards such as STEP, which has been continually delayed in committees that must negotiate an industry consensus. Even with Microsoft's power in the desktop market, however, OLE D&M may realistically require at least four years to achieve STEP's level of maturity. End users cannot tolerate additional delays given today's intractable problems sharing CAD data. Consequently, CAD vendors can neither abandon their STEP efforts nor can they ignore OLE D&M.

Performance and scalability of OLE D&M functionality also pose concerns. While OLE works nicely for sharing objects among office productivity tools such as Excel, Powerpoint, and Word on Intel hardware, it will be rigorously tested in the mechanical environment given the typical size and complexity of designs and applications. To a degree, the continued dramatic explosion in hardware performance will likely alleviate at least some of these concerns. Even so, OLE imposes significant performance overhead with its layering approach that leaves considerable room for optimization and alternative approaches.

Finally, OLE D&M represents a remarkable turn-around for Intergraph's posturing on openness. As recently as 1991, Intergraph was repeatedly chided for proprietary hardware/software solutions. Along with Microsoft, Intergraph now challenges the rest of the industry to match its level of openness.