A new role for OLE.

We never quite understood all the fuss about Microsoft's OLE ("object linking and embedding") technology, perhaps because most commercial applications of OLE have been fairly trivial. But a few weeks ago we paid a visit to Redmond, where we got some eye-opening briefings on a quiet but profoundly important shift in Microsoft's positioning of OLE. In essence, OLE is emerging--along with Windows NT--as the underpinning of Microsoft's enterprise and client-server strategy. That new role puts OLE in head-to-head competition with the biggest and most entrenched players in the high-end market. But the odds are pretty good, we think, that Microsoft will end up extending its desktop dominance into the corporate applications development environment as well. If that happens, OLE may come to define a whole new set of standards for interoperability and interface design. And, just as Windows and Office compatibility have turned into de facto standards for the rest of the PC software industry, we suspect OLE will soon become a test of admission to much of the enterprise market.

To be sure, OLE's success is hardly a done deal. But the repositioning of OLE raises some interesting issues that we're watching closely:

How does a desktop-oriented technology like OLE suddenly become an enterprise architecture?

In its earlier incarnation, OLE was simply a pipeline that let independent, off-the-shelf applications talk to each other. Now, it's become a specification for designing components that end users can assemble into complex, customizeable applications. Microsoft envisions a world in which enterprise solutions are assembled out of a wide variety of pre-built objects (some from commercial developers, others created by users themselves). All of these objects will communicate through common APIs, their behavior and appearance will be consistent, and they will support industrial-strength transaction processing and distributed computing.

Is OLE really ready for prime time?

As of today, OLE has critical shortcomings in such areas as network support and object inheritance. But Microsoft has been recruiting top programming talent and simultaneously evangelizing third-party component developers. Many of the earlier component and object contenders--NeXt, Taligent, OpenDoc--have failed to deliver timely, well-supported solutions. Our guess is that Microsoft won't make this mistake.

Won't OLE essentially force users to adopt a Microsoft-specific standard?

OLE APIs are open, and Microsoft has lined up a few partners (notably Software AG) to port OLE over to Unix, MVS, and OS/400 environments. But the truth is that OLE remains closely tied to Windows NT, Microsoft's only high-end platform. If customers haven't first made a pretty strong commitment to NT, Microsoft is going to have a tough time selling OLE as a development architecture.

However, NT has made impressive gains in the high-end market, while these days rival platforms like OS/2, Unix, and Netware seem to be losing momentum and direction. So finding customers who have bought into NT is less of an obstacle than it once was.

Doesn't OpenDoc provide much of the same component architecture, but through a more open environment?

Right now, it's difficult to take OpenDoc seriously. Key members of the OpenDoc consortium have been conspicuously reluctant to incorporate the technology into their own products, and the rationale for OpenDoc increasingly sounds like the "Stop Microsoft" rhetoric that we heard from the anti-Windows crowd a few years ago. Anti-Redmond politicking didn't make OS/2 a success then, and we doubt that it will save OpenDoc now.

What about Microsoft's notorious lack of experience with enterprise customers.

That's an out-of-date story. These days, Microsoft has a very respectable infrastructure in place for support, consulting, developer training, compatibility testing, and the like. There's now a dedicated Enterprise Customer Unit sales force and a network of third-party "solution providers." Maybe we've drifted too far into the Microsoft reality distortion field, but we're especially struck by the sense of mission these issues inspire among Microsoft people. If that enthusiasm is even slightly infectious, OLE customers aren't going to have too many doubts about Microsoft's commitment to the enterprise market.