OpenDoc vs. Network Ole: The New Internet Wars

Just when you thought you had it all figured out, along
comes another zinger to mess up your Internet planning (in 1995).

Francis Vale

The internal use of the 'Net + Web and their standardized client/server protocols within your organization offers many attractions. Why reinvent the wheel over and over again, or be locked into a single desktop OS vendor, when you have a portable, highly consistent, client/server approach that can encompass all of your current and future systems?

But given today's volatile rate of technological change, an organizational decision favorable for the 'Net + Web might mean that Microsoft is the big loser. This point is not lost on major users. Steve Heit, the Web site development manager at UPS, notes that this potentially major loss is much on the minds of many in his organization. And right into the middle of these concerns steps a new marketing war that may significantly shape corporate IT/IS decision making within UPS, and elsewhere.

This results of this marketing brouhaha may help block the fall of the Microsoft Client King, and thwart the budding ascendancy of the Web princes. The sound and the fury centers around the recent dash to componentware. This new technology offers a means to recast specialized application functions as easily plug-in, plug out, distributed object modules.

With widely available network componentware, the business case will go away for buying huge software application monoliths, which, typically, possess features beyond most users' requirements. E.g., application suites. Rather than buy a bloated set of tools, componentware users will be able to quickly assemble customized applications that suit their ad hoc needs.

There are now two competing network componentware technologies vying for market acceptance. The technology which emerges victorious will likely have a big impact -- for good or for ill -- on the future development of Web-based client systems.

One of the big contenders is OpenDoc, a specification backed by Apple IBM, Novell, and many others. Given the system's numerous backers, it is no wonder that OpenDoc implicitly supports the linkage of distributed objects across heterogeneous systems. IBM's system object model (SOM) plays a key role in OpenDoc. Apple is preparing its own Internet-enabled application variant of OpenDoc, code-named 'Cyberdog.'

The other main contender is Microsoft's Network OLE. The newly announced Network OLE spec draws heavily on Digital Equipment's Common Object Model (COM). Microsoft's spec obviously outlines a means for extending its products/business model across a network. Significantly, the heterogeneity so fundamental to the 'Net and the Web is not crucial, nor probably even beneficial, for the working of MS Network OLE. A closed, proprietary network would do the job for Network OLE just as well as an open one.

In fact, it might even be argued that a proprietary network would work even better for Network OLE because it could be optimized for Microsoft's systems. (Some may recall that this particular scenario was already played out once before -- It was called VMS and DECNet.) MS has said that its next version of NT 4.0 beta, due now in February/March 1996, will include Network OLE.

Regardless, Microsoft is also making moves to make Network OLE 'Open'. Microsoft has recently tagged Software AG to make OLE cross platform capable; thereby supplanting Digital's developer role. [This assumption of OLE development duties by Software AG is a sad commentary on Digital's ability to keep technical pace even with its own software.]

Software AG plans to support its Network OLE variant on 18 different platforms, including OS/400, MVS, and several varieties of UNIX, in the first half of 1997. The question then becomes if Microsoft will let Software AG do its own marketing thing; or will Microsoft stick to its tried and true ways, and actively discourage the use of non-MS operating system OLE implementations?

If OpenDoc should win the day, the heterogeneous 'Net might well become the preferred means to deliver and assemble these mixed-vendor objects. And the most viable means for doing OpenDoc assembly at the desktop might then also become the OS independent Web client. It is unlikely that either Apple or IBM are thinking in Web/OpenDoc client terms, wedded as they are, like Microsoft, to their own operating systems. But events have shown that a new technological force, once unleashed, usually follows its own paradigm, and the sputtering vendors be damned.

Only time will tell as to who the winner is of this all important new battle for the distributed desktop client. But given the tremendous popularity of the 'Net and the Web, current trends may favor market acceptance of the more 'open' system; i.e., OpenDoc.

As Heit of UPS noted, "The outcome of events over the next twelve months will tell us a lot."

Back to main article: Net + Web Business Model

Copyright 1995, Francis Vale, All Rights Reserved


21st, The VXM Network,