IBM this week shored up its end of the OpenDoc technology, only to find Apple Computer Inc.'s end starting to sag.
IBM on Tuesday finally released a Win32 version of OpenDoc, for which developers have been yearning for years. However, the number of developers, at least on the Apple side, has been diminishing rapidly.
A year ago, Apple said 300 companies would release OpenDoc components in 1996. This summer, the Cupertino, Calif., company pared that number to fewer than 100. Now, Apple said it is shooting for a dozen Macintosh releases by January's MacWorld Expo in San Francisco.
IBM's OpenDoc 1.1 may win back a few developers, since it delivers on the long-promised support for Windows 95 and NT, as well as OS/2 and AIX.
But many developers have already abandoned the OpenDoc camp, discouraged and disillusioned by the consortium's failure to deliver on the glowing promises made back in 1993.
"If we made a mistake, it was [that] interest peaked prior to the technology being available,'' acknowledged Scott Hebner, IBM's manager of application development and object marketing. "Ideally, we would have wanted the technology to be available earlier.''
A Mac developer agreed that OpenDoc may have received too much attention too soon.
"Last May, with Mac OS 8 down the tubes, the focus turned to OpenDoc, and the expectations for what OpenDoc could do in a year may have been too high,'' the developer said.
The OpenDoc consortium agreed that Windows support was key, but in the summer of 1995, Novell Inc. defaulted on its original commitment to build OpenDoc for Windows and dropped out of the consortium. That left IBM with a half-finished set of code that took more than a year to complete.
Meanwhile, developers were turning their attention to the component-building potential of Java, and IBM officials were edging in that direction, too.
John M. Thompson, senior vice president and group executive of IBM's Software Group, earlier this fall distanced himself from OpenDoc, saying, "My own belief is that Java has far more appeal and is more likely to be the base for development.''
Even OpenDoc 1.1 fails to keep all the original promises. IBM's product does not support the Open Scripting Architecture, one of the key elements of Apple's version of OpenDoc. Instead, it uses technology based on Lotus Development Corp.'s LotusScript to support Microsoft Corp.'s ActiveX and OLE Automation.
Hebner said IBM's priority was to build something that would appeal to Windows developers first rather than stick strictly to Apple's implementation.
Another problem that OpenDoc faces is the lack of container applications. Components are common among small companies, but so far no major ISV has released a container or enabled an existing application as a container.
Claris Corp., of Santa Clara, Calif., initially said it would ship an OpenDoc version of ClarisWorks by the end of the year. The company now says the software will not ship until next summer.
"I expected there would be containers by now, but they haven't materialized,'' one OpenDoc developer said. "It's tough to make the case right now for OpenDoc parts because the containers aren't there.''
Hebner said IBM's OpenDoc 1.1 includes a set of utilities that lets developers turn existing applications into OpenDoc containers and lets Netscape Communications Corp. browsers become containers.
But similar efforts, such as Apple's CA Lib, have not impressed developers. Claris used CA Lib in its OpenDoc work but had to rewrite 80 percent of it.
Apple, however, is working on another conversion utility, OpenDoc Internet Adapter, which the company said is easier to use than CA Lib.
Hebner, who insists that the industry is only at the beginning of the adoption curve for component-based software, says OpenDoc is still viable.
"I don't think anyone should be panicking about the future of OpenDoc,'' he said. "There is no overwhelming adoption of component software right now. It is just not OpenDoc. The cycle for components and objects is just now ramping up.''