Cherish Development Choice

Opinion: Innovative energy will flow where developers can create with the least constraint.

Olympic speed skater Chad Hedrick, one gold medal in hand and others likely to follow, reminds us that talent first seen on one platform may find its best outlet somewhere else. Raised on roller skates, Hedrick only started skating on ice three years ago—but his unorthodox technique propelled him to an all-around championship and new world-record achievement just 14 months later.

I thought about Hedrick while trying to decide how I feel about Borlands announcement last week that it will seek to divest its line of integrated development environments. For many years, Borlands development tools have seemed to me like great talent that was always happy to take advantage of any stage where it could shine. When DOS on the IBM-standard PC was clearly the place to be, Turbo Pascal and even Turbo Prolog shook up peoples basic ideas of how development should be done; when Windows and Java created new opportunity for developers, Borlands Delphi and JBuilder each offered developers brilliantly crafted visions of how to turn that potential into products.

I spoke last week with Borland CEO Tod Nielsen, who told me that his priority in the divestiture is "to make sure that the community is taken care of—that the NewCo is investing in that community," he said, using the "NewCo" placeholder that Borland executives have coined to let them talk about the future home of the tools line before they know what that entity will be called. Thats a good thing to hear, but I hope that "the community" doesnt wind up being more narrowly defined by those tools new owner. Borland has taken risks, at least from time to time, with projects like the Linux-based Delphi/C++ toolkit Kylix—now a "Borland Classic" product receiving no further development or support—as well as useful if somewhat tenuous support for JBuilder on Apples OS X. I hope that NewCo will be at least as daring.

Similar thoughts, but with a less optimistic spin, came to mind upon the news that Skypes new IP-network voice conferencing services will require Intel Centrino Duo, Pentium D, Pentium Extreme Edition or Intel Viiv machinery to use. This doesnt seem to me like a gesture of confidence from either player. If Intel thought that it could deliver a superior platform, it wouldnt need to engage a service provider in an exclusive arrangement, any more than AMD had to do a deal with Microsoft to make AMDs 64-bit x86 extensions the reference standard for 64-bit x86 Windows. If Skype thought that its services were clearly superior, and going to stay that way, it would offer them on every platform under the sun.

In the current environment of top-tier developer tools (like NetBeans 5.0, which I review in eWEEK this week) and delivery options (like AJAX, powerful although not a panacea) available at virtually zero cost, developers are able to pick their targets with precision and become the worlds best at what they do—then invite any and all potential partners to collaborate with them in doing it. I like that model better than one in which dominant players try to make up for lack of agility by linking arms to form barriers to their competitors.

I hope that developers will agree with me in that preference, and choose paths that lead toward adding technical value rather than reducing marketplace risk.

Peter Coffee can be reached at


Check out eWEEK.coms for the latest news, reviews and analysis in programming environments and developer tools.