Rich parents who pay tuition to ten different private schools don't solve the problem of where to send Junior to kindergarten. Buying vacation packages to both the Caribbean and Hawaii over New Year's leaves you with one too many options.
Hewlett-Packard just bought a piece of Taligent, guaranteeing yet another option but little guidance for its customers. On the surface, the buy makes sense. Given the disproportionately huge revenues that come from hardware sales, it is a small price to pay to place a bet on every possible operating-system combination to make sure that the winning combination in the UNIX lottery is on one of its tickets.
Further, this move gives HP compatibility with IBM customers, providing HP reps with direct access to competition for future equipment upgrades of several large OS/2 customers, according to industry observers.
Large customers tend to make their own technology decisions anyway, so why shouldn't HP be the WalMart of computing?
The danger is that HP will fail to capitalize on each of the technological advantages of its options. Operating systems need integration into complete customer strategies. Sun, for example, is serious about NEXTSTEP. It is committing a raft of resources to this option and sending clear messages to customers about its vision for the future.
HP is also serious about serving customers, but perhaps it underestimates the trouble required to integrate each of these options into complete software solutions. Most puzzling is HP's technology vision, which shows Taligent as an object layer underneath CDE. This organization puts a clunky procedural front end on Taligent objects.
HP should present one procedural and one object strategy. The procedural strategy can remain as it was before Taligent, with some support and interconnection to the object side of things.
Meanwhile, HP should use its expensive influence at Taligent to make certain that Taligent objects interoperate completely with NEXTSTEP and all other CORBA-compliant objects. Then, when Taligent finally ships its development tools (or whatever product it's going to ship), its product can be complementary to NEXTSTEP. Sure, HP will offer its customers a panoply of options, but at least those options will all work together within different portions of enterprises that might use different systems.
HP's low-end workstations are now commodity priced, but it must maintain its edge in high-end machines. A combination of object-oriented tools, such as PDO, combined with a coherent strategy linking together the various UNIX options, just might do the trick. NeXT, for its part, must nurture this important relationship. HP's stiff reply to the Sun announcement would suggest that HP and NeXT had differing impressions of the "full briefing" that NeXT claimed it provided HP in advance of the announcement.
Both the Sun and HP relationships are treasures to be managed independently, not played off against one another. NeXT must realize it doesn't have the whip hand. NEXTSTEP is strategic, but both companies could walk away from the NeXT deal and not feel any serious, damaging effects for several years.
Naturally, there are political considerations operating here. But all the companies involved need to keep in mind that the best way to succeed is to serve the customer. Serving up a series of conflicting and confusing messages is hardly the way to achieve that.
Dan Lavin comments on business issues in NeXT Ink.