It seems that the Common Open Software Environment firms have pretty much worked out between themselves how far they think they can go in technology terms, and the talking now concerns the more difficult task of how to make it all happen, according to Hewlett-Packard Co’s European Computer Systems Marketing Manager Bernard Guidon. COSE’s Xhibition debut a few weeks ago caused more than a little of consternation among users who felt they had been left out of the discussions about what vendors have in store for them, but Guidon believes that as the effort progresses, COSE will increasingly fall into the shadow of X/Open, and that’s where users should participate. The common desktop environment specifications should put X/Open Co Ltd back into the frame, he says. As he sees it, X/Open has been unable to resolve the long-term question of a common or standard interface and that it has taken an initiative like COSE to create the momentum required to bring these issues to book. Because COSE has no organisation and therefore no marketing presence as such, X/Open is expected to become the marketing interface for COSE specifications.
We’re 90% there
Guidon believes Hewlett-Packard will be first out with a COSE-compliant desktop product – by the end of the year because with Motif and the Visual User Environment, we’re 90% there already, he claims. IBM Corp will follow soon after, he believes, with Sun Microsystems Inc having probably the most to do to bring its Open Look-based interface technology into line. Many of the basic technology issues that COSE tried to settle in its early days were highly emotive issues for Unix System Laboratories Inc while it was still with AT&T Co. As soon as the planned acquisition by Novell Inc was anounced, we were able to proceed rationally, says Guidon. He maintains that COSE will not address the issue of a common Unix kernel per se, but agrees that the adoption of a range of kernel strategies by vendors has restricted the ability of users to integrate different systems, ultimately slowing the uptake and sales of new technology. Application programming interfaces must be the way forward to unifying Unix, Guidon believes, as existing kernel technology developments are now too entrenched – or the issues still too emotive – to be settled any other way. No-one is working on unifaction of the kernel, he says, and in any case the kernel by itself isn’t important to users, he argues. Hewlett-Packard’s kernel strategy for its own HP-UX Unix implementation remains something of an unknown. Although the firm has indicated its intent to offer more compliance with Unix Labs’s Unix System V.4, that shouldn’t be taken as an indication of any intent to move the HP-UX kernel in that direction. We’re not saying we’re using the Open Software Foundation kernel, or the Unix Labs kernel. We want the best technologies to be able to deliver what customers want. Even if we find a technology in Windows NT then we’ll put it in there [HP-UX]. If no-one can offer what we want then we’ll do it ourselves or try and find third parties to do it. Choice of kernel is nevertheless a critical element in any operating system strategy, admits Guidon. To develop it and put it on hardware takes years of work and millions of dollars. You make that decision once every 10 years and you don’t make it lightly, because you don’t want to change it after two years. Consequently we’re making that decision very carefully. Hewlett-Packard has great momentum at the moment, argues Guidon, but developing on or for a new kernel would mean that resources would have to be taken away from maintaining that momentum, and that’s something the company does not want to do. On the other hand the firm obviously doesn’t want to continue with the same version of HP-UX for the next 10 years, and consequently the timing is crucial in terms of momentum and internal development.
By William Fellows
Guidon says Hewlett-Packard is also evaluating whether to go with a microkernel technology for the next generation HP-UX system, or an object-oriented sy
stem. That’s the decision we’re trying to make at the moment – it hasn’t been made yet, he says. Hewlett-Packard is doing object development work with IBM and is looking at various microkernel systems from a business and technical perspective, but Guidon says it’s far too early yet to think about going with the IBM-Apple Computer Inc Taligent Inc opeating system for instance. Hewlett-Packard’s problem is that while users aren’t too concerned about that decision at the moment, it is nevertheless one that Hewlett-Packard has to make. It is therefore certainly not these kinds of long-term technology plans that will enable Hewlett-Packard to maintain its current momentum in the short term. The key to maintaining momentum, Guidon believes, lies in being able to create the right channels of distribution to match customer needs with products, learning from other industry models where necessary. The crucial element that determines the success or otherwise of being able to maintain growth is what Guidon calls Hewlett-Packard’s win/loss ratio: the firm’s ability to win new customers against the competition, something that’s definitely restricted to its Unix business. Guidon believes Hewlett-Packard has the products necessary to win the business on which it competes in most situations. Where it doesn’t win, he argues, is because the company doesn’t have enough feet on the street – OEM customers, resellers, distributors, value-added resellers or direct sales people – to be able to talk to everyone. When we do compete, we win, says Guidon, but we’re not able to compete enough because we don’t have the channels. Some vendors compete everywhere but don’t win enough. They need to increase their win/loss ratio. We have a fantastic win/loss ratio but don’t compete enough. From a technology point of view, however, we’re fine. The pressure is most definitely on Hewlett-Packard to sustain its recent momentum. To broaden its channel exposure Guidon plans to sign a slew of resellers and partners and expects to see more application providers and integrators offering their wares on Hewlett kit – telesales and telemarketing operations will be ramped up too. The company will increasingly focus on what it perceives as the largest sector to which its win/loss model has yet to be applied – the AS/400 market, where it has recently signalled its intent, striking a deal with Synon Corp. Hewlett-Packard’s recent agreement to put the NeXT Computer Inc NeXTStep environment up on its systems for the financial and banking markets is a good illustration of the way forward, says Guidon.
That’s open systems
NeXT has done well in specific financial accounts, outside that niche many users don’t know or care about NeXTStep. Steve Jobs wanted out of the hardware business, we wanted to sell into those markets. So it was win-win. It also makes sense, argues Guidon, for Hewlett-Packard to become an integrator for what he perceives to be the current object-oriented market leader. There’s a lot we can learn from the relationship with Jobs, and we get a new set of customers. The important aspect is not the return on the investment we are able to get from an accounting point of view, but the exposure it gives us to new technologies plus the niche accounts. The relationship with Next is not, however, a strategic technology development partnership like the deal with IBM, Guidon emphasises. Guidon doesn’t expect Windows NT to become a dominant force or threaten Hewlett-Packard any time soon. Microsoft wants to take on the world with NT, the marketing was unbelievable. So NT is a desktop, a client-server system or a local network server. One size fits all. I don’t believe it can be successful from the start, because customers must be confused, they don’t know what it is good at and what it is not. It would have been better to say this is a high-end desktop and a year later say it is also personal computer network server. One step at a time. Partners don’t know what to do with it and MS-DOS and mainframe users may have it but don’t know what to do with it. Mea
nwhile, Hewlett-Packard continues to strike alliances with and at the same time fight against the same adverseries. That’s open systems, argues Guidon, at the same time you are a customer of each other, you have co-operative agreements and you are a competitor too.
