Commentary — Opposite to Phil Windley, I’d tend to trust Steve Ballmer: Longhorn will make rich interfaces more readily available through network-based applications.
Quoting the ZDNet story:
So next time you hear Steve Ballmer telling you that rich clients are important and you’ve got to have Longhorn to make them work in your enterprise, only believe half of what you hear.
On the contrary: Longhorn has great potential to let developers build rich network-based applications, compared to today’s gadgets.
Today, we have Citrix-based solutions to get on-demand applications, but this requires a decent bandwidth. And a few gadgets such as the Webservice HTC, HTAs, and XUL to build network-based applications that are richer and more streamlined than simple web forms. None of these gadgets, however, look even remotely credible when compared to XAML.
In many ways, the purpose of XUL is to do more or less the same thing as XAML. But then, XUL lacks a decent IDE, uses bitmaps rather than vectors, is borderline when it comes to implementing an application that uses SOAP calls — consider comparing the .Net way and the XUL way to call web services using SOAP — and requires a Mozilla-based browser.
You might argue that Avalon and Indigo are X-Windows reinvented. And I think you would be right, in some respects. Yet, I do not recall X-Windows is service-oriented, XML serializable and running on top of a semantic file system.
In the end, I’d be quite amazed if Linux, its pathetic usability and its dinosaur-age technology choices survive Longhorn. Let’s hope Haiku or a similar project will be finished and that it will have gained more support by the time Longhorn delivers.