>Could you replace xp with 95? No? That's because certain programs rely on new technologies in XP. Like new theme, BITS, external desktop, text-to-speech, ...

Of course not, but then XP (NT 5.1) brings something to the table that '95 didn't: Stability. I remember beta-testing 2000: it's like 98, but has the stability of NT! That is the only reason we upgraded. And you got admit, stability is killer (after dealing with win98 for any length of time).

NT 4.0 just didn't jingle with me.

>The problem is you're making these statements based on what you saw from this Developer Preview of Longhorn. I've worked with this preview too and every time I get the same feeling as you: "where's the killer?". The difference is that I know that this is just a preview and that all the changes were under the hood. The version that we got to see was not even beta.

Aright, I'll concede this point. But, Longhorn better have something killer, because at this point, MS has solved ease-of-use (95-98) and stability (2000-XP-2003).

Server 2003 was derived from the codebase of XP, did you see the Luna style in 2003? Read again:

Sadly enough, I have it running on the servers at work (Themes service->Luna). Who says I can't sacrifice a few processors for good looks? Wink

>I don't think we will see a revolution in computers, but ms's products will always be an evolution.

Indeed. But small evolution isn't cost justifiable. I do I need to use words like ROI and TCO?

>Based on what you saw from the developer preview: yes. But when Longhorn goes RTM the whole graphical system will be changed.

Wait, it isn't RTM? Then what was that party MS has for Longhorn a while back...

>As a customer you would think this is "just another feature". It sounds true, but it is a big feature for developers.

I am a developer/network admin (lot's of hats, I've worn them all). The developer side says Longhorn will be fun with all the framework in it, the network admin said says 1.) with people still using win98, developers will still need to support the others for quite some time and 2.) with that in mind (and no immediate need to upgrade) Longhorn with its new technologies will take a while to pick up.

>If winfs is used in an application then those apps can't run on xp or 2003. If One-click-deployment is used then those apps can't run on xp or 2003. If Indigo is used then those apps can't run on xp or 2003. Those custom apps are the reason one should buy Longhorn.

Ah, but WinFS is essentially metadata (service), like SQL. So while it isn't prefect fit, you can easily make it rely on SQL (actually for enterprise apps, SQL is better). It's that whole circular thing: developers build apps for the largest number of people. If the largest number of people see no reason to upgrade to Longhorn, developers must provide applications on the platforms that the people have.

>I think it will be a bigger success than with WinXP. Microsoft is already giving very early access to Longhorn for developers. With this in mind a lot of commercial apps will be available sooner for Longhorn.

Of course, and having developers on board is the reason MS trounced IBM (IBM charged for kits, funny stuff). However, and unless developers make Longhorn their exclusive platform for release, Windows 2000/XP/2003 will be around for quite some time (as people see no reason to upgrade to Longhorn, because there's no killer).