Microsoft senior VP for Windows Brian Valentine has brought a certain pleasing symmetry to Longhorn roadmapping by telling the good people at IDG that there may well be a server version of Windows Longhorn after all. This, says Computerworld, "despite claims to the contrary."

Indeed, we have one of these claims quite handy, made last November by, er, Microsoft senior VP for Windows Brian Valentine. Actually Valentine doesn't seem to have been trying to convince Computerworld that Microsoft never suggested there would be no server Longhorn in the first place; he describes Microsoft's statements to this effect as being "a bit premature," saying that Longhorn will require additional services at the server end, and that it has yet to be decided how these will be packaged and delivered.

Which is almost entirely consistent with what he was saying in November, the only difference being that Microsoft is no longer ruling out the possibility of shipping a Longhorn server, aka Windows Server 2005 in order to deliver this extra functionality. The alternative, no Longhorn server solution, would mean shipping some form of 'service pack from hell,' so the decision in part will depend on which one of these routes Microsoft decides is less painful.

Another consideration Valentine covers with Computerworld is the cost to customers of new server OS deployments, and therefore the difficulties inherent in releasing them too close together. He figures the cycle can't be any faster than two to three years (which will still seem pretty fast to the luckless server purchaser), but note that 2005 is dangerously close to the two year end of the scale. At the other end of the scale this starts to run into Blackcomb, which will need a new server version, and which is scheduled for 2007-2008.

And you thought world domination was easy? ®