I've been on the arse end of an XP migration over the last couple of years. It was an unmitigated disaster although the PR for the event says otherwise. Typica

I've been on the arse end of an XP migration over the last couple of years. It was an unmitigated disaster although the PR for the event says otherwise. Typically a lot of the ActiveX vendors that were relied on to prop up the browser functionality that had previously been migrated from IE5/Win2000 are gone. There is no source code so the only result is to hire in contractors to rewrite/replace functionality from scratch as a winforms app (as browsers still can't do this sort of stuff, Silverlight is canned and the contractors don't know WPF). This is not a bad thing but it has a cost. Add to that, the server side, a completely insecure hodge-podge of ASP that only worked on Windows 2003 and VB6 COM components. This needed scrapping and starting again. Typically they did this with contractors who just about knew ASP.Net web forms, some WCF and a bit of SQL so we're talking circa 2005-2008 tech for the rewrite. Then there's the hardware which had 10baseT networks so the entire core switches were replaced, all the cabling was replaced and the workstations were all replaced, 25 servers were replaced, the network appliances were replaced and a SAN was added. Ten years down the line, a rewrite will be due and it will require porting to Windows X (10) and all the tech will have changed again. So £15.5m every decade for this platform + maintenance on top. That's a lot of bottom line out of the window for a tech choice. At least with other platforms, incremental change is a hell of a lot easier and cheaper (Java EE / generic Unix).