Time you geeks started looking at this from the end user point of view, and the industry point of view, not from your own narrow little point of view.
"Interoperability" issues most likely start centered on Outlook+Exchange combination versus the messaging/collaboration client of choice in the Linux deployment with Exchange. Every user would touch that element of desktop. Calendar delegation, shared folders - mundane to techies, but the sorts of things every stressed EA needs to work NOW, based on even unreasonable time pressure from their execs. Those execs are paid to do a job, their workflow habits are in part based on what Outlook+Exchange *combination* is capable of. Asking people to do different is like driving a RHD car on the RHS of the road: you can do it, but it is more difficult especially under pressure. Ultimately that RHD car isn't actually cheaper to own. It's a hobbyist's dream ... and a workplace's limitation.
Before any of you say "why not replace Exchange with <insert messaging server de jour here>" have a think about what *could* be involved with that for a government *department*. Shared cross departmental services, multiple non-overlapping outsourcing deals, unique departmental security considerations. Major major business issues not just about ripping out one piece of technology for another. This is business reengineering. It isn't going to happen on the same cycle as a desktop refresh.
Next on the list in terms of how many would bump into how much... yeah probably Sharepoint, the first time they need to collaborate with someone outside their department (say another agency which has not contemplated going down the different path). Then the %^&* proliferating Access Databases. Then Visio. Then complex Excel Macros. This is simply not in their control. When one agency shares with another methods, data & analysis for chasing some crook or an active potential security threat, no one can depend on an IT guy to mess with macros to make it all hang together. Collaboration happens on a time line, it isn't a tech problem to be solved FFS get that through your heads. It's a capability that has to be *operational*.
It's not about the technology. The technology is fine, as far as it can go, OO is good for me and I am luck I don't need Visio, my workplace doesn't use Exchange. I wish it could go further, but amazinge we still have critical apps that depend on IE/Windows and no schedule to fix, because they work as is and licenses are in place :(
A whole-of-government commitment for both desktop AND server is the only one that is going to generate shareable patterns for deployment and the business-requirements-driven level of interoperability with legacy systems: what capability, for how long, and with what risk mitigation ? Which (probably smaller) national government is going to take the risk to develop such IP for the greater good of Europe (true independence from MSFT tax) ? To be the guinea pig. Which consulting and/or IT services companies have this goal in their interests ? Probably none. Why do they care whose tech they implement or manage ? The mercenaries of IT industry don't need another weapon to master. So why would they kick in in kind when the IP will be available to all competitors ? Early access to the new IP ? Doesn't mean they will have resources to do the job everywhere at the same time, it just makes for a wages explosion while competitors skill up (poach from the same pool of people for a while).
Gee I wonder why it won't ever happen successfully ? Stick to public school student desktops (eg Spain?) they have less set-in-stone dependencies on the back end services, do home desktops which can work with any number of cloudy emerging back end services. Work really long term from there ... if you want. You must bide your time like the rodent pre-mammals of the Jurassic. Someone like Apple can afford to pick and choose where and when they fight for 'the business desktop' (and they can also afford to BUY any extra pieces for their own niche walled garden of Desktop Value). And yet they mostly choose not to, going for a 'flanking' strategy via iTunes, pod, phone and pad to be 'a permanent beachhead on' rather than 'compete head on for domination with'. Alternatively you could see Oracle buy Citrix to complete their own walled garden from Desktop to Tape :) Wouldn't that just be snowball earth pleasant to live with.