Just saying that the web interface on Microsoft platform doesn't work, because ADO is blocked at firewall level, ActiveX being blocked. I've never used ActiveSync, because I don't see what added functionality it has to offer, and I don't believe in protocols that can be unilaterally changed because they don't follow the open source principles. Look at the twists and turns SuSE has to pull off to keep compatibility with EAD.
There are many solutions for groupware application linking. A convenient one because it does everything, and it ties in to the Kolab backend as well as SQL, is Kronolith. It's just an example, there are at least a dozen similar solutions, all open source, and there are even more open source based but proprietary solutions. It would also work just fine on a very small power device though, it's not resource heavy at all. It allows shared calendars with easy rights management, and supports mobile as well as desktop specificities. In my opinion, most users of the site wouldn't need the functionality of shared calendars. Kronolith is the CalDAV module developed for the Horde framework, which also offers a 5-minute install for people that need the extra compatibility/functionality. It's just an example. Enterprises that have a need for this kind of installs, usually get full custom solutions, and luckily, there is noone in the business that wants to put how-tos for that on the web, because that's where the money comes from, in return for the quality of a real custom solution of course. I just think that it's very hard to provide a how-to on extended configs that will be safe to use by everyone that just follows the how-to, because every environment is different, and the reason why these open source applications are the way they are, is because custom tailormade configs are the norm in open source, and they work really well. Some scenarios will make good use of ActiveSync or EAD, others won't, but in my opinion, those environments where this is even an issue, should not roll out their groupware solution based on a how-to using bog standard quick-install packages. Kolab is also a commercial entity, they make specific things happen for customers, they have the technology and the knowledge to implement anything, there is a good reason why the default Kolab install is limited to certain features, with the option to extend the install depending on specific needs, but no two deployments that require these extras, will be the same, so they're not preconfigured or comprised in the standard 5-minute install.
I don't know what it's like in the US, but over here, most companies have not evolved beyond OL2003. Even the German government is paying for extended support for MS-Office 2003 and XP, with the view of migrating away from it, and that's just it, ever less people are using recent Office products, with the official publications and correspondence being in Open Document format. Google and Microsoft do everything they can to keep people away from ODF, but it's just not working, and in the end, however crappy LibreOffice may be at times (but not all office applications are equally crappy, again, there is plenty of choice in open source for every application), there is no reason to use Microsoft products any more, let alone Microsoft protocols.
Things like clamav are also extras, if there are no Windows clients on the network, there is no real risk. Again, this is a thing that would be useful for large hybrid enterprise environments, where a custom roll-out would definitely be the way to go in my opinion.
If people would just know how to roll out Yunohost (which is Debian based) on a RasPi, the possibilities they would discover would already be amazing in my opinion, and the last thing they would be worrying about would be whether or not it's AS or CalDAV based, and whether or not it's EAD compatible or not. The main feature of Yunohost is that it is a pretty good solution already, easy to deploy, and that it is truly a zero-cost solution, with no vps costs or high energy costs, with the exception of a 40 USD PC-on-a-stick or RasPi, that would then last at least 10 years if not more, and could literally run of a few AA batteries. Even for SOHO's, 99% of the added value comes from that zero-cost 5-minute solution, and after that, to add value it starts to proportionally become really expensive, to the point that - in my opinion - it's overkill for most users of this forum, and more an enterprise thing.