Hi Joachim,
two years ago we were planning on implementing an Outlook integration with Petra, the OM in-house product. We looked at ways of doing it and the best way for our purposes seemed to be to buy an inexpensive third-party DLL, written in C++, that had a full .NET wrapper. It was supposed to get round all the problems a programmer faces with the integration of various versions and configurations of Outlook, all of which have their own set of functions and bugs.
Unfortunately we ran into one or two serious bugs, which the author of the DLL was not able to reproduce and therefore couldn't resolve. He claimed they were Outlook bugs, or more specific, problems with our configuration of Outlook on Exchange. We ended up spending a lot of time on it, and left it incomplete as we ran out of time for that feature and it never got implemented. Sigh!
For OpenPetra, a solution that would be GPL-compatible would be much preferred as we could then distribute it with OpenPetra. At the time when we were looking into implementing that feature there were only very basic implementations of such libraries available that weren't commercial, and they either didn't meet our needs or were unfinished/unstable, but that might have changed in the mean time.
What we learned from our experiences two years ago is that takes a lot of time to test the many configurations (e.g. on Exchange [in cached or non-cached mode], or not on Exchange) and versions Outlook comes in to be confident that the chosen approach works - and that accessing Outlook isn't as straightforward as one would hope as Outlook exposes a huge API which isn't easy to use.
Wishing you all the best in your efforts,
ChristianK
