Which D, 1 or 2? I have used D for several years and know a few people that also used D for a longer period of time. Consensus is that D is dead for all practical purposes (especially D2).
Also, good luck with any kind of support that is not in the interest/hobby of D developer. I once wrote a fairly large program that was for a commercial purpose (wrote as an order) - and I got stuck on a fairly large bug in standard library which I couldn't resolve myself because I have lacked the expertise to do so, I even think it was not open sourced then also (basically free() never returned memory to the OS). That bug is still open after several years. I offered money, anything basically to get this resolved since my application crashed due to it - if there weren't for helpful Tango guys I would have to rewrite the damn thing from scratch in another language.
D1/Tango had so much potential, but it got nowhere, unfortunately. I am still bitter I spent so much time and effort in D at all, just to return to C (and some C++).
D2 of course... "Consensus is that D is dead for all practical purposes (especially D2)." BS..
There is a chicken and egg situation with D... if somebody big doesn't put some backing behind it then it will go nowhere...
Tomasz wrote it better than I can right now http://h3.gd/devlog/?p=22 and I respect his view because, if nothing else, he is the man that wrote the largest code base there is written in D (for his Nucleus engine and tools). Main problem, among MANY, is that people who are working on D these days are not coding in D at all, not serious projects at least - otherwise they would see fairly quickly what is wrong with it. This topic has been covered among D community by large now, you can dig up fallout of tango team from D2 and numerous sagas alike on digital mars newsgroup.