Just because you can, doesn't mean you should. I caught wind of this twitter the other day that encapsulated perfectly the kind of fear I have about getting cornered (by someone else, or my intention) into making a "community site" that seems to beg for drupal.
For the not-as-geeky among you, feel free to skip to the end where I hope to extrapolate on the larger lessons about audience and technology users.
Let's get into it, shall we:
Drupal is great software. There's a lot of developer energy, there are a lot of tools, and by god this is a program that can do everything. You want to host a community of bloggers? Done. You want a discussion forum? Done. You want wikis and/or collaborative tools that you can use to write structured documents? Done. You want to tie into other popular services? Done. You want to create custom content views and insulate certain kinds of content with very granular access menus? Done.
And so forth. Oh, and by the way it's free software/open source software to boot. Who can argue with that? Certainly not me. Though I wouldn't consider myself a drupal master, and I don't use it on any of my sites, I've played around with it a bit, and while I liked what I saw, there were a couple of major problems for me.
The administrative interface was complicated and non-intuitive.
To be fair, I really don't like to interact with programs through a web browser, and it seems to me that all I was doing with drupal was clicking through page after page to try and find the one bit that I needed to flip, but I'm hypersensative to these things. The cause of this, as near as I could tell was the fact that the admin interface and the front end of the site are differentiated by the users access level not anything in the interface. This is confusing and not, I think good form, visitors to the websites and admins need different things.
Drupal sites reek of drupal.
Most web programing frameworks reek of themesleves, so this doesn't set Drupal apart, but I think it's particularly bad in the case of Drupal. What do I mean like reek? Well, it's that feeling of "oh they're running drupal/rails/etc," here that you get 7 seconds after clicking on your first link (if not sooner). While this isn't a bad thing, and I suspect most users don't notice this very much, but I think if most independent site runners have sites that seem to work the same, the reader will be left with the impression that the sites are the same. We must think about total experience.
The level of programing know-how needed for setup is non-trivial
While I think Wordpress, which I use a lot and rather like, has gone too far in the other direction,  the fact that it's so much more complicated to get drupal from the box to production, means that people who want to start a site now need to break momentum to do programing (or hire someone.) The truth is that I think the know how needed to really use drupal, is like 10% less than the know how needed to build something in Django. Maybe I'm overplaying it, but not too much.
To be fair, I've not used the 6.x series of Drupal, but I think there's a few broader lessons about creating a product/content and audience, that comes from that nagging dread that drupal seems to create:
- It's not clear that drupal knows weather it wants to be a web framework or a CMS. Know what you're trying to do, and be clear about it.
- Different tools for different jobs, rather than one swiss-army knife. It's good to have related project that are integrated, of course but that doesn't mean that you need to only have one program for everything in your project. 
- Create low barriers to entry. I didn't stay with drupal because after a weeks of tinkering with it, I still didn't like what I was seeing, so First impressions count.
|||there are templating features and capabilities that were around in b2/cafelog that have been abstracted and removed, and while I understand the rationale, I still miss them.|
|||I'm such a unix dweeb, so your milage may vary.|