It's obviously valuable to evaluate your options, but there's a point where it becomes useless. Like most other things, this just needs to be done in balance; you can spend all day arguing why OO is better than procedural and vice-versa, or whatever you're arguing about, and in the end a decision still has to be made.

It's much more useful to get out there and do something, basically anything, than it is to wax on about that thing's implications for an unreasonably long time.

Everything should be considered, but then a decision should be made pretty quickly.