Note, I am still very new to GTD. Before GTD I used tadalists (and nothing else!) to manage my tasks. After reading GTD, I played around with a few systems and then settled on vitalist. Vitalist lasted all but a week. The problem with most apps is that they don't fit exactly with your own needs and intuition. The intuition part is VERY important IMO. Any system should be easy and fun to use, and you shouldn't have to think about it. So based on that factor (and after doing some reading to back up my thinking), I settled on a paper based (moleskine notebook) system to manage tasks and projects.

MY SYSTEM

Moleskine - capture, action lists, project lists, 'small' project planning/brainstorming, someday/maybe lists.

Gmail - email and contacts. Caveat: I'm seeking a more flexible system for managing professional contacts. I want the ability to keep notes and manage communication history. Something like Highrise is what I need, but I hate the idea of being tied to yet another web app (at least google is established). Maybe a custom built system is what I'll settle on.

Calendar - GCal. It has it's faults, but is pretty well integrated with gmail. I'm sure there are better calendar solutions, but using less tools simplifies my life.

Project support and general reference - No external apps! I use a combination of paper and standard files (txt, rtf, spreadsheet, pdf) to manage almost all of my project support material. I love paper for brainstorming/mind mapping and planning. I've considered things like setting up an information retrieval/storage system using a database, but the complexity of my data has not reached that point (yet!).

MY ADVICE

Start with paper. Do as much as possible with paper until it becomes unmanageable. My biggest problem with high tech apps is that I feel disconnected from my work. It's easier to reorder lists, categorize tasks, and generally promote procrastination by hiding the key information under a layer of software complexity. Furthermore, paper gives the most flexibility. This is a big plus for me, since I depend on intuition within my system to facilitate getting things done.

Consider YOUR needs. Decide on what you need your system to provide. Then look for software that fits these needs. Unfortunately, most software won't fit your needs perfectly (see below.

As you can see, I dislike using software (despite the fact that I have a tech background). Software is good for certain things, but sucks at doing other ones. Current software is not even close to being complex enough to handle knowledge based work flow or personal information management. Thus, my advice would be to seek new software tools only when the complexity in your life warrants it. And commit to tools only when they fit your needs very well. Finally, think about portability of your data.