Liam - We'd need to know a bit more to be able to offer advice.
I can tell you some things that matter to me, and that might help you.
1. I rejected mbed for several reasons including your online "cloud" compiler. I help folks do robotics, and it can be a real problem to access a "cloud" compiler at a robotics event.
2. I want the flexibility to be able to use low-cost devices. Boards that rely on C# or Java need too many resources to scale down. I'm happy to use them, I just don't want them to constrain the cost of device I use. Ideally, I'd have a range of programming technology available on the boards I use because, IMHO, one size is a poor fit on most.
3. I teach and help folks make open-ended projects. IMHO it is more exciting for them to know that, if they really wanted to, they could make the PCB and electronics themselves. That rules out boards based on proprietary designs, which is quite a lot of boards. I look for Open Source hardware.
4. I work outside schools, and I want folks to be able to get the software development tools at a low-cost/free without needing to be "in education". That rules out many tools which are either proprietary and too costly, or only available at a low-cost to people in full time education at a reasonable price. I look for Open Source software.
5. I introduce microcontroller and programming technology, and hope that folks will take it in directions which interests them. I don't want to use something that requires them to learn assembler, or understand deep hardware details, to use the microcontroller for reasonable things. That rules out a whole bunch of processors, including ARM7. I don't know enough about ColdFire to be definitive. Cortex-M3 is good.
6. I have been using Arduino's. I understand pretty well what I can make them do, and also I understand that I need quite a lot more computer power for some of my projects. I am happy to use Arduinos's and Maple's. The fact that the tool chain underlying the two programming environments is very similar (gcc), and there is commonality in the libraries, makes it feasible for me to stay sane while using two different boards with different microcontroller technology. Most people obsess about choosing the 'perfect one'. I don't. I try to focus on having choices that give me a larger range of options than any single board, or microcontroller family, could possibly provide.
So what is it you want to do?
Edit (I thought I'd add the list of my core requirements):
0. Must be very easy to use. Preferably simple enough for a year six old to program and use. Must be simple enough for an eleven year old in a classroom environment, or at home, by themselves.
0.1 Hardware is just the board and a USB cable, no programmers, no expensive cables. It must be easy to wire up a quick prototype. The board should clearly document the connection between the physical world and software, because that's what the board is! The board may present the common and popular subset of its features to avoid muddle.
0.2 Software must have simple, reliable, robust 'one button' upload.
0.3 Programming must be in a high level language, which is either industrially relevant, like C/C++, with very well designed and implemented libraries, or something much easier for beginners like Scratch.
0.4 No flow-charting languages. Flow-charts were broken in the '70's, they are still broken. IMHO they have the wrong properties for programming. Scratch is very different from flow-charts. Scratch has normal programming semantics but with reduced syntax complexity.
