To add to Chris's comment, with which I mostly agree, I'd expand on his comment about the "UTXO" model.
In our view (I'm the CTO of R3), one key design decision which distinguishes blockchain platforms is whether they use what I call the "replicated virtual machine" metaphor or whether they are based on the idea of transaction outputs (the "UTXO" model).
Mike Hearn wrote about this in the context of Corda's design: https://www.corda.net/2016/12/09/rationale-tradeoffs-adopting-utxo-style-model/
The key point I wanted to make, however, was that the UTXO model - at least as implemented in Corda - enables full-function code to be run... the set of supported operations is not in any way limited.
The differences become manifest mostly in the context of the non-functional characteristics of the platforms.
As Chris says, platform selection should be driven by use-case: I believe that the UTXO model we've used in Corda makes it particularly well-suited for a wide variety of financial services scenarios (that's why we made the choice to adopt that architecture!). But there is no one size that fits all.
