Fail safe or no fail safe, is not the issue. The issue is, get it running. Test it on a server. Let users experience it. And based on the feedback, change it. If its a disaster, reverse the change and move on.
(Building a fail safe) Is like saying we have to develop for performance when performance is not (yet) an issue. You don't know if it will be an issue. Deal with the problem when its there.
Software/Game development is not thinking it out into minute details and then implementing it. It's more like: I want this feature implemented, then you take a small step(easy to revert) towards it, evaluate it and then take the next small step.
Development is an iterative process, if you do nothing, nothing will happen. And if you make it easy to contribute, it will happen quickly if not it will last forever and people will loose interest.
To be honest, I lost interest... I'll probably keep playing now and then but wont be much interested into contributing to it. But, then again that is normal! Maybe later on someone else has an idea and wants to contribute but then realizes its hard to actually contribute and see changes, then he/she will also loose interest.
The above is just my view on things, and is not ment as criticism. I don't actually know how much effort it costs to change things in Arma. I messed around with its scripting language and its not pretty... But I do firmly believe its possible to automate development, to make it easier to contribute and try things out.