Personally I have to think about this in my mind before I go and vote.
On one hand: A New plugin system as described above where most functions are modular, there might be a noclip/godmode plugin/function for sourcemod, but if I don't want it I just turn it off. It works on a fairly 'old' and 'incomplete' and in some instances 'inadequate' architecture, but, it's one that's already known.
So in other words option 2 has it's positives, and it's negatives.
However, option one is more ideal (in my opinion), and in the long run being able to support your own coding structure and system is a bit easier than dealing with another one and waiting for their developers to fix something you knew how to fix weeks ago. Plus there are other potential benefits to both the individual developers of it (future job opportunities), and to the gaming community.
Oh and, CSS (and other HL2 games) attribute to 50% of the valve games that are being played (according to serverspy.net) and the total numbers still outweigh that of battlefield 2 and a few others, and those 'few others' have plugins being developed for them (though lacking and extremely limited)
But as a long time server administrator myself, and a very short role coding very simple plugins for AMXX, I would agree with KaszpiR, roll out a simple 'option 2' only do bug fixes and major feature requests while continuing working on option 1, while it makes for a longer development time, it may be worth it.
Yeah yeah, there's probably a lot I don't entirely understand, I'd like to read the logs (and I will) but it's 3
0 AM I'm tired, I'm going to sleep.
__________________