A solid first step would be defining and formalizing, and then proposing alternatives once the problems are known. That several folks have attempted something in the past is a clear indication that this needs to be done.
The Major part isn't the actual coding, it's reworking the old files for any changes. But it's hard to attempt buy-in for the Major work until the defining and formalizing has been done.
I don't know whether the final product should be C or PHP or both, and won't know until problems have been identified and buy-in has been accomplished.
I'm in favor of this project.