Count me in, will send you a PM. But... before we start the hard work, let us think a few more days about the following.
I suppose the main driver for a moderated RFF lens DB is the fact, that collecting new lenses from EXIF data or manual input will lead to many similar miss-spelled names for the exactly same piece of glass, especially entering modern electronic zoom lenses, existing in different edition and also specifically for APS-C and full format.
Cautionary tale: LensModel="AF-S Nikkor DX 18-200/3.5-5.6 G IF-ED VR II"
One small typo and you have an additional entry for the same lens.
Given this problem I have to agree on a moderated own RFF DB.
Why exact data in a DB:
If we really want to have lens statistics and a dropdown list to the single glass,we need a correct moderated DB. The downside is the big never-ending-work to keep the DB in sync with new equipment. Are we able to accomplish this on the long term?
Fuzzy logic approach:
As already mentioned by furcafe we can forget about exact data and have the lens data as a text field without restrictions. Provide the search function (for lenses) with a normal DB query (e.g. "Nikkor & 18-200" would find all similar spellings of the above complicated example).
Additionally you could use the - mostly - correct EXIF data from the images for an initial load of the lens DB. Do we accept having completely wrong and many similar entries? Do we accept having no dropdown function for existing lenses (interesting for film users).
Are there other pros and cons to these alternatives?