best evidence.

I seem to often hear those opposed to FRBR/RDA for whatever reasons say things like: “The FRBR user tasks were not tested with real users. We have no idea if the FRBR model actually supports what users want to do.” Those more strident simply aver that the FRBR model does not support real user tasks.

But check this:

With expressions, editions/manifestations, and works…. indeed LT might become the most full implementation of FRBR out there, as Tim suggests. Which is perhaps ironic because I’m pretty sure LT/Tim don’t care at all about implementing FRBR for the sake of FRBR,or for the sake of ‘standards’. LT is implementing what makes sense to meet their user’s needs. [Which is perhaps why initially Tim was resistant to accepting that he was 'doing FRBR' at all]. (src: Bibliographic Wilderness)