>>Scatter Name is not totally obsolete, but fallback pattern if there are reasons a table buffered cursor is not enough - but the idea of scatter memvar in vfp ***is*** obsolete.

I have ancient code predating SCATTER NAME or even inheritance in VFP. Obsolete? Maybe, but so is VFP and if it ain't broke...

I also try to use mdot every time on the right side of any assignation or test. Not the left, and not at all if it's implicitly a variable. Nor have I always used mdot for THIS or objects. Some might call this inconsistent, but a lot of this was written when mdot was out of fashion and benefit versus potential bug injection doesn't really support remedy. Especially when clusters of more than 3 this. or object. are said to benefit from WITH which I've used consistently.

IMHO there are so many use cases and "what about"s that there's unlikely to be single universal answers to many of these concerns... especially since you need fairly tight loops to be able to measure most differences and some of positive or negative examples do come across as trick shots.

I did notice a comment from Dragan along one of these threads, that prg-based forms and classes instantiate quicker than scx or vcx. Did I get that right, Dragan? Definitely the case? Last I heard (admittedly many years ago) was that dbfs or views in a visual data environment load visibly quicker than a series of USE but again, that's likely to be eclipsed by better hardware over the decades...