Neil Conway wrote:
> Perhaps you're suggesting shell variables were used as the design
> model for psql's variables (although I can't be sure, you didn't
> elaborate). If so, what I'm saying is that this model is not very
> friendly for setting psql-internal options, and we'd be better
> changing it as I've elaborated on elsewhere. Do you agree?
Shell variables did serve as a design model, mostly because I found it
better to use *some* model rather than inventing behavior out of thin
air. Consequently, I am sort of biased on this.