RE: [boost] Boost.Format syntax

From: Brey, Edward D (EdwardDBrey_at_[hidden])
Date: 2002-02-18 10:26:32

> From: David Abrahams [mailto:david.abrahams_at_[hidden]]
> 
> I'm curious about something. Of all the people with whom 
> you've corresponded
> about using operator[] in format, what percentage have 
> immediately responded
> positively to your view of operator[] as an analogy for projection?
> 
> I think op[] has some clear technical merits, but unless my 
> antennae are
> seriously mistuned, almost noboone thinks that it makes any sort of
> "intuitive sense"... and that's not something you'll be able 
> to convince
> people of. It's like a joke: either they get it, or they 
> don't. If they
> don't get it, they won't laugh even if they understand you 
> when you explain it.

The realm of "intuitive sense" is a difficult one to discuss.  I didn't
think that I was trying to defend whether or not someone would understand
the logical soundness of format::operator[] compare with the ease of
understanding why a newspaper is black and white and re(a)d all over.  I was
just responding to the good questions that James had posed regarding the
technical merits of the proposal.

Fortunately, in this case, the technical validity of the project is more
important the intuitive aspect.  For an intuitive side, the question is
whether the syntax clearly indicates that what it inside the brackets will
be substituted into the context string.  Knowing that this substitution
happens to logically map onto existing C++ usage of operator[] is kind of
like knowing that the word "sincere" comes from Latin for "without wax",
from the days when unscrupulous potters would use wax to hide defects in
their work, and so the reputable merchants would hang a sing over their
shops advertising their abstinence from the practice: it's fun and
interesting theory, but doesn't really affect your usage of the construct.