Subj : Re: curve for verbosity in a language
To   : comp.programming
From : Jon Harrop
Date : Sat Aug 06 2005 07:43 pm

pantagruel wrote:
> I don't think that LOC is verbosity,

LOC is certainly not an ideal measure. It is just the best I have come up
with (and it is still highly subjective).

> however your suggestion is close, how about:
> 
> if the character count between two programs doing the same task differ
> then the program with the greater character count is verbose in
> comparison to the first with each character being a point of verbosity.
> I think this should include a normalization of whitespace.

I have a couple of problems with that. Firstly, in some languages you must
put in more whitespace (e.g. "=!" and "= !" are not the same in OCaml, and
you must write "list<list<int> >" in C++) and in some languages characters
are often put in by convention (e.g. "f (a+b)" rather than "f(a+b)" in
OCaml). Secondly, it fails to take into account the importance of seeing a
lot of information on the screen at once.

Also, OCaml code is easier to navigate than C++ code simply because it has
more characteristic shapes. In particular, programs written in C-like
languages are mostly unnecessary braces and loops that add nothing to the
syntax.

> Performance differences in the task between the programs affect the
> verbosity, the question then is how is a difference in performance
> calculated into program verbosity.

Yes. That is why I plotted LOC vs speed graphs for four different versions
of my ray tracer.

> At that point one could begin to make reasonable statements such as
> language x has been shown to be verbose in comparison to language y for
> problems of type z.

Have you read the conclusions on the page about my ray tracer?

-- 
Dr Jon D Harrop, Flying Frog Consultancy
http://www.ffconsultancy.com

.