Subj : Re: curve for verbosity in a language
To   : comp.programming
From : Vesa Karvonen
Date : Sun Aug 07 2005 03:16 am

Jon Harrop <usenet@jdh30.plus.com> wrote:
> Vesa Karvonen wrote:
[...]
> Whitespace is important to both the machine and the human. The fact that the
> amount of whitespace is irrelevant for the machine has nothing to do with
> the notion of verbosity (which is for the human).

Your notion of verbosity does not agree with any widely used definition I
have seen. Again, according to Merriam-Webster, verbose means

 "containing more words than necessary" .

Whitespace is not considered a word by any strech of imagination in most
programming languages.

> I can type "wc" in under 1sec. So that is ~5 orders of magnitude faster than
> writing and testing my own lexers.

What would you estimate would be the time for you to type the name of a
token counting program after someone has written it? For the sake of
argument, let's assume that the name of the token counting program would
be called `tc'.

The token counting program only needs to be written once and extended once
per language --- not every time you use it. You don't roll your own `wc'
each time you want to use one.

> > *NOT (number of tokens) is a well defined and objective metric*.

> No. Token count is also subject to personal style. For example:

Let's see.

> let f a b =
>   let c = a + b in
>   c * c

Here I count 15 tokens and 3 lines.

> let f a b = (fun c -> c * c) (a + b)

Here I count 18 tokens and 1 line.

The token counts differ by 17% and the line counts differ by 67%.

*Evaluation*: Looking at the example, my opinion is that the latter
definition is slightly more verbosely written; the latter definition has more
puctuation than the former one. Since both functions do the same thing, I
would expect that a good complexity metric would give similar results for
both. The token count metric agrees with my expectations in this respect
clearly better.

However, let's look at a more interesting example:

  let f a b=let c=a+b in c*c

  let f a b =
    let
      c = a + b
    in
      c * c

Between the above definitions there is no difference in token counts
and the line counts differ by 80%.

*Evaluation*: The definitions are considered identical in Ocaml. I
personally consider them equally verbose, although they are formatted
differently (and neither according to my personal preference, but still I
wouldn't be surprised to see either style being used). Since the
definitions are identical, I would expect an objective metric to give
them, from a practical point of view, the same result. The token count
metric succeeds. The line count metric fails miserably at matching my
expectations.

> Indeed, verbosity is inherently subjective.

To some degree, yes. Particularly when you choose a lousy definition of
verbosity. However, there are more and less subjective ways to define and
measure verbosity as I have hopefully demonstrated above. Based on the
above, I consider the token count metric to be a clearly superior metric
of verbosity compared to the line count metric. Line count metric is at
the best a good metric of personal preferences in formatting code and at
the worst a poor metric to base salary upon.

> Regarding the personal note, OCaml is my favourite language and do I prefer
> it to the other languages that I have learned. Consequently, I advocate it.
> Not the other way around.

I think we need a new group: comp.lang.ocaml.advocacy.

I have nothing against Ocaml. I find it a very nice language, based on my
experience in writing Ocaml code for a couple of years. However, I find
your arrogant style of advocating it in terms of highly subjective and
biased metrics quite revolting. There is a reason why there are so many
advocacy groups. Those groups are for people like you.

-Vesa Karvonen

.