scripting languages vs statically compiled ones

[Alex Martelli]

Richard Blackwood <richardblackwood at cloudthunder.com> wrote:
   ...
> >Language       Type of Language     Execution Time Relative to C++
> >C++            compiled             1
> >
> Now since when was C++ as fast as itself?

Well, this is ONE line in the table nobody's gonna disprove!-)


> >Table 4-1 of the same book shows the "ratio of high-level-language
> >statements to equivalent C code" (higher is better):
   ...
> >Java                    2.5
> >Microsoft Visual Basic  4.5
> >
> Higher-level than Java?  Well, you surely pay for this in language 
> features and language design sensibility.

That's another issue, and harder to quantify.  Counting average number
of statements per function point, which is what "language level" means
in this context (and the table's title is very explicit about it), it
does seem right -- of course the precision of these numbers is dubious,
it could be, say, Java=3 and VB=4, &c.

> >Perl                    6
> >Python                  6
> >
> Not above Perl?  I shall have to think about this one.

In statements per function point?  Idiomatic Perl tends to be way more
cryptic and terse than idiomatic Python; that Python pulls back to rough
parity (and I do agree with this table that it does) is, in a sense,
surprising.

Note that this need not translate to same-productivity, although "fixed
number of LOCs/hour" is a long-standing hypothesis of the Function
Points theorists.  Prechelt's work couldn't _disprove_ the hypothesis in
a statistically significant way, but if you eyeball the language medians
in the LOCs/hr graphs in his IEEE Computer article (available as a PDF
on the net), you get the strong visual impression that, while most
languages do cluster in that measure, Java tends very much to the low
end of the cluster (20 LOCs/hr) and Python to the high end (40).


> >http://page.mi.fu-berlin.de/~prechelt/Biblio/jccpprt_computer2000.pdf
> >"An Empirical Comparison of Seven Programming Languages", by Lutz
> >Prechelt, IEEE Computer, October 2000, 23-29.

Yep, THIS paper!-)


Alex