Re: [Gnu-arch-users] Re: darcs vs tla

[Bruce Stephens]

Andrew Suffield <address@hidden> writes:

[...]

> I am willing to accept as a hypothesis that some languages may be
> faster to write than others, but there is no more than
> circumstancial evidence in both directions.  Furthermore without a
> way to quantify the skill of a programmer in a given language, in a
> manner comparable to the skill of a different programmer in a
> different language, I don't think it is possible to have real
> evidence either way.  So you won't really get much mileage out of
> it.

I'm reasonably sure I've seen cites of research comparing productivity
across languages and individuals.  I agree that it seems impossible to
eliminate all of those issues; however, the issues don't look that
different from those in other fields, so I imagine researchers have
developed approaches to investigating such things.

> Subjectively, the programmer appears to be the primary factor in
> determining productivity, not the language.

Oh yeah, I'm pretty sure the available evidence supports that.

Joel Spolsky argues (in a small part of this article) that languages
do make a difference, but that the largest proportion of it is simply
automatic memory management
<http://www.joelonsoftware.com/articles/APIWar.html>.  It seems likely
that he's really comparing languages which aren't that different from
each other though, like C++, Java, C#, rather than more radically
different languages where you might hope advanced things (higher-order
functions, etc.) would make an impact.

[...]