On 10/30/07, Elf <elf at ephemeral.net> wrote:
>
> this is a query directed at those in support of r6rs:
>
> does r6rs represent a manifestation of the Right Thing?
>
>
> reason for this query: it seems as though the major difference between the
> various factions boils down to the Rightness or lack thereof of r6rs. do
> those defending r6rs claim that it is the Right Thing? this is unclear.
I think this question suffers three major problems, and therefore
isn't really a good question to ask about programming languages.
First, there is no "Right Thing" in the programming language space.
Some design decisions are better than others, but there is no perfect
programming language. Some languages are better for some tasks than
others, but they cannot be order hierarchically. Even if they could,
that hierarchy would not be "Rightness", since there is no right or
wrong in these matters.
Second, even if there was a "Right Thing", the attitude that has been
taken is that the only things that should be added to Scheme are
things that are components of the "Right Thing". But programming
languages should not be designed by incremental addition of features,
but instead by careful thought and design of the whole system.
Sometimes other concerns prevent redesign, but the addition of one new
concept every 10 years will not produce the "Right Thing".
Third, the criteria by which many people who want Scheme to represent
the "Right Thing" seems often to be "fundamentalness", that is, the
ability to reduce the language to the minimal set of features from
which everything else can be implemented. While this is often a
desirable goal, it is certainly not the only one, and it also can
hinder the development of other features. Especially when we do not
yet know what the "Right Thing" might be, demanding it will hinder the
development of the language. The perfect is the enemy of the good.
--
sam th
samth at ccs.neu.edu
Received on Tue Oct 30 2007 - 12:14:32 UTC