[mercury-users] constructor class syntax

[Samuel Sean Watkins]

On Tue, 2 Nov 1999, Ralph Becket wrote:

> > ... Has any 
> > research been
> > done on whether token-rich or token-sparse languages are easy for a
> > *human* to write, and subsequently parse? :)
> 
> I'm averse to excessive overloading (e.g. the madness in C++), but I'm
> also averse to Too Much Syntax.  I liked Lisp and Prolog because I didn't
> have to keep a Big Book of Syntax next to me.  Mercury, I think, has
> strayed too far from the simple-syntax camp, since I frequently have to
> refer to the language spec.  Mind you, it is doing an awful lot more than
> Prolog did.

I think that Lisp has the right approach - define an absolutely minimal
'built in syntax' for the language (i.e. the brackets!), then do
everything else with functions (built in or not), or relations in the case
of a declarative language.  Operator overloading is good if the 
meaning of the operator stays the same.

It's pretty important to keep the *language* as simple as possible (but no
simpler) and then to build a huge library on that language.  I think
'syntactic sugar' should definitely be a library feature rather than a
language feature if that is possible - how about a language feature for 
describing additional syntax?  Perl's approach to syntax is a 
pretty interesting one.
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------