[mercury-users] Bertrand Meyer's opinion

[Ralph Becket]

Personally, I doubt that it's realistic to expect widespread adoption
of `academic' languages - however, it is the case that some of the
core ideas are filtering through into the mainstream.  For example,
strong(er) and more powerful type schemes, generics (~= polymorphism) 
and delegates (~= closures) are making itinto contemporary languages
(C++ templates, the various generic Java proposals, Microsoft's .NET
framework, C#, etc.)

Joe programmer, it seems, finds the notion of `state' easier to understand
than `function' and prefers a big bag of mostly-okay tools (e.g. C++, Perl)
to a small, clean, set of primitives (i.e. Mercury, Haskell etc.)  In a
word, all syntax and no semantics.

Moreover, academic languages are still thought of as inefficient, although
personally I doubt that most large applications would really suffer - my
prejudice being that the relative simplicity of implementation, reduced
development and debugging time etc. probably do a good job of hiding the
performance gap, such as it is.  But nobody wants to listen.

On the performance issue, one of the big problems today for larger 
applications is the appallingly low processor utilisation that is obtained
through lack of exploitation of concurrency and poor cache behaviour (the
twin curses of (a) the difficulty of handling asynchrony and (b) the way
thread based approaches trash the cache).  I wonder whether Mercury et al.
provide some means of overcoming these problems...

Real-world applications are probably the best advert' a language can have.
Next up would be statistics such as development time, maintenance
requirement
and lines-of-code.  I don't think touting groovy theoretically motivated
programming practices should go in the sales brochure - that stuff just
turns
the pleb's off.

Assuming all goes well, it'll be interesting to see how well Mercury does in
the code size and efficiency stakes in next week's ICFP coding challenge.

Ralph

--
Ralph Becket      |      MSR Cambridge      |      rbeck at microsoft.com 
--------------------------------------------------------------------------
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
--------------------------------------------------------------------------