[mercury-users] Arguments against Mercury usage??

[Ralph Becket]

J?rg Roman Rudnick, Wednesday, 18 October 2006:
> 
> So what can I tell safely to a person wanting to express something about 
> Mercury but still stand on the safe side? What to consider right to 
> prevent overly ambitious promises?
> 
> + Mercury was mentioned to have a steep learning curve, making it not 
> the first choice in introduction to logic programming

Learning Mercury probably takes less effort than learning C++.  The
ideas underpinning Mercury (type systems, mode systems, predicate logic)
are not fundamentally hard; it is their combination in the same language
that can seem overwhelming to the newcomer.

> + the type class system still seems to need some time...

The type class system is quite usable, although we do have plans to
extend its functionality.  But that can be said of nearly everything in
language design.

The drawbacks are only apparent with some more adventurous uses of
polymorphism and multi-parameter type classes.  Observe that most OO
languages do not support multi-parameter classes or parametric polymorphism.

> + for network connectivity, I had to extemporize a TCP socket(!) - not 
> the thing `ordinary' people like, I assume... ;-)

Have you looked at posix.socket.m in extras/posix?

These things have all been done in the past, but have not necessarily
made it into the main Mercury distribution.

> + if the mode for vi and the Prolog mode for Emacs don't count, there 
> seems to be nothing like an IDE...

Real programmers use vim and make.

Personally, I'm skeptical of productivity claims associated with fancy
IDEs.

> + ... anything else??

You tell me!

> Obviously, you declared Mercury to be a zero-version yet - what do you 
> expect should be added so you would call it a one-version?

At the least we would like to support uniqueness properly.

> You made 
> almost no `advertisement' yet; Mercury is almost completely unknown in 
> business - I guess you discussed the whole a lot and wisely chose to do 
> so: What are your reasons?

There are at least three companies out there using Mercury: Mission
Critical, Yes Logic, and Logical Types.

> What would you like people NOT to promise prematurely about Mercury - 
> just to prevent tears and, even worse, undeserved bad press?

Mercury will not make weak programmers into strong programmers.

Mercury can (we believe) help *good* programmers be more productive and
to write much more robust programs.

Mercury is a general purpose language, but there are niches where it
would not be appropriate (e.g., hard real-time computing).

-- Ralph
--------------------------------------------------------------------------
mercury-users mailing list
Post messages to:       mercury-users at csse.unimelb.edu.au
Administrative Queries: owner-mercury-users at csse.unimelb.edu.au
Subscriptions:          mercury-users-request at csse.unimelb.edu.au
--------------------------------------------------------------------------