Re: Seeking for papers about tagged types vs access to subprograms

["Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr>]

Le Sun, 12 May 2013 09:14:41 +0200, Dmitry A. Kazakov  
<mailbox@dmitry-kazakov.de> a écrit:

> To put it simply: if a type is designed in a way that limits it use, it  
> is
> a poor design.

The sense or meaning, both in natural and formal languages, comes from  
restricting possible interpretations.

> You are conflating the
> semantics of a type with the semantics of *a* program that uses this type
> somewhere, somehow.

Precisely, I see types as constituents of a domain, so also as  
constituents of applications of the domain. I'm guessing I'm not the only  
one.

We may diverge here, in short I would say: I don't want bricks giving the  
illusions I may always be able to use it to build something which may or  
may not fails, I want bricks which will tell the sooner I will not be able  
to use it in some invalid ways if I want something which will not fail (or  
at least reduce potentials for failures, as program proof is more than  
that).

What I don't understand is: how something which give the illusion it has  
no prerequisites, can help in constructing safer and more trustable  
architectures? Reusing is nice, however only where it, and in a way which,  
make sense for the domain (I don't believe an exception is meaningful,  
that's not a value, that's a runtime error). Seems the invariably `True`  
precondition wants to give the illusion of universal reuse without  
conditions. With all precondition moved to the postcondition, the program  
has more potential for erroneous interpretation, because that may turns  
into a lot more decision paths in the program after each invocation.

I will not argue that much on this issue, as that may involves personal  
feeling.


-- 
“Syntactic sugar causes cancer of the semi-colons.” [1]
“Structured Programming supports the law of the excluded muddle.” [1]
[1]: Epigrams on Programming — Alan J. — P. Yale University