[Haskell-cafe] Is Haskell a 5GL?

[Andrae Muys]

On 30/09/2006, at 6:15 AM, Nicolas Frisby wrote:
> Software engineering is as of yet misnamed. A professional engineer's
> design work should never include figuring out why the first attempt
> exploded/collapsed/failed--professionals in mature engineering fields
> only debug catastrophes.

That is only the case when engineers are performing pure  
development.  When they are involved in research, engineers first  
attempts routinely explode/collapse/fail. Building a house, bridge,  
refinery, family car, etc, stopped being research a long time ago.   
However F1 racing cars, and interplanetary autonomous vehicles are.   
Software Engineering is so new a discipline that we still don't know  
how to render any but the most trivial of problems 'pure design'.

> My intended takeaway is that design in software engineering does not
> yet compare to design in the mature engineering fields. In my
> engineering-centric opinion, the goal of computer science is to enrich
> the design principles of software engineering so that it does compare.

I couldn't agree more.  So I'm going to quote the passage from Robert  
Dockins that you objected to.

On 29/09/06, Robert Dockins <robdockins at fastmail.fm> wrote:
> In a similar way, if a someone refuses or is unable to learn the  
> mathematical foundations of computation, I don't think I really  
> want him programming any systems that I'm going to be relying on.   
> He don't need to be an expert in category theory, but if  
> programmers aren't learning the skills they need to understand the  
> basics and mathematical notation of PL theory, then something is  
> very, very wrong.  (Not to beat you over the head with my point,  
> but what I'm saying is that programmers really ought to know this  
> stuff and the fact that most do not is a terrible state of affairs).

Of course we will have hobbyist programmers, in much the same way we  
have home handymen or backyard mechanics.  However anyone claiming to  
be a professional programmer has to deal with everything they do not  
being pure design, but containing a non-trivial element of research.   
For that reason I have to agree with Robert.  I too am very concerned  
at the number of programmers writing code that I rely on who don't  
have a basic understanding of the mathematical foundations of what  
they do.

Andrae

-- 
Andrae Muys
andrae at netymon.com
Principal Mulgara Consultant
Netymon Pty Ltd