scsh-users
[Top] [All Lists]

Re: Ousterhout and Tcl lost the plot with latest paper

To: Bill Eldridge <bill@rfa.org>
Subject: Re: Ousterhout and Tcl lost the plot with latest paper
From: Ben Black <black@zen.cypher.net>
Date: Wed, 2 Apr 1997 13:40:03 -0500 (EST)
Cc: scsh@martigny.ai.mit.edu
> 
> > >The strong typing of object-oriented languages encourages
> > >narrowly defined packages that are hard to reuse. Each package
> > >requires objects of a specific type; if two packages are to work
> > >together, conversion code must be written to translate between the
> > >types required by the packages.
> 
> If you define the interface well in OOL's, you can handle a variety
> of typed inputs.  It's much easier to extend that typing in OOL's than
> with something like C (and have deterministic output).  I would guess
> serious programmers do develop their own libraries, or work with
> corporate standard ones (what does Adobe use? they can't be rewriting
> every graphics routine from scratch).  Having classes inherit methods
> and types is much nicer than having everything default to a string
> and thinking that's sufficient.  MFC is a help for OOP exactly by
> defining libraries that simplify programming.  The dangers on
> simplification
> include that it might come at the expense of too much speed, that
> it might not provide the tools to do everything a lower level 
> approach would, or that in doing more complex tasks, the "simpler"
> language actually becomes more obfuscated than the lower-level
> languages.
> 

having started to use the C++ Standard Template Library (STL) recently, i 
must say it totally disproves ousterhout's assertions.  the interfaces 
are extremely general and because they use templates instead of virtual 
members, all the overhead is compile-time.  there is no performance hit 
for the genericity.

also, i think MFC is perhaps a bad example of how to use C++.  M$ has 
never impressed me with their API design ability.


b3n


<Prev in Thread] Current Thread [Next in Thread>