Path: utzoo!attcan!uunet!hsi!stpstn!cox
From: cox@stpstn.UUCP (Brad Cox)
Newsgroups: comp.software-eng
Subject: Re: Reuse and Abstraction (was: reu
Message-ID: <5152@stpstn.UUCP>
Date: 2 Jun 90 17:05:32 GMT
References: <4979@stpstn.UUCP> <102100009@p.cs.uiuc.edu> <19820@duke.cs.duke.edu> <12085@june.cs.washington.edu> <81111@tut.cis.ohio-state.edu>
Reply-To: cox@stpstn.UUCP (Brad Cox)
Organization: Stepstone
Lines: 33

In article <81111@tut.cis.ohio-state.edu> Bruce Weide <weide@cis.ohio-state.edu> writes:
<This gets to Brad Cox's point about "common sense." I guess the above
<example helps illustrate where Brad and I might disagree (if I
<understand the argument he's making).  Brad seems willing to describe
<the intended behavior of stacks by saying the implemented objects "act
<like stacks" and relying on my common sense to understand what he
<means.  As I've pointed out before, though, even THIS seemingly rather
<precise statement means different things to different people.  It's
<not that hard to say it rigorously, and (IMHO) no harder to say it
<formally than rigorously.  I still don't understand the resistance to
<doing so.  People keep implying or explicitly stating that formal
<specs are inherently such a pain, but why?

"Most" people (certainly not myself) aren't saying that they're a big pain
for stacks. A lot of them *are* saying that their usefulness is less obvious
than their costs for bigger objects like Sets (see Ralph London's article),
word processors, or hamburgers.

To restate what I'm saying once more, I'm interested in a paradigm shift,
a software industrial revolution. We'd start applyingw3 an organizational
principle called *division of labor* (or *reusability*). In the end
(which might take as much as fifty years, which seems to be the historical
time constant for such paradigm shifts) we'd have software components
marketplaces, in which building a word processor from trusted components
might be done with about as much effort as it is to build a stack from
raw software materials (ints, pointers, etc).

This path, while *clearly* not easy, might get us to the point where the
formal approaches you advocate for Stack might also apply to word processors.
-- 

Brad Cox; cox@stepstone.com; CI$ 71230,647; 203 426 1875
The Stepstone Corporation; 75 Glen Road; Sandy Hook CT 06482