some comments for Python 3000

[Robin Becker]

In article <3999E27E.5A187A07 at san.rr.com>, Courageous
<jkraska1 at san.rr.com> writes
>Aahz Maruch wrote:
...
>> Interesting.  Bertrand Meyer seems to have a different POV in Software
>> Development Magazine, July 2000:
>> http://www.sdmagazine.com/uml/beyondobjects/s0007bo.shtml
...
The thrust of the article seems to espouse design by contract and yet I
get a 404 on http://www.designbycontract.com/research/papers.html
and Linux the language sic leads to a blank page, in the original the
link http://www.trustedcomponents.org is busted as well. A typical case
of the rather abstract and lofty view leading to pride leading to a
fall. The claim that any single design methodology will result in better
software is doubtful at best. Engineers have know how to design and
build bridges for a long time, but they still fall down. The problem
with software isn't in the design domain, but in the
problem/implementation domains. We still can't guarantee that the chips
won't burn out and crash the train or that the specification won't allow
the aircraft to land with the undercarriage up etc etc etc. Attempting
to build 'perfect' systems is pointless (and fate tempting) we should
try to build 'robust' software which crashes more softly.
-- 
Robin Becker