[
Lists Home |
Date Index |
Thread Index
]
1/23/02 9:37:13 AM, Mark Baker <distobj@acm.org> wrote:
>My objection is simply the name and the meaning that the name invokes,
>that it's somehow a good idea to be "worse", because sometimes "worse
>is better". Considering the whole system when designing, rather than
>just part of it, can in no way be considered "worse". I don't
>disagree with the conclusion of "worse is better", but I disagree with
>the unfortunate choice of wording.
I much prefer the way Kernighan and Plauger phrased the concept back in 1976 in _Software Tools_:
something like (I don't have a copy handy) "something that does 90% of the job today is often better
than something that promises to do 100% of the job sometime in the next 6 months."
IMHO, if you come (as I do) from a math background, it's easy to overemphasize completeness and
internal elegance. Both of them are highly desirable properties for certain classes of abstract
concepts, but I'm a realist in the philosophical sense, and therefore view abstractions as
(necessarily imperfect) models of reality, rather than reality as imperfect realizations of
abstractions. It's certainly a Good Thing that Turing came up with an elegant and complete formal
model of computation, but that doesn't imply that his formalisms are the Best Way to implement real-
world systems.
|
