On 10/27/2013 6:54 AM, Miles Fidelman wrote:
Not in any way disputing your basic premise that it's hard, or
impossible, to do polymorphism and inheritance in C, with which I
agree.....
And not to weigh in on whether an "object" is a "machine" or vice versa
(though a pretty good arguement can be made that objects can be viewed
as finite state machines - and you can find such arguements in the
literature)......
Sure, you can find all kinds of arguments in literature. But the
discussion was not about finite state machines. And, of course, you can
call ANYTHING a "finite state machine".
But it is also not call "finite state machine oriented programming"
because "object oriented programming" is more appropriate.
And also not to weigh in (much) about what's actually important about
OOP (Alan Kay has more than once pointed out that it's message passing
and isolation that are important, polymorphism and inheritance are just
what seem to get the attention.) By the way: really nice discussion of
the "state of OOP" at
http://www.infoq.com/interviews/johnson-armstrong-oop
(on a completely un-related matter, I happened to be looking at Erlang
from an OO perspective and stumbled across this rather nice interview
with Rolph Johnson and Joe Armstrong - of "Design Patterns" and Erlang
fame, respectively).
That is Alan Kay's opinion, nothing more, nothing less. Others consider
inheritance and polymorphism to be quite important. But then those are
people who have actually used polymorphism and inheritance effectively.
But then such is the opinion of another academic who has had no (or very
little) real-world business experience.
Academics get paid to write papers. Business people get paid to do work.
Jerry Stuckle wrote:
And no, I did NOT refer to "reading a few pages on wikipedia and
writing a little php". I do not consider either to be reliable.
Rather, I referred to recognized experts in the field such as Booch,
Rumbaugh and Stroustrup.
Kind of flip-flopping to cite "recognized experts" after dismissing guys
like Donovan, Saltzer and Corbato as "But they never were that highly
regarded except in academia" - when discussing operating systems and
systems programming.
Yes, all academics who have never (AFAIK) done any business programming.
And then to cite Booch, Rumbaugh and Stroustrup re. OO programming (ok,
Stroustrup wrote C++) - but if you want to cite experts - how about Dahl
and Nygaard (Simula, pretty much invented software objects) and Alan Kay
(Smalltalk, pretty much invented OOP). Maybe Joe Armstrong (Erlang) for
a countervailing view.
Booch and Rumbaugh also were real world programmers, not just academics.
Separately they created OO design patters which were later merged into
UML - the most commonly used design pattern today.
Others are academics.
And BTW, it isn't limited to programming where the business world thinks
little of academics. It goes for many professions.
But you're too caught up in your own little world to even try to
understand REAL experts. Your mantra is "I have my mind made up and
no one will change it".
Sounds more like you're describing yourself, Jerry.
First, you've got to understand who the "real experts" are, rather than
finding ones who simply backstop your pre-defined opinion (and.. experts
are nothing without citations). Beyond that, why is it that you always
seem to draw from narrow confines of IBM (or, where Booch and Rumbaugh
are concerned, Rational, now part of IBM). IBM is an important part of
the computing universe, no contest; but the field, and it's leading
edge, are much broader than just IBM.
Miles Fidelman
Sure. Real experts have real experience in the field, using the tools
to write real programs. They get paid to do things which make the
company money. They don't just sit around and write papers and maybe
teach a class now and then.
I would love to see some of your "experts" in a real-world business
environment. My bet is they would all fall flat on their faces.
Jerry
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/526d19e7.3080...@attglobal.net
