C is fastest
Lennart Sorensen
lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org
Thu Oct 15 14:06:41 UTC 2009
On Thu, Oct 15, 2009 at 02:02:04AM -0400, Yanni Chiu wrote:
> Christopher Browne wrote:
>> You'll find a number of us that haven't "taken the OO kool-aid."
>
> It's sad if your impression of OO is formulated from C++ and Java. OO
> was originally conceived of in Smalltalk, and in that environment,
> objects have a "liveness" characteristic that was a design goal. This is
> missing in a dead code environment like C++ and Java, where the code is
> just text in a file. In Smalltalk, objects are all available to be poked
> and prodded via inspectors. Even the class object itself can be
> inspected, just like any other object you've created.
Actually simula-67 had objects and classes and all that long before
smalltalk existed. It is considered the first OO language.
> To pick up on an earlier critique of OO from someone else (namely that
> you have to create a class be able to start writing code), this idea has
> been explored in the Self language. In Self, there is no distinction
> between classes and non-classes. Everything is just a thing with slots
> to hold other things. What I've been told is that you start out creating
> new objects without thinking about classes as you develop your system,
> and by the time you've finished re-factoring and such, you end up with
> some objects that pretty much function like classes. These class-like
> objects end up holding the behaviour that's common for a group of
> objects. There can still be some objects that don't have an associated
> class-like object though.
But I don't want everything to be an object. It doesn't fit a lot of
problems well at all. Objects can be useful tools (although far less
often than people try to use them). They should not be required in
a language.
--
Len Sorensen
--
The Toronto Linux Users Group. Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
More information about the Legacy
mailing list