[using gcc book] ch1 objective-c blurb

Matt Austern austern@apple.com
Tue Aug 26 17:59:00 GMT 2003


On Monday, August 25, 2003, at 9:04 PM, Chris Devers wrote:

> On Mon, 25 Aug 2003, Matt Austern wrote:
>
>> On Monday, August 25, 2003, at 08:37  PM, Chris Devers wrote:
>>
>>> More importantly, based on another branch of this thread, I'm not 
>>> sure
>>> that the lack of a C intermediate is a significant assertion anyway.
>>
>> I don't think it is an interesting assertion.  It's a true statement,
>> but you can make lots of true and uninteresting statements about
>> intermediate representations that the compiler doesn't have.  (e.g.
>> the C compiler doesn't use Fortran as an intermediate language.)
>
> Great, you agree, but what then *would* be an interesting assertion?

A brief discussion of what kind of language Objective C is, what kinds
of problems it's supposed to solve, a link to Brad Cox's original
work on the language (Object-oriented Programming, An Evolutionary
Approach), and maybe a brief discussion of the ways in which Objective
C and C++ differ.  Roughly: Objective C is an object-oriented extension
to C that makes it possible to combine C with Smalltalk-style designs.
The Objective C object model, and the message-sending syntax, were
deliberately modeled on Smalltalk.  Objective C differs from C++ in that
it favors a more dynamic programming model with less static type-
checking, and in that it is a much simpler extension to C.  Objective C
is used by the free GNUStep application framework and by Apple's
Cocoa framework.  There is also a combination of Objective C and C++,
called Objective C++.  [Objective C++ hasn't yet been integrated into
GNU gcc, but by the time this book comes out it may have been.]

You shouldn't use what I wrote verbatim; you should probably reword it,
expand it, maybe correct some details, and maybe include some details
about the ways in which ObjC has evolved since Brad Cox's original
work.  But I think that's the sort of information that users would care
about.  Most users have no reason to care about low-level implementation
details like the compiler's intermediate representation.

			--Matt



More information about the Gcc mailing list