Re: Universal Binaries, x86 and compatibility...
Re: Universal Binaries, x86 and compatibility...
- Subject: Re: Universal Binaries, x86 and compatibility...
- From: Henry McGilton <email@hidden>
- Date: Wed, 8 Jun 2005 23:12:55 -0700
On Jun 8, 2005, at 1:14 PM, Keith Blount wrote:
<<<<<<< munch >>>>>>>>
So... am I safe in assuming - from reading various
comments and looking at the docs published with Xcode
2.1 - that the chances are that my app, which doesn't
do much more than use some custom views and use basic
AppKit/Foundation classes, will most likely just run
on an Intel machine once it is compiled as a Universal
Binary? It's not a big deal if the compatibility
issues are likely to be relatively minor and if it
will just mean ironing out some small bugs when the
new machines come out; what worries me is the idea of
having to rewrite hundreds of lines of code I'm
investing in right now a year or two down the line.
I started developing on NextStep sometime in 1991.
By around early 1993 I had about ten shipping
applications. Around that time, NextStep was opened up
and ported to Intel 486, HP Snakes, Sun Sparc, and (never
released) the Motorollah 88000.
When the Intel port was available, I selected the appropriate
switches in Project Builder to compile for Next and Intel,
generated the (fat) binaries, and all ten apps ran quite
flawlessly without my having to change a single line of code.
There was no time spent on chip-specific code, and no time
spent on debugging. Everything Just Worked . . .
Around the same time period, the OpenStep project started,
running on Next boxen and Sun Sparcen. Same old story
(modulo Sun not knowing what the hell they were about):
Everything Just Worked . . .
Here we are, ten or so years on, and the Great Mandala
continues turning. One can make several assumption-based
analyses of how this twelve-year old 'transition' will
likely play out:
o the underlying Mac(h) kernel, BSD layer, and all that stuff
are very mature, and have been ported to multiple chip
architectures over a period of around fifteen years . . .
o the old NextStep/OpenStep AppKit and Foundation Kit are
very mature, and have been ported to multiple chip
architectures over a period of around fifteen years . . .
o for the past four years, I have been writing code to a set
of APIs (NSObject, NSView, NSWindow, for examples) that look
pretty much the same to me today as they were fifteen years
ago (Object, View, Window) . . .
o the Apple engineering team --- many of whom came from Next
in the first place --- know what they are about, have been
through this process in the past, and understand the issues
very well . . .
o Even on Panther, if you look though the .h files for
the system, you will see massive numbers of references
to i386. That indicates to me that the Apple engineering
team have been worrying about this for some time . . .
At the end of the day, I have minimal concern about the transition
to Intel chips. What I am spending my time on right now is
developing new apps, with confidence in the ability of the Apple
engineering team to provide me a programming platform that will
Just Work . . .
Cheers,
. . . . . . . . Henry
===============================+============================
Henry McGilton, Boulevardier | Trilithon Software
Objective-C/Java Composer | Seroia Research
-------------------------------+----------------------------
mailto:email@hidden | http://www.trilithon.com
|
===============================+============================
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Cocoa-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden