On 7/26/06, Wade Curry <[EMAIL PROTECTED]> wrote:
Stewart Stremler([EMAIL PROTECTED])@Wed, Jul 26, 2006 at 03:58:52PM -0700:
> begin quoting Carl Lowenstein as of Wed, Jul 26, 2006 at 03:40:41PM -0700:
> > On 7/26/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> > >Assembly is useful to learn. It helps you understand CPUs
> > >and ultimately what your high level code is doing under the hood.
> > >
> > >I'd like to give
> > >an intro to assembly programming class if anyone is interested.
I'm inclined to attend if time, schedule, and life allow.
<snip/>
> >
> > Do you have any particular type of CPU in mind?
> > Or do we just assume that all the world is Intel x86?
>
> >From the discussion on IRC, it's the latter.
>
> Personally, I don't see the point. If you're going to give in to
> the dominant paradigm, why bother with Linux at all?
>
OK, I have to say that this is an obvious non-sequitur. I don't
run Linux because of anything to do with dominance. Linux is my OS
of choice because it allows me to control *my* computer more
entirely (Gnome notwithstanding ;-) ). Learning to cope well with
the dominant paradigm, even learning the paradigm itself, isn't the
same as giving in to it, either. Since it's the _dominant_
paradigm, it's not like you won't run into it at every turn. It
makes no sense to pretend otherwise, and learning it may be the
best way to cope (assuming that it's a bad paradigm).
It seems likely that Intel CPUs have a less than attractive
assembler, and architectural features that irk or completely
anger the knowledgeable. I'm not arguing the actual quality, but
just pointing out that abandoning everything dominant out of hand
only because of its dominance doesn't make sense to me.
This brings me to my real question. What is it about Intel's X86
assembler langage that is so awful? That isn't a challenge; it's a
sincere question. I have stolen a glance at snippets of assembler
in books and articles, maybe even in some source I've on my own
computer. With that said, what is it that could vary between two
assemblers so much? Obviously two CPU architectures will have
different instruction sets, but what does it mean that X86
assembler is convoluted? Is it a matter of CISC v. RISC? Is it a
matter of the assembler language? Peculiar requirements of the
Intel branded chips?
As usual, people speak of the assembler when they are really referring
to the instruction set. It is the x86 instruction sat that is badly
convoluted. The assembler merely allows direct reference to the
instruction set.
carl
--
carl lowenstein marine physical lab u.c. san diego
[EMAIL PROTECTED]
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
