Re: What's the best language to learn? [was Re: Should I learn C or Pascal?]

[bokr@accessone.com (Bengt Richter)]

mdw@excessus.demon.co.uk (Mark Wooding) wrote:

>Szu-Wen Huang <huang@mnsinc.com> wrote:

>> Lofty, ungraspable concepts like:
>> 
>>   void swap(int *a, int *b)
>>   {
>>     int c = *a;
>> 
>>     *b = *a;
>>     *a = c;
>>   }
>>   ...
>>   swap(&a, &b);
>>   ...
>> 
>> ?  swap(), by the way, is a primitive for just about every algorithms
>> text I've ever read.  Does knowing the computer must:
>> 
>>   LOAD r1, a
>>   ADD  something_else_totally_unrelated
>>   LOAD r2, b
>>   INC  some_counter_from_another_place
>>   STOR r1, b
>>   INC  that_same_counter_but_we_just_need_to_fill_the_slot
>>   STOR r2, a
>> 
>> in order to swap two integers aid in the understanding of swap()?
>> I agree that we need to break an algorithm down to primitives, but
>> are you actually saying swap(), for instance, isn't primitive enough?

>I think you're deliberately trying to choose a nasty architecture.
>We're talking about teaching, I think, so why not choose an easy one?

>; --- swap ---
>;
>; Just like void swap(int *a,int *b)

>swap		LDR	a3,[a1]
>		LDR	a4,[a2]
>		STR	a3,[a2]
>		STR	a4,[a1]
>		MOV	pc,lr

>Simple.  (And now there's a 200MHz version of this processor, so it's
>not /that/ old-fashioned.)
>-- 
>[mdw]

>`When our backs are against the wall, we shall turn and fight.'
>		-- John Major
 ^^^^^^^^^^^^^^^^^^^^^^^^--I chuckle every time. Is that a real quote?

Here's what BCC32 -v -O2 shows via TD32 on a Pentium:
-----------------------------------------------------
Turbo Debugger Log
CPU Pentium
swap: void _fastcall swap(int *a, int *b){                 
:0040107C 53             push   ebx                        
#tswap#7:  temp = *a;                                      
:0040107D 8B08           mov    ecx,[eax]                  
#tswap#8:  *a = *b;                                        
:0040107F 8B1A           mov    ebx,[edx]                  
:00401081 8918           mov    [eax],ebx                  
#tswap#9:  *b = temp;                                      
:00401083 890A           mov    [edx],ecx                  
#tswap#10: }                                               
:00401085 5B             pop    ebx                        
:00401086 C3             ret                               
:00401087 90             nop                               
-----------------------------------------------------
Or, taking advantage of 386+ architecture by assembler:
Turbo Debugger Log
CPU Pentium
#swapx#11:  mov ecx,[eax]                     
:00401150 8B08           mov    ecx,[eax]     
#swapx#12:  xchg ecx,[edx]                    
:00401152 870A           xchg   [edx],ecx     
#swapx#13:  mov [eax],ecx                     
:00401154 8908           mov    [eax],ecx     
#swapx#14:  ret                               
:00401156 C3             ret                  
-----------------------------------------------------
Either way, there are additional architecture and compiler/
assembler-specific things you have to understand beyond the
C code representation, which in turn has things extraneous
to the essential algorithm in the abstract. I view the imperfect
match of expression language and idea as a necessary evil you have
to put up with to communicate. Ideally the language will be
appropriate and adequate, and both parties will be fluent.
But some algorithmic poetry will probably not survive translation
to assembler, so a language may have to be learned -- nay, mastered --
before the original can truly be appreciated.
Assembler is only one language, with numerous dialects. But
whatever language one uses to express something, it can be
said of the expression, "Ceci n'est-pas une pipe."

Regards,
Bengt Richter
I wish I knew Latin. I could pontificate so much more elegantly ;-)