🚀 go-pugleaf

Charles Richmond <nume...@aquaporin4.com> wrote:
> On 4/27/2018 9:08 AM, Walter Banks wrote:
>> On 2018-04-26 7:47 PM, Al Kossow wrote:
>>> On 4/26/18 4:23 PM, manh...@gmail.com wrote:
>>>> Currently how many programmers coding in machine code? Currently
>>>> how many programmers coding in Assembly?
>>> 
>>> These are unanswerable questions. No one even knows how many
>>> programmers there are.
>>> 
>>> 
>>> 
>> I have a couple data points. I worked on the ISA of a co-processor for
>> automotive and general aviation engine controllers. We wrote a C
>> compiler for this processor. We did not initially provide an assembler
>> for this processor (compile directly to machine code) This a heavily
>> used processor with hundreds of new applications per year (10's million
>> devices per year) Two years after release we released an assembler. No
>> one had ever complained to support.
>> 
>> Since then I know of a single piece of application code written in
>> assembler that has been released for shipment by a customer.
>> 
> 
> I know Walter Banks is *not* discussing an x86 processor....  but...
> 
> People are even more discouraged from learning or using machine or 
> assembly code by the *awful* instruction set of x86 processors.  So lets 
> say that *fewer* people code in assembly language than would code in 
> assembly language...  if the processor had a decent, reasonable, 
> logically devised instruction set.
> 
> 

The thing with the x86 is that the instruction set has improved over the
years. I don't know about x86-64, but with -32 you can largely ignore a lot
of the 8-  and 16-bit cruft, just stick to the 32-bit instructions, and
it's not a completely terrible architecture.(IME)

-- 
Pete