[Simh] VAX/VMS

[Johnny Billquist]

On 2016-02-16 17:25, Clem Cole wrote:
>
> On Tue, Feb 16, 2016 at 10:54 AM, <lists at openmailbox.org
> <mailto:lists at openmailbox.org>> wrote:
>
>     Every new IBM machine and OS was designed to preserve the investment
>     in all
>     ​ ​
>     the software and development skills the customer already had. In
>     terms of
>     ​ ​
>     architectural and implementation purity with compatability as a
>     fundamental
>     principle and a 52 year track record of success, IBM wrote the book.
>
>
> ​I agree, but I suggest that you don't sell Intel short.  To be honest,
> I used to think the folks @ Intel had to be brain dead until I thought
> it about and looked more carefully.  Once I started to work for them, I
> really understood.   They deal with with an ugly architecture, but they
> make the best of it for their customers.
>
> You are right Intel killed compatibility many times between lots of
> different impure attempts ( 432, ​
> ​all of their RISC systems etc...), but to Intel's credit - they always
> did a good enough job on compatibility in the
> 4004-8080-8086-80386-INTEL*64 transitions to move the customers programs
> over somehow (again you are right, sometimes easier than others)​.
>
> But the trick is that economics of the Intel family, along with
> compatibility - drove the price of computing down. And Intel was
> compatible enough to have people keep doing it.  The fact is you can
> still boot and old copy of DOS and the programs will run.
>
> As was brought up in this thread, if you take the last VAX made it will
> not boot VMS 1.x and I suspect it will not even run user code compiled
> for it for any really sophisticated user code.
>
>
> More over before DOS, Intel (while hardly perfect) did manage to bring
> 8080 programs to 8086 systems (and 4004 to 8008 and 8008 to 8080).
> Yes, IBM and DEC did it better than Intel did early on and on many
> threads, but over the long haul of their flagship architecture, stuff
> just works.
>
> The bottom line is ugly as it may be, Intel did, does it and the
> ecosystem for their compatible architecture is frankly worth a great
> deal more than S/360 or anything DEC did.  The economics puts Intel in
> the leadership position here.
>
>
> You can agree or not, but my point is not that 8080/x86/INTEL*64 is a
> great architecture (it is not); but that Intel has done an incredible
> job of moving it forward, with binaries continuing to "just work" and
> all while dropping the price all the time.
>
> Unless you are using a cell phone, I'm willing to bet that you are
> typing your messages on a INTEL*64 architecture system, even if the
> processor is not made by Intel.

I agree with all Clem wrote here. Intel can't be blamed for the effort 
in keeping compatibility working. It does work.

However, maybe Intel is not to credit for this (or at least not fully), 
sine the x86-64 is actually a creation from AMD. Intel was trying to 
kill the product by introducing the incompatible Itanium. AMD showed 
that it was feasible to extend the old pig to 64 bits, and the market 
went with them.

In a way, Intel is trapped by its own success. Nothing can replace the 
x86, no matter how ugly the architecture is. The point is, that any new 
CPUs needs to be backwards compatible all the way to the 8088, or else 
they will not fly.

	Johnny