[rescue] Bigger Iron at Home (was: SNMP, Baby!)
Geoffrey S. Mendelson
gsm at mendelson.com
Thu Nov 13 03:28:13 CST 2003
Francisco Javier Mesa-Martinez wrote:
> As I tried to explain myself earlier in a previous thread, I happen to do
> architecture for a living... so I am very very very picky. Just because my
> perspective is a little bit different, i.e. most people take an
> architecture for granted, as that is whay they have to work with. I have
> to go deeper and understand why is that the designing team was doing what
> they were doing.
I guess I'm an old fart. I like architchures that program in a cool way.
IBM 360 The instruction set is so rich that except for type
conversions and I/O you could translate COBOL 1 for 1.
Which was what the machine was designed for.
Buroughs cica 1968-1970 Operating system was written in ALGOL.
HP 3000 circa 1974? Operating system written in PL/I or ALGOL
Halfway there: CDC 6x00. Good assembly language, except that it had that
silly 60 bit word.
Apple II. 6502 wasn't bad, but the different roms for
different slots sucked.
Macintosh. 68000 was much better processor, but hardware
was limited, originaly only Pascal available.
Unexamples (flame proof suit on): anything little endian. I could never
understand why except that "IBM did it the other way".
Commodore 64. Assembly language sucked. Too many forward references produced
good object code in the listing, bad code in the executeable. Return codes
not documented. Example 1 = file not found, 2 = i.o error, anything else,
either success or one we missed. :-(
Best current one: The AMD low power chip that is destined to fail. It has
a MIPS core, with the power draw of an ARM. Too bad no one has any software
for it. If you wanted to start a company and build a device around it they
would probably give you a million or two chips to play with.
Geoffrey S. Mendelson gsm at mendelson.com 972-54-608-069
Icq/AIM Uin: 2661079 MSN IM: geoffrey_mendelson at hotmail.com (Not for email)
More information about the rescue