Re: Seeing VERSIONINFO under Vista?

In article news:<dt6c53t5nqoaf5kch8g1qegljd6e3kl2t1@xxxxxxx>, Joseph M.
Newcomer wrote:
Don't blame IBM; they had very limited options at the time.

Oh come on ... you're telling me that IBM couldn't have come to a deal with
Motorola whereby they (IBM) would have set up a new fab line for 68ks if
they had really wanted to? IBM have done deals on fabrication with Motorola
in the past, so I don't suppose there would have been any objections to
such a deal on principal from either side.

The problem is that IBM, corporately, didn't think that the PC was
important enough to warrant the investment.

Yes, I blame IBM.

The other side of the story, though, is that most business software for
desktop computers (PCs in the broad sense) was written in assembler for
CP/M-80 running on 8080, 8085 or Z80 chips. The 8086 family had a register
set and an instruction set that -- although not identical to those of those
8-bit chips -- was sufficiently similar that automatic translation of
working applications could be effected to bring software quickly to market.
Intel even marketed a tool (conv86, IIRC) that achieved this translation
from intel 8080/5 assembler source to 8086 source.

If IBM had gone with the 68k much of that business software (wordstar,
visicalc, etc.) would have had to be rewritten from scratch, and the
time-to-market would have been much longer.

The company I worked for at the time sold interpreters for the Prolog logic
programming language on CP/M (and unix). We wrote our own translator in
Prolog that translated Z80 macro assembler to 8086 (non-macro) assembler,
mapping registers, changing opcodes, and expanding macros as it went -- we
even did some peephole optimization. We assembled the result on digital
research's 8086 assembler for CP/M-86 and the code just ran.

[Actually, it didn't run /correctly/. There were two places in our code
where it mattered that a DEC instruction on an 8080 may set the carry flag
but a DEC instruction on an 8086 doesn't, and we had to correct those by
hand. One case was in the integer divide routine which we replaced by an
8086 DIV operation anyway, the other led to a subtle garbage-collector bug
that took about 10 minutes to find.]

Cheers,
Daniel.


.

Relevant Pages

  • Re: SHARE Requirements and IBM
    ... assembler features (duplicate and overlapping USING flagging, ... rejections were announced. ... community both as a user and as an IBM representative. ... Committee regulars in early 1991: ...
    (bit.listserv.ibm-main)
  • Re: SHARE Requirements and IBM
    ... assembler features (duplicate and overlapping USING flagging, ... rejections were announced. ... community both as a user and as an IBM representative. ... Committee regulars in early 1991: ...
    (bit.listserv.ibm-main)
  • Re: Best practice
    ... I am amused/concerned that IBM has obsoleted many of its own assembler ... For IBM-MAIN subscribe / signoff / archive access instructions, ... send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO ...
    (bit.listserv.ibm-main)
  • Re: Non-Standard Mainframe Language?
    ... Phil Nesbit if anyone remembers him - over the possibility that the Assembler compiler may not be delivered with the operating system which was to continue to be free - or "at no additional cost with the hardware". ... Didn't I see something recently about IBM getting more than half its revenue from software these days - or something like that? ... For IBM-MAIN subscribe / signoff / archive access instructions, ... send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO ...
    (bit.listserv.ibm-main)
  • Re: Assembler - any recent IBM manuals have quick-start tutorial?
    ... When I first used assembler it happened to be with S/360 DOS ... I guess IBM didn't publish a tutorial manual since it was protecting ... IBM manuals have a quick-start example, sort of similar to like how ... I'm digging my way through the IBM MVS manuals, ...
    (bit.listserv.ibm-main)