Hi Clem,

-=> On 29 Jan 97  07:01:56 Clem Clarke said to All <=-

 CC> WHY LIFE STOPS AT 64-KB
 CC> I  OCCASIONALLY INDULGE WHAT IFS ABOUT THE COMPUTER INDUSTRY...

 CC> lntel's 8088 with its horrible segmented memory
 CC> architecture, IBM consigned the PC industry to a fate
 CC> equivalent to that of those unfortunate Chinese girls who
 CC> had their feet bound at birth, and it's been hobbling ever
 CC> since.

Incidentally the segmented mess was ripped off from DEC who invented
it to retain compatibility with early machines that could only address
48k.
 CC> It's because of 64Kb that you keep running out of
 CC> resources when you try to run more than one  component of
 CC> It's those two innocent words 'backward computability'
 CC> that are the cripplers. 

Correct BUT all those manuf's that sacrifices backward compatibility 
ALWAYS lost their entire customer base - usually within a year.

 CC> It hindered the introduction of GUIs to IBM hardware
 CC> for years.  and it blighted the chances for advanced
 CC> programming languages like LISP and Prolog. Worst of all, it
 CC> was almost certainly responsible for the commercial triumph
 CC> of the C programming language, a disaster from which the
 CC> software industry may never recover.

The C language succeeded for reasons unconnected with segentation:
Primarily becase it is so low-level that experienced assempber
writers can predict the instructions that will be generated just as
accurately as writing assembler. 
The other languages mostly failed because:

a) Its extremely easy to write code that is unbelievably inefficient

b) The language is so un-human like that programmers having learned it,
   forget all they have learned before they can finish their first program.

c) They don't support the concept of language-independent run-time libraries.

d) IBM decided to ignore them because of NIH. 

The language that solves all known problems was developed in the late
'60's - Algol-68. This had all the benefits of all the languages known
at that time (and no new features have been invented since AFAIK). It
died because IBM wanted to ship PL/1 instead - which ahd all the defects
of all known languages of the time instead. Algol-68 was not written in 
America, and not by IBM - so it was a dead duck.

 CC> What prompted this particular outflow of spleen was
 CC> reading the 'Report of Inquiry Into the London Ambulance
 CC> Service (Feb 1993)'.  On 4 November 1992 London lost all
 CC> effective ambulance cover because of a catastrophic crash of
 CC> its new computerised dispatching system.  This software  was
 CC> written in a mixture of Visual Basic and C, running under
 CC> Windows 3 on a network of 486 boxes.

Yes but there's more to this story than meets the eye. When the bid to
develop this system was submitted, every consultant in the whole of the
south east said the system would not fly, but the buyers said 'its cheap'
and bought it anyway. There are strong suspicions of nepotism and 
corruption in the selection process, even if not voiced in public. I
worked for a company that submitted a competitive tender, having 
succcessfully shipped a number of other similar systems at the time. 
(I did not work on that kind of equipement myself.)


 CC> The inquiry team concluded that the crash was caused by
 CC> a memory leak - I quote from para 4039: '...the programmer
 CC> had inadvertently left in the system a piece of program code
 CC> that caused a small amount of memory  within the file server
 CC> to be used up and not released'

Systems are always cheaper if you don't bother to test them. Unfortunately
the old engineering saying 'if it ain't tested, it don't work' is
generally found to be more accurate than code written in C.

 CC> In my alternative What if... universe that system might
 CC> have been written in Oberon 2 (the current descendant of
 CC> Pascal featuring an automatic garbage collector which
 CC> prevents memory leaks) running under Bortech Windows 11.0 on
 CC> a 250MHz PowerPC 9064.

The alternative systems would almost certainly been written in C on either
Intel SBCs or DEC kit. I think you will find that all the competitive bids 
proposed such solutions.

Andrew



... Old age is 15 years older than I am.
--- Blue Wave/Max v2.30 [NR]