[Oberon] Hardware driver construction 'capacity building'.

eas-lab at absamail.co.za eas-lab at absamail.co.za
Fri Nov 29 14:32:46 CET 2002


--------
Newsgroups: comp.lang.oberon
Subject: Re (2): Using REAL/INTEGER or LONGREAL/LONGINT

 
> Stefan Salewski wrote:
> > I wonder if the use of REAL and INTEGER datatypes is obsolete on 
> > modern computers.
> > 
> > Any comments?
> 
Michael Griebling <mg at comp-inspirations.com> wrote:
> Embedded micros could still find the INTEGER & even the SHORTINT
> types useful.  Certainly the REAL numbers are much more efficient
> on the 8- and 16-bit micros in common use in these applications.
> 
> Michael G.

Well purely on a practical level, related to oberon S3 - is this
include in "modern computers"/systems ?
You need a capability to keep on generating new drivers for
the new hardware. (This used to be a big factor that kept MS.
ahead of linux).

As part of my effort to salvage dying S3 usage for possible
minimum resource 3rd world usage I want to investigate 
the possibility of telco-inter-communicating between
S3 machines, by 'manchester modem'. 

My reasoning being that manchester modem technology
wrote/read data to cheap cassette recorders and that
the 4 Khz telco bandwidth should also handle this ?

I thought this task could be profitabily intergrated into a
general tutorial (using succesive refinement) on developing
hardware drivers for S3.   The first can-of-worms that I fell
into was cause by exactly the desire to ignore 'word size'
and other lo-level realities for hi-level usage.  eg. 

(* Switch speaker and amplifier on/off *) 
PROCEDURE Tone(b: BOOLEAN);  (* Loudspeaker port address: 61H *) 
VAR s: SET; 
BEGIN 
   SYSTEM.PORTIN (61H, SYSTEM.VAL(CHAR, s)); 
   IF b THEN s := s + {0, 1} ELSE s := s - {0, 1} END; 
   SYSTEM.PORTOUT (61H, SYSTEM.VAL(CHAR, s)) 
END Tone; 

I didn't find a tutorial on S3 style 'assembly', and hardware-type-persons 
don't want to analyse the transformation from: 'complement bits 0,3,5 and
output to port(61)' to the corresponding S3 syntax.

I few lines of 'macro/template' so that one doesn't need to switch the
mind out of machine-code-thinking to oberon-implementation-of-sets
-thinking would be a great asset !?

So yes, if you pretend the lo-level universe does not exist, that's OK
provided someone else handles the reality for you.

-- Chris Glur.




More information about the Oberon mailing list