some HLL comments

[Jecel Mattos de Assumpcao Jr.]

On Sat, 11 Mar 95 2:06:34 MET Francois-Rene Rideau wrote:
> Jecel said:
> > Strange... I still haven't received back the last few messages
> > I sent.
>    Err, I do not control listserv very well, and I don't know what is its
> buffering algorithm. If you're talking about your Tuesday messages, its ok now.
> If you have any suggestion about a better mailing list handler, tell me !

I don't know any good listservers. The problem could have been with the
intermediate nodes on the way to ens.fr ( I can't send it mail directly ).
As you said, it is working now.

> > Byte Magazine, vol. 8 no. 9
> > pp. 356-373, September 1983
>    Well, this will be difficult for me to check out ;( ;( ;(.
>    Is there any way to scan and/or OCR/type the page ? Ark.

Too bad. I'll type in an Echonet program from the magazine so you
can see what it looked like ( I got a message from the author saying
it is now calles Purma and that he is still working on it ):

Vacuum one room

 repeat for 15 minutes
   adjust floor brushes for Vacuum
   check Vacuum nozzle
   while Vacuum..isStopped do
     rotate Vacuum clockwise by 90 degrees
   move Vacuum forward for 3 seconds

Capitalized words are objects. It is very verbose, right? Here is what
the Sieve looks like - much more conventional:

benchmark

 uses {{False True}}
 Console line <- "10 iterations.", ConsoleBell

 repeat 10 times
   clear Count
   PrimeSieveFlags__size bytes of PrimeSieveFlags <- True

   for PrimeSieveFlags__size Index's starting at 0 do

     if PrimeSieveFlags [ Index ]. true then
       Prime <- Index + Index + 3
       SieveIndex <- Prime + Index

       while SieveIndex < PrimeSieveFlags__size do
         PrimeSieveFlags [ SieveIndex ] <- False
         bump SieveIndex by Prime

       bump Count
       ;; Console line <- Prime " is a prime."

 Console line <- Count " primes.", ConsoleBell


This code generated 216 bytes for Z80 code that executed directly
on the hardware ( no OS necessary ).

> > [ ideas for changable syntax ]
>    I think that by separating the syntax from the semantics, semantics from
> internal representation, and having automatic rewrite, we can manage all
> these difficulties cleanly.

Maybe having separate HLL and LLL is not such a good idea? How about
extending the LLL to the HLL ( and beyond :-)

-- Jecel