Re: 4p-1.5-wip with just in time compilation

On Aug 29, 10:07 pm, werty <we...@xxxxxxxxxxxxx> wrote:
BTW , sorts are to display data on LCD or maybe make MCU searching
faster .

BS !

I am the worlds greatest programmer , we DO NOT sort , we link ,
because

the MCU is not human , and thus can access stuff fine , using links
rather

than sorted data . Duh , bet they didnt teach you that in the

U. of P.C. !

Besides , if i leave the data unsorted , i can use it´s position as

valuable Kernel info , like where did it come from .

The next level is kernel analyses it , compresses it and places and
links

data . Now i can trash the ¨form¨that i recieved that data in !

You are right.
The other thing is to waste space with informations usually not needed
in a users program.
I understand your concerns.
I dont really sort, I use a hash. I'm not really happy with it. But
the alternative would be to waste a lot of memory to link the things
instead of to collect knowledge about things.
You are completely right.

Btw. there is nothing wrong imho, if kernel info needs some additional
action to get it. See reflect4p and near-dxt to what I mean. For
example the "primitive" compilers of 4p 1.5 are not named. If you
define it with

prim: xyz ... ;


the name xyz is only used to associate a value in an array. It's
nowhere noted, what name is associated with the code position. You
would have to search the code position in that primitves-array to get
the number of the primitive which you can convert to the name using
dxt>name . But that is extra work of course. The representation in
memory is oriented on the user program, not on the programmers needs.
The programmer has a fast enough machine to inspect the array to get
the number and to convert this number into a name. He will also be
hurt soon if he does not have implemented this lookup if he was to
lazy with some of his primitive definitions. Well, I did not need it
until now - I did make a dummy definition:

: JIT-PRIMITIVES ;

That shows

JIT-PRIMITIVES + ...

if you use the .dxt from reflect4p . This helps at least to estimate
where where something was wrong. For a programmer this is enough
information - a user does not need it.
So not everything that is not "linked" is evil.

Regards,
-Helmar


Regards,
Helmar

.

Relevant Pages

  • Re: "Sorting" assignment
    ... is the fastest possible way of exchanging two ... implement the exchange in a sort. ... In your Employee example, you never have to copy actual class ... programmer reasons that if his code is doin' a lotta data movin' (like ...
    (comp.programming)
  • Re: ways to update PS files?
    ... method you can even do random access I/O operations. ... multi-user access (without some sort of higher level manager) would ... And finally, if it is an application programmer type asking, the ... Behalf Of john gilmore ...
    (bit.listserv.ibm-main)
  • DFSORT - simple formula for SORTWKn size?
    ... allocate the SORTWKnn datasets dynamically. ... From my looking, SORT ... ICE084I 0 EXCP ACCESS METHOD USED FOR SORTOUT ... explanation to the programmer who was the one that started me up on this ...
    (bit.listserv.ibm-main)
  • Re: custom sort
    ... Without knowing the real reason, I think it's because two different programmers ... VBA's help for Sort shows this for ordercustom: ... just a guess) is that the sort programmer expected the custom list programmer to ...
    (microsoft.public.excel.programming)
  • Re: Is it possible to delete an element from a sorted array with O(1) time?
    ... Programmer B: Yes. ... Continuation 1 ... Programmer B: Just google radix sort. ... There are some sequence that it ...
    (comp.programming)