Message-ID: <4318@utzoo.UUCP>
Date: Fri, 14-Sep-84 20:24:49 EDT
Article-I.D.: utzoo.4318
Posted: Fri Sep 14 20:24:49 1984
Date-Received: Fri, 14-Sep-84 20:24:49 EDT
Organization: U of Toronto Zoology
Lines: 56

     I have read some of the recent news on net.micro.ti and have noticed that
some discussion of CALL PEEKs and CALL LOADs has taken place.  I know of a few
memory locations that may be useful to anyone reading the news for this  news-
group.   There  is a two-byte address which can give you information about how
much memory you still have available, when you  have  XBASIC,  without  memory
expansion.   I  do  not  know  if  it will report memory available when memory
expansion is attached, as I do not have this.  The advantage of  this  command
over  SIZE is that it can be incorporated into a program to report memory free
during the actual run of a program.  I obtained this information from  an  old
issue of COMPUTE magazine.

            CALL PEEK(-31974,A,B)::PRINT A*256+B-2455 should work.

     The 2455 refers to the number of bytes of VDP RAM  used  by  the  machine
when  XBASIC  is attached (i.e.  16,384 bytes - 2455 = 13,929).  The addresses
are in the CPU Scratch Pad RAM.

     Someone was interested in the fact  that  you  could  use  a  number,  by
itself,  in the IF statement.  In case no one else comments on this, I thought
I would tell what I know.  When the condition following the IF is evaluated, a
-1 is returned for a true condition and a 0 is returned for a false one.  If a
single variable follows the IF, its value is examined.  If the value is -1 (or
not 0) then the instruction after the THEN is followed.  If the value is 0 the
instruction after the ELSE is followed (or if there is no ELSE,  the  instruc-
tion in the next program line is followed).

     Another useful thing to know if you are  interested  in  compacting  your
BASIC  or  XBASIC  program  is that relational expressions can be incorporated
into equations.  An example that I have used  this  for  is  the  rounding  of
numbers.   You've probably heard of the rounding method in which you shift the
decimal point to the right of the digit at which you wish rounding  to  occur,
then add .5 to the number and take the INT of that number, and then shift back
the decimal point to its correct position.  This method is fine except for the
fact  that  numbers  ending in exactly .5 (after the decimal point is shifted)
are always rounded up.  If you had a large number of such  values,  you  would
bias  later  calculations  based  on those numbers.  I have incorporated rela-
tional expressions into an equation that permits  the  computer  to  round  up
values  ending  in .5, if they have an odd digit to the left of the .5, and to
round down such values, if they have an even digit to the left of the .5.  The
same equation permits the normal rounding up or down of other numbers as well.
All of this is achieved in a single program line.  The line looks like this:

                IR=IR-(R-IR>.5)+(R-IR=.5)*(IR/2-INT(IR/2)<>0)

     The R variable is assigned the value of your  number  after  the  decimal
point  has  been shifted, the IR variable is previously assigned the INT of R.
In this equation, IR is assigned the value of the number  after  it  has  been
rounded  off, but the decimal point has still to be shifted back to its origi-
nal position.

     There are probably plenty of other uses for such a technique, which might
save  you a line or two of BASIC code here and there.  If anyone has a simpler
method of rounding numbers, I'd like to hear about it.  This method may not be
the  best and I'd appreciate any good suggestions.  Now to send this before it
becomes far too long.