
***********************************************************************

                       The Elliott 903
                   Algol library facilities
                   ------------------------

An error has come to light in INV.DAT, the procedure wrongly adapted
from what T.J. Dekker wrote at the Matematisch Centrum in 1965.
It gives silly results, of the order of ten to the seventh.
Thanks to the help of W.L. van der Poel and Herman te Riele who
kindly sent me a copy of the original, the error has been found.
They also mentioned that the series AP204, AP205, AP206 were
dated July 1965 despite what they say in their introductory
comments.

It was the built-in treachery of the "independent compilation"
process, not the code itself, and is caused by the appearance
of the "forbidden" constants 0.0 and 1.0. It is therefore easily 
corrected:

  ALL THESE CORRECTIONS HAVE BEEN DONE (30-JUN-1998)

(1) Replace three occurrences of "0.0" by "0".
(2) Replace one occurrence of "1.0" by "1".

You could also tidy it as follows:

(1) Replace a "<" by a "," and "1963" by "1965" in the
    introductory comment.
(2) Drop the declaration of "l" (little L).
(3) Replace six uses of "l" by "i".

Place it in the library by the following recipe, given here only
to make this note self-contained:

C:\903> copy INV.DAT EIN.DAT
C:\903> Overwrite EIN.DAT (Yes/No/All)?Y
C:\903> sim900al > INV.OUT

Use NOTEPAD to transfer just the numbers from INV.OUT (omitting
the first row of all zeros and all the rows after the one
containing "255" characters) to the end of MODULE.DAT, replacing
what is already there.

C:\903> copy MODULE.DAT EIN.DAT
C:\903> Overwrite EIN.DAT (Yes/No/All)?Y
C:\903> sim900al > INV.RLB

This could take 5 minutes or so. Now insert into LIBRARY.DAT the
numbers from INV.RLB except for the row containing the sequence
... 255  0  0
which acts as a terminator for the library.


DGNH [ed CPB]  08-DEC-1997

***********************************************************************

