Any programmer who has worked on more than one
computer system will tell you about subtle
differences among them. As standard as COMAL is, it
also must deal with system variations. ANYONE
PROGRAMMING ON ONE COMPUTER FOR PERSONAL USE DOES
NOT HAVE TO WORRY ABOUT THIS. However, those who
use several systems should make a point of learning
the differences. For example, graphics were omitted
from the COMAL standard because of the
incompatibility of the screen and color resolution
on various computers. HOWEVER, THERE IS A COMMON
SET OF GRAPHICS/TURTLE KEYWORDS.

One difference comes with devices and filenames.
COMAL usually supports the naming conventions of
the specific computer system. Commodore filenames
can contain as many as 16 characters, while IBM and
CP/M names are restricted to eight characters plus
a three character extension. IBM and Amiga also
support subdirectories and time/date stamping of
files. Commodore disk drive names use numbers; CP/M
and IBM use letters; Amiga uses words such as DF0:.
All COMALs call the screen and printer DS: and LP:,
except IBM, which uses the name CON: for the screen
and several names for the printers.

Another difference concerns data storage. Most
computers can use 5 or 3 inch disks, yet each
computer has its own method of storing information
on that disk. Thus an IBM PC cannot read a disk
written on a Commodore 1541 drive.  However, this
is not the only problem. There are ways to transfer
data between different computer systems (including
modem transfers). But even if the computers can
share data, there still is one more factor to
consider.

Different methods are used to represent data and
programs. The user never needs to understand this,
that is one of COMAL's jobs. What is important to
the user is knowing that COMAL has two methods for
storing programs and data on disk. One uses compact
files for quick access. WRITE FILE and SAVE create
these efficient files. They can only be understood
by the version of COMAL that wrote them. COMAL also
supports ASCII text files which can be read by most
computer software including the other versions of
COMAL. LIST, DISPLAY, and PRINT FILE produce ASCII
files.

A debate has been going on for years concerning the
null string and the IN operator. For example, what
should "" IN "abc" return? The COMAL standard
originally stated that the result was to be TRUE
(true has a value of 1). A slight enhancement on
this theme returns the length of the second string
plus one. This value is accepted as true, and it
allows a program to differentiate between the null
string and a string matching the first character in
the second string. Now, Common COMAL specifies that
the null string is not part of any string. Thus

  "" IN "text"

will always be FALSE.

A similar situation involves KEY$. What should it
return if no key is pressed? Power Driver and C64
COMAL cartridge return CHR$(0); other COMALs return
"" (null string).

Another situation involves PRINTing several items
on one line. There are 3 common ways to do this:

  * with no space between the items
  * with one space between items
  * using a preset tab zone system

If you do not use the ZONE statement to set up your
own tab zone intervals all COMALs will be
identical: a comma (,) gives no space between items
and a semicolon (;) gives one space.

However, the tab character has changed from the
comma to the semicolon, and the default zone
changed to match. IBM and AmigaCOMAL use the new
standard. CP/M and Commodore COMALs use the
original method.

BENCHMARK TIMINGS

Here are some benchmark run times from benchmark
tests mentioned in the Compatible Article,
converted into bar charts.

AHLS
A-C   3.1 ######
A-B  12.8 ########################

SIEVE
A-C   6.6 #####
A-B  16.0 ###########

SUBSTRING REPLACEMENT
A-C  0.2 #
A-B  5.1 ###################

A-C=AmigaCOMAL
A-B=AmigaBASIC
Base model Amiga 500 was used for the tests.