Kevin Doyle
One of the key elements in AMPLE composition is experiment.
Instruments can be altered, different mixes used, and parts added and deleted.
However, the end result of this can often be a cluttered program
as well as a good piece of music. And if memory is getting short then extra, unused
words need to be identified and deleted.
This utility will show user words which are not in use
elsewhere in the program - you can then decide which ones can be deleted.
The utility is in EXEC format and can be read into an
existing program using the command
*EXEC U.FREE
In order to disturb the existing contents of the screen as
little as possible the screen display is turned off as the utility is read in
and switched back on at the end. (If you press Escape whilst the utility is being
read in you can turn the display back on with 6#OUT or any MODE command).
To use the utility type in the command
FREE
This will then list (in the same form as the SHOW command)
all user words which are not used elsewhere. In a standard music piece some
words (such as 'RUN') will always appear in the list. The 'mix' and 'part'
words will also appear and the use of these can only be checked by examining
the PLAY string in 'RUN'. The utility words 'FREE' and 'FREED' will also always
appear.
Other words, which are recognised as redundant can then be
deleted using
"wordname" DELETE
To delete the utility itself use the command
FREED
The utility uses 6 words out of the maximum allowed in AMPLE
of 125 and therefore it can only be used with user program containing fewer
than 120 words.
The utility works by identifying each user word and then
searching for its occurrence in every user word. If the word is not found in
use anywhere it is displayed on the screen. The time taken for the utility to
do its search is therefore a function of the number of user words and their use
in other user words. For programs with a large number of words, therefore, the
search may take some time - perhaps 30 seconds for a program with 60-70 words.
For those interested in the programming, the utility makes
use of some AMPLE Nucleus words which are not normally user-accessible. To
access these words their tokens have been written directly into previously defined
user words ('#3' to '#6'). These words are initially defined as variables
(GVAR). This means that they occupy three bytes (one for the GVAR token and two
for its data) and that their addresses are known (since GVAR leaves the address
of the data bytes). The contents of the two data bytes are set to &0F00 -
the AMPLE tokens for a zero-length comment - and then the GVAR token is
overwritten with the AMPLE token required.
There are four such AMPLE words used:
"#3" (&2E) - this returns the address of the
next occurrence of a specified byte (or token) in a word. The target byte and
the starting address for the search are input to the word and it returns either
the address where the byte is found and the ON flag or, if the byte is not
found, the OFF flag.
byte startaddr &2E ->
foundaddr ON
or
-> OFF
or
-> OFF
"#4" (&59) - this returns a string containing
a word name from the address of the byte before the word name. Once the address
of the start of a user word definition is known (see below) then 4 bytes can be
added to it to give the address of the byte before the word name. This word can
then be used to return the word name as a string.
wordnameaddress &59 -> wordnamestring
"#5" (&73) - this returns the address of the
next user word in alphabetical sequence. The address of the start of a word
definition and a code (6 for user words, 9 for modules) are input to the word.
Its outputs are the address of the next word plus an OFF flag, or, if no more
words are found, the same address and an ON flag. To find the first word the
input address should be set to 0.
wordaddr code &73 ->
nextwordaddr OFF
or
-> wordaddr ON
or
-> wordaddr ON
"#6" (&B8) - this displays a word name string
padded with spaces to the right to produce an aligned display. It is used by
words such as SHOW and FIND. Word names of less than 10 characters are padded
out to 10; those of 10 or more are padded out to 20.
string &B8
Related file on this disc:
Published in AMPLINEX 013, September
1989
