This is an ED command file script that creates a modified version of
the american.accent file that comes with translator42. The new file,
USA.accent, fixes the incorrect word separator processing of that library.
The CBM translator treats any non-alphabetic character as a word separator,
however, Translator42 only recognizes space, new-line, tab, comma, period,
left and right parentheses, and the dash.
Here is a comparison table for some common word combinations
word-punctuation combinations | American.accent | usa.accent
-------------------------------|------------------|------------
[you, 'you, `you | eey'ow | yoo
tene', tene] | tenna | teen
8K | eightk | eight kay
K8 | keight | kay eight
---------------------------------------------------------------
The problem was fixed in the usa.accent by definition of a special
class "ns" of alphabetic "non-separators" and replacing the default
separator requirement, "$", with "(ns~)", "not a non-separator",
requirement in all statements.
The original american.accent was totally incompatible with my program,
WreadFiles, however the usa.accent restores WreadFiles functionality with
translator42.
The original american.accent file is retained. If you unpack this file
in ram:, all you have to do is type:
ed new with ram:makeUSA.cmd {'new' is a dummy file-name for ed}
to create the new USA.accent file. Use translator prefs to select the
usa.accent as a default in place of the american.accent. If the script
fails, type Esc, then q (enter) and answer Y (no file will be affected).
This same type of change may be made to the other accent files. Just
make sure that you do not change the [\$]=" DAA2LER " equivalent entry.
Note that the same commercial distribution limitations that apply to
the american.accent file also apply to the derived USA.accent file.
IMPORTANT: Caution About Translator42 Usage With General Text
If you are going to use translator42 to read general, un-processed
text, BE SURE to disable the in-text accent specification feature. To do
this, open the new translator prefs and clear the BACKSLASH from the
Specify [ ] Accent name box. Make SURE that all these boxes are BLANK
(delete, delete, enter, in each box and save).
Otherwise each backslash encountered in your text will be interpreted
as a command to load the accent file specified by whatever text follows.
If that text is not an accent file, then a requester comes up stopping the
program or a software failure may occur if it loads an invalid file.
<>MLB<> 23/June/96
......
|
aminet net>