Information on WDSPELL: Jeremy Hughes, 27/04/87


WDSPELL folder should contain:

README         This file
WDSPELL.PRG    Spelling checker
WDLIST.PRG     Dictionary editor
WDCOUNT.PRG    Word counter
WDICT.DAT      Dictionary, currently containing circa 6000 words



WDSPELL.PRG

   This program checks the spelling of every wordform within a text
file against a dictionary of wordforms (WDICT.DAT), and also against a
user-defined glossary of names and technical terms (WGLOS.DAT) and a
second user-defined list (WLIST.DAT).  The latter is intended to
contain wordforms that are used frequently in a particular text but
are not worth including in the dictionary or glossary; it may also
serve as a temporary holding place for words that you later wish to
incorporate in the dictionary or glossary.  WDICT.DAT must be
available for the program to run but the other word lists are
non-essential.  Word lists may be created or supplemented using
WDLIST.PRG.  Wordforms which WDSPELL fails to recognize are listed
alphabetically in a file named after the text file and having the
extension 'WDS'.  Misspelt words will need to be corrected by hand
using an editor or wordprocessor, but any future version of WDSPELL
will probably incorporate a search and correct facility.

   WDSPELL was originally written to check the spelling of a 95,000
word doctoral thesis written with FinalWord, which uses Ascii files,
but it will also work correctly with FirstWord files and possibly with
non-Ascii files of other wordprocessors.  It uses a hashing technique
to sort words and is therefore reasonably fast.  The main limitation
at present is that the dictionary currently contains something in the
region of 6000 words and therefore fails to recognize a large number
of fairly common words, but new words can easily be added.  Note
incidentally that words such as 'realize' and 'recognize' are spelt
with 'z' in the dictionary (this is the preferred spelling of the
Oxford English Dictionary); if you wish to use 's' you will have to
add such spellings yourself.  

   [In case anyone is interested in extending the dictionary to a more
useful size (15000+ words), and making this publicly available, the
following guidelines are suggested: 
   1) Use British rather than American spellings (unless you are
      producing an American dictionary.)
   2) Use '-ize' in preference to '-ise', or consistently enter both
      spellings if you have a strong personal preference for '-ise'.
   3) Omit technical vocabulary which can be included in specialized
      glossaries of technical words.
   4) Exclude all proper names.  
   5) Try to avoid incorporating incorrect spellings.
The obvious way to increase the present dictionary size would be to
run WDSPELL over a number of large text files, and edit the lists of
unrecognized words before incorporating them into the dictionary using
WDLIST.]

* NB WDSPELL does not distinguish between upper and lower case, since
it needs to be able to recognize 'and' and 'And' as instances of the
same word.  All unrecognized words, including proper names, will be
listed in lower case forms.  You should also note that WDSPELL treats
hyphens (and other punctuation) as word boundaries, so you may get some
curious unrecognized forms if you use hyphens at line ends; for examp-
le the word broken at the end of the last line will be read as 'examp'
followed by 'le'.



WDLIST.PRG

   This program may be used to supplement or create word lists for use
by wdspell.prg; type RETURN after 'Listfile: ' if you are creating a
new list.

   The words which you wish to include in a list should be placed in a
file whose name should be entered after 'File to merge: ' when the
program is run.  This list need not be in alphabetical order, and any
duplicate forms are automatically eradicated.  No more than 1000 words
(approx) may be added at a time.

   The program normally abbreviates wordforms to save disk space and
speed up spell checking.  However, it is possible to disable this
process in order to produce a list in readable form (which may then be
edited using a wordprocessor or editor).  In this case you should enter
'y' after 'Disable file abbreviation?  '; hit RETURN (or any other key)
if you do not wish to disable the abbreviation procedure.  If you wish
to produce an unabbreviated list without adding any words from a second
file you should press RETURN after 'File to merge: '.

***** NB It is not advisable to add words using a wordprocessor or
editor, since a word inserted in incorrect sequence will destroy the
effectiveness of the list you are editing.  However, words may safely be
deleted from a list in this way.

   If you wish to write the merged list to the file you have specified
as your listfile you may press RETURN after the 'File to write: '
prompt.  WDLIST automatically backs up your listfile if it is the same
as your output file.

***** NB You should discard your output file if the message 'File write
error' occurs.  This probably means you ran out of disk space, and will
inevitably mean that the output file has been truncated.  Your original
file will be preserved as *.bak and can be re-edited after you have
cleaned up some disk space.  DO NOT RE-EDIT IT without changing the 'bak'
extension or you will probably lose it altogether.



WDCOUNT.PRG

A simple word counting program which doesn't count apostrophe s (or
apostrophe t) as a separate word.  Words are defined as a sequence of
characters containing at least one alphanumeric character and bounded
by white space.  This should work with Ascii or FirstWord files.
