[aspell-user] Re: Suggestion for aspell: per-document dictionaries

[Kevin Atkinson]

On Tue, 5 Jun 2001, Reuben Thomas wrote:

> > Aspell doesn't have a per document dictionary mainly because there is no
> > real good way to store them and also because it will add more complexity
> > to the user interface.  However, if you wish you can have a per-document
> > personal dictionary and also use your normal personal dictionary by doing
> > the following:
> >
> > aspell --add-extra-dicts=~/.aspell.english.pws --personal=<per-doc-dict> ...
>
> Bingo, that's what I needed. I was really thinking of this for use with LyX,
> and if there's some standard way to add extra dictionaries on the command
> line, then the LyX crew can work out the other bit of the problem.

Oh OK.

>
> > Perhaps in the future I will have an option to save the session word list.
> > Which will probably be closer to what you are looking for.
>
> Hadn't thought of that; I'd tend not to use such an option, as I can't rely
> on myself in the heat of spellchecking not just to click "Accept for this
> session" on all sorts of junk. Mind you, that begs the question of how I get
> the per-document wordlist in the first place...I agree that yet another
> button "Add to per-document dictionary" is not attractive. Mind you, with
> per-document dictionaries, perhaps I'd not often use a personal dictionary,
> and hence would be happier to have two-stage dictionary maintenance, so that
> during spellchecking I only get the option "add to per-document dictionary"
> and not "add to personal dictionary", and then afterwards I can choose
> another action to add words from the per-document dictionary to my personal
> dictionary.

That is an interesting way to look at it.  Most of the time when someone
clicks "Ignore All" they generally don't want the word to be ignored for
this document which is why I came up with the idea to save the session
word list.

However, Aspell will certainly allow you to due the two stage maintenance as
you suggested above.  To add words to a personal word list all you need to
do is add them to the list.  You can also create a new personal word list
my simply adding the header line "personal_ws-1.1 english 0" to a top of
a word list.  The format of the header line is
  personal_ws-1.1 <lang> <count>
where <count> is the number of words in the word lists.  However, this
number is used for optimization purposes (manly to know how much
memory to allocate for the hash file) and does not have to match the
number of words in the list.  When Aspell saves the word list this number
will be updated automatically.

> This sort of thing might also be a good way to reduce the size of the main
> dictionaries, particularly to eliminate all sorts of words that you might
> not want (you give the example "wether"). Ideally, one would like *only* a
> personal dictionary, which would be quite small (most people, as I'm sure
> you know, have an active vocabulary of under 10,000 words, and even if
> that's root words, that's not much more than 20,000 in total, much smaller
> than even the small English word list).

Having a large personal dictionary (more than 100 words or so) is not a
good idea because personal dictionaries are designed to make adding
entries easy and do not use memory effectually at all.

Making a normal word list is rather easy.  See the manual for what to
do.

If you are interested in using a small base word list than you might want
to check out SCOWL.  SCOWL allows you to creating word lists or all sorts
of sizes.  The default Aspell word lists use 65 as size.  The idea behind
the default word list is to make it complete, not small.  If the word list
is two small most people will get really ignored when Aspell marks
correctly spelled, but slightly uncommon, as misspelled.

---
Kevin Atkinson
kevina at users sourceforge net
http://www.ibiblio.org/kevina/