To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=45913
                  Issue #:|45913
                  Summary:|Spellcheck context menu anf large number of installed
                          |languages
                Component:|framework
                  Version:|680m84
                 Platform:|All
                      URL:|
               OS/Version:|Linux
                   Status:|NEW
        Status whiteboard:|
                 Keywords:|
               Resolution:|
               Issue type:|DEFECT
                 Priority:|P3
             Subcomponent:|code
              Assigned to:|tl
              Reported by:|tl





------- Additional comments from [EMAIL PROTECTED] Thu Mar 24 00:48:55 -0800 
2005 -------
The problem with the context menu of the spellchecker are the entries "Words
is..." and "Paragraph is...".
In order to guess what language it is in the worst case all currently 25
languages get initialized any maybe without even finding a suitable language.
The time problem for loading all this language is not the rela problem. The
problem is that now 20+ langauge are loaded into memory which usually don't get
used since the user
does not use them. And currently nothing is done to get rid of them later on.

For example it would be possible to release those language data when for example
it is not used for some time. If this is good idea is a completely different
question though since the may easily get loaded again when the context menu is
displayed anew.

If memory consumption really is a problem and how big it is needs to be
evaluated though.


Currently, aside from getting rid of those entries, I see three different ways
to improve this:
a) The entries "Word is...", "Paragraph is..." should moveinto a seperate 
submenu,
 and thus the language data should only beloaded if the user explicitly enters 
that
submenu in order to get a language suggested. This way the "normal" use of just
getting suggestions for the wrong word would not implicitly get all languages 
loaded
by default.

b) The list of installed languages should be limited to a reasonable small 
number.
The problem with this is that we currently have our installable langauge modules
be organized by languges without the ability to select only one of those. Thus
the number of actual languages installed may be about twice the number of 
modules.
For example in the english module are en-US, en-GB and en-CA.
That way having even installed only 4 languages modules may easily result in
already having 8 different languages data files.

c) Lernout&Hauspie (where we had aquired the IPR spellchecker) had a single
rather small module that was able to 'guess' the language of a given text.
Having such a small module would completly eliminate the need to use the
lingistic as a workaround to guess a language and thus the memory (and maybe
time problems) would not be available at all.
If we're going with a new technology it would be quite reasonable to ask the
vendor if it at least for SO 9 would be possible to have such a small memory
neutral language guessing tool that is capable to guess/suggest exactly the set
of languages that can be obtained from that vendor.

---------------------------------------------------------------------
Please do not reply to this automatically generated notification from
Issue Tracker. Please log onto the website and enter your comments.
http://qa.openoffice.org/issue_handling/project_issues.html#notification

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to