Issue 79445

Summary:	Shouldn't manual hyphenation affect spell check matches?
Product:	Writer	Reporter:	wbaetens <wimbaetens>
Component:	editing	Assignee:	AOO issues mailing list <issues>
Status:	CONFIRMED ---	QA Contact:
Severity:	Trivial
Priority:	P4	CC:	issues
Version:	OOo 2.2.1
Target Milestone:	---
Hardware:	All
OS:	All
Issue Type:	ENHANCEMENT	Latest Confirmation in:	---
Developer Difficulty:	---

Description wbaetens 2007-07-10 15:05:10 UTC

Shouldn't manual hyphenation affect matches found by spell check? For example,
when the AutoSpellCheck is switched on, I type a long word (e.g. in Dutch)
douaneprocedures. AutoSpellCheck will underline this word as unknown. This is a
combination of two existing nouns "douane" and "procedure", which I have both
already added to the dictionary.

If I add a manual hyphenation (Ctrl-Hyphen) between douane and procedures,
AutoSpellCheck still underlines the word. Wouldn't it be better if at that
moment, AutoSpellCheck looks for matches for the two separate nouns, or is this
undesirable?

Comment 1 michael.ruess 2007-07-10 16:24:03 UTC

Reassigned to SBA.

Comment 2 stefan.baltzer 2007-10-10 17:38:24 UTC

SBA: Note: a question is a bad summary. But at least this one did not start with
"WHY" :-)

SBA: Depending on the language, a dash is an important part of the word OR a
word seperator. Last not least: Both in some languages. See issue 64400.

In German language, the problem of "compound words" is similar. Almost (but only
ALMOST!) all words can be "legally" combined. So this is a problem of the
dictionary developer who must built-in the rules for compund words.

Comment 3 stefan.baltzer 2007-10-15 15:52:58 UTC

SBA: Note that "adding words" to user dictionaries NEVER makes their
"combination" a valid word. However, the dash in the middle should meke it 2
words for Dutch.
In OOo 2.3, the word "douaneprocedures" is still marked as wrong, but
"douane-procedures" is accepted (without having the two parts added manually, so
they seem to be in the current dict.).
So this issue could be closed as "worksforme".

SBA->Simonbr: I am a little undecided... Maybe your update (see issue 79051)
already adressed "this example" ("douaneprocedures"). Or even the compound thing
for Dutch "in general"...?

However, please take note of this one and let's see if there is better ways to
add "singel words" to spellcheckers WITHOUT having to file an issue for each in
the future...

Comment 4 simonbr 2007-10-15 21:04:45 UTC

Actually, the latest Dutch spell checking dictionary still doesn't include this
particular word. 

However, this issue is already being addressed in two ways. 

In the OpenTaal project we are working towards using the compound feature of
Hunspell, where this word would be accepted if "douane" and "procedures" are
marked as potential constituents of a compound word. This is a bit tricky; we
want to avoid that this causes spelling mistakes such as "kruitkoek",
"bruidengom" and "rodenkool" (correct spellings: "kruidkoek", "bruidegom",
"rodekool") to be accepted as well. We have some ideas how to achieve this.

We have also started to prepare the next version of the dictionary. It will be
expanded by adding many less frequently used words, which may well include
"douaneprocedures". 

I have my doubts regarding the suggested feature, because manual hyphens can
also be inserted between syllables that are not words on their own.

Therefore I propose to leave the status of this issue as NEW (as I don't plan to
implement the suggested solution) and to resolve it as a WORKSFORME once we have
a new version of the dictionary available that accepts the word
"douaneprocedures" one way or the other.

By the way, it should not be necessary to add words such as "douane" and
"procedures" to your user dictionary, because the Dutch spell checking
dictionary (in any case since the july 2005 version) already contains these.