Re: [libreoffice-users] Dictionaries.

On 2024-07-15 19:23, Brian Barker wrote:
> At 13:57 15/07/2024 -0400, Chris Johnson wrote:
> I can't say I'm surprised. It doesn't appear in most dictionaries: 

> > I have found, and added, a couple dictionaries, and lots of words.
>
> I'd counsel against that. The more words you include in your spelling 
> dictionaries, the more chance there is of misspellings of some ordinary 
> words failing to be flagged because they happen to be detected as 
> correct spellings of obscure words that you are never likely to use.

On the flip side, nobody wants to see a page of red underlines that 
imply the correctly spelled word is incorrectly spelled.

> Did you see the University of Texas's glossy commencement brochure for its "School of Pubic Affairs"? See 
> https://tinyurl.com/yc8f4cdt . If you are unlikely to use the perfectly  proper word "pubic", it is helpful if it > is *not* in your dictionary.

That is (probably) the result of sloppy proof reading.

> I remember someone on one of these lists constantly bragging about how 
> many words he had managed to add to his spelling dictionary, evidently 
> oblivious of these niceties. 

That might have been me. Between the various dictionaries i utilize, I 
have somewhere between 700,000 and 1,000,000 words listed. Three of them 
are language independent. (Numbers, Place-Names and People-names.) All 
of the others are language dependent.

> No: bigger is not necessarily better,

I constantly throw correctly spelled words into my dictionaries, because 
they are red underlines.  In a page of red underlines, the incorrectly 
spelled word gets lost in the sea of correctly spelled red lined words.

> the ideal spelling dictionary should be the right size, not simply as large as possible.

Back when spelling dictionaries were separate programs, the dictionary 
of one of the organizations had space for 520 words --- the five hundred 
most commonly misspelled words, and 20 slots for the user to add their 
most commonly spelling errors.

At times, I've pondered the usefulness of a LibreOffice extension that 
has that functionality.

> As you increase its size, you decrease the number of false positives whilst at the same time 
> increasing the number of false  negatives. False positives are dealt with by review, but false 
> negatives  will be left as errors.

The LibreOffice extension _Linguist_ will create a list of all words in 
the document, with the number of times it is used. A quick scan of that 
output should pick up words that are out of place. Using that UT School 
of Public Affairs brochure as an example, both pubic, and public would 
be listed. The pubic is obviously out of place.

The other technique is to read the document _backwards_.
Once upon a time I knew how to configure a screen reader to do that.

jonathon

--
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/users/
Privacy Policy: https://www.documentfoundation.org/privacy