learning spelling in rest of document?

For all things Mellel

Moderators: redlers, Eyal Redler, Ori Redler

Post Reply
FA1
Knows everything, can prove it
Posts: 66
Joined: Sat Sep 02, 2006 11:36 am

learning spelling in rest of document?

Post by FA1 » Sun Sep 17, 2006 12:14 am

I right-clicked a word and chose "learn spelling." The rest of the instances in the document remained highlighted even if I checked the whole document for spelling. Is that intended? When I right-clicked on another instance of the word, without doing anything else, then it was no longer highlighted.

sloehr
Got the styles thing figured out
Posts: 6
Joined: Sat Sep 16, 2006 3:47 am

Post by sloehr » Wed Sep 20, 2006 12:06 pm

I have the same problem, but I find it is something that happens also in Mail.app, especially when I can languages in the dictionary. May not be a problem with Mellel?

alexwein
Read the guide, knows everything
Posts: 54
Joined: Tue Oct 25, 2005 9:31 pm
Location: Portland, Oregon, USA

Post by alexwein » Wed Sep 20, 2006 4:12 pm

I notice this happens in programs that use Apple's spell checker. Once you have told it to learn a word, try clicking on an instance of the word that still has red highlighting or underlining and the highlighting should disappear. I think it's just a bug in the system spell checker, but someone may know more about that.

rpcameron
Knows everything, can prove it
Posts: 978
Joined: Wed Oct 26, 2005 12:48 am
Location: IE, CA, USA

Post by rpcameron » Wed Sep 20, 2006 5:41 pm

The system scans the typed word when it first appears, and then checks it against the dictionary. If the dictionary is modified later (such as adding a new learned spelling), then the word needs to be rescanned against the dictionary. Because that would be incredibly intensive on the system to constantly check every word non-stop, it is only checked at two points: first when the word is first typed, and second when the document is scanned to check spellings.
— Robert Cameron

joewiz
Knows everything, can prove it
Posts: 199
Joined: Sun Oct 23, 2005 9:42 pm

Post by joewiz » Wed Sep 20, 2006 7:04 pm

A similar issue came up in a Chinese-Mac mailing list thread in Dec. 2005. Maybe this info will help:

I was fed up with the built-in spell checker's inability to learn words and sent the following to Apple through the Mac OS X feedback page:
Mac OS X's built-in spell-checking system suffers from a major bug:

1. Words that contain accent characters (such as "Tōkyō") are not
recognized as *single* words. TextEdit, using "Tōkyō" as an
example, underlines "ky" but not the rest of the word. "Ōsaka"
produces an underlined "saka". "Běijīng" produces "ij".

2. As a result, words like this that are split up by the spell
checker cannot be "learned". Control-clicking on these words
brings up a contextual menu that only allows the underlined portion
of the word to be learned - "ky" or "saka" or "ij", in these
examples. [Counterexamples: "fèihùa" is completely underlined and
can be learned.]

This is a long-standing bug in the spelling system, and it
seriously inconveniences my work, which requires the use of
Romanized Chinese and Japanese text. Please, please correct it
soon! I would really appreciate it.
Then Tom Gewecke wrote to me:
Joe -- It's perhaps not totally surprising that Apple's spell checker
doesn't work right for these, since it doesn't include any languages
that need more than Latin-1 (and Apple has so far shown zero
inclination to expand it). Have you tried CocoaSpell, which does
have a broader range?

http://people.ict.usc.edu/~leuski/cocoaspell/home.html
Then I realized that while CocoaSpell helped, it had its own problems:
Thanks so much! CocoaSpell did the trick! I thought CocoaSpell would only contain a different set of words; I didn't realize it had a different logic set that could get around the built-in spelling engine. What a relief.

Upon closer inspection, CocoaSpell appears to ignore accent marks entirely. For example, it will underline the entire word "Yosaburō", which I can then tell it to learn; but if I type "Yosaburo" sans micron, it doesn't flag the word as misspelled. I think I'll write to CocoaSpell's developer to see if the accent marks issue can be addressed. Still, I much prefer CocoaSpell (2.0.2 beta) so far to Apple's buggy Latin-1-centrism.
So I wrote to Anton Leuski, the developer of CocoaSpell:
A fellow Chinese-Mac mailing list member pointed me to CocoaSpell to solve a problem I was having with Apple's built-in spell-check engine.

The problem was that Romanized Chinese and Japanese text, which contains a fair number of accent marks, wasn't being recognized well at all and could not be 'learned' by the engine. I wrote to Apple about the issue today.

I'm very happy that CocoaSpell doesn't suffer from the same bug. But there is one issue that I'd be interested to know your opinion about. CocoaSpell appears to ignore accent marks entirely. For example, it will underline the entire word "Yosaburō", which I can then tell it to learn; but if I type "Yosaburo" sans micron, it doesn't flag the word as misspelled.

Is sensitivity to accent marks something that could be toggled via CocoaSpell's preferences?
Anton replied:
does it happen only with the words you learn or with the words in the standard Japanese dictionary? I mean if "Yosaburō" was a common word in the language what would be the behaviour? I'm trying to figure out, if it's something that is inherent in the dictionary, or I'm doing things in my code...
And I answered:
Thanks for your reply! I am actually using the English dictionary; I'm typing my dissertation in English, and I commonly type Romanized Chinese and Japanese names. Yosaburō is a Japanese given name, and the macron over the o is very common in Romanized Japanese text. Does this help? I would be happy to try to be more specific, if I'm not giving the right information.

I also just noticed something interesting as I was typing in Mellel. I typed: "wasn’t", and I noticed CocoaSpell flagged "wasn". I realized that CocoaSpell isn't recognizing the curly apostrophe ("typhographer's quotes" as Mellel calls them). When I force-typed a straight apostrophe, CocoaSpell didn't show any problem with "wasn't". Do you think it would be possible to expand CocoaSpell's tolerance for curly apostrophes?
And that's where our conversation ended. So using CocoaSpell will get you 'learned' words, but it doesn't understand typographer's quotes very well.

Post Reply

Who is online

Users browsing this forum: No registered users and 9 guests