344687 - Spell checker has no "ignore capitalized words" option

Status: RESOLVED DUPLICATE of bug 214518

(Core :: Spelling checker, enhancement)

Description

[Maury Markowitz]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060714 Minefield/3.0a1
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060714 Minefield/3.0a1

Many words flagged as incorrect by the spell checker are actually proper nouns. Under normal circumstances (and all the ones in my particular case), these are capitalized.

Firefox should include an option to ignore the spelling of capitalized words, with the additional sub-option "unless they are the first word of a sentence".

I note that the program already ignores words that start with a number, like "3rd", so it is surprising this option does not already exist.

Reproducible: Always

Comment 1

[Brett Wilson]

I would argue that we should not have this option.

Comment 2

[Maury Markowitz]

Are you sure? Do you happen to drive an RX-7? Do you use a MacPro? Ever heard of the famous Cray-1?

All of these are "errors" according to the spell checker.

Spurious spell checking markup makes the user ignore the real errors due to the forest of false hits.

Comment 3

[Brett Wilson]

No spellchecking program I'm aware of does this. At the beginning of a previous bug post, I typed Thhe, which would be marked as correct using your scheme. This is unacceptable. So you would need to know about the beginning of sentences, which is basically impossible.

To do this properly would require completely understanding sentence structure. If we can't do it correct the vast majority of cases, we should not do it, since it is better to mark right things as misspelled than wrong things as OK. People are used to this. Word has been doing this for 9 years.

Comment 4

[Maury Markowitz]

> No spellchecking program I'm aware of does this.

Argument from ignorance.

Thunder, first released circa '89, offered this feature. I'm pretty sure others did earlier (DOS TSR's), but this was the first realtime checker I used. Since then practically every program has added logic for detecting proper nouns. For instance, my text editor, TextPad, offers these options... 

Auto Correct
Case sensitive
Ignore capitalized words (e.g. England)
Ignore single letters (e.g. b)
Ignore words in capitals only (e.g. ASAP)
Ignore words with mixed case (e.g. TextPad)
Ignore words with numbers (e.g. Win98)

Note that the spell checker in Minefield reports TextPad is a spelling mistake, along with TSR and realtime. With the options from TextPad, two of these would be correctly eliminated with no side effects.

This sort of heuristic is so common that most spell checkers don't even offer it as an option any more. One counterexample is Aspell, which allows the user to turn off its case sensitive algos.

> At the beginning of a previous bug post, I typed Thhe, which would
> be marked as correct using your scheme.

If the word in question was "at the beginning of a previous bug post", that means it WOULD be flagged as an error. Did you correctly parse my report?

Before you complain that it would not flag this error in the middle of a sentence, consider how the error was made: you held down shift and hit T, released shift and typed the rest of the word. The vast majority of times one does this in the middle of a sentence is to type a proper noun or other "special case". Do you disagree with this statement?

> So you would need to know about the beginning of
> sentences, which is basically impossible.

Pfft. Methinks you doth protest too much. If a 8MHz ~1 MIPS Mac could do it, I'm sure this ~20 GFLOPS machine would have no problems. Pattern matching is what a spelling checker is all about, after all.

A sentence starts with a period and then one or more non-vertical breaking whitespace chars (nbsp's do not break). The counterexamples to this rule are abbreviations, but since they are in the dictionary, the rule is modified to consider these.

Before you make silly arguments like this, perhaps you would first take the time to educate yourself on the state of the art in spell checkers. The Wikipedia article is a fine place to start, at least for pointers to other resources. Newer spell checkers do indeed break down the sentence structure, in order to find words that are properly spelled, but incorrect (like "Witch car should we drive?").

Comment 5

[Peter Kasting]

*** This bug has been marked as a duplicate of 214518 ***