Spell check question

[phossler]

In the sample text

From 8:00 a.m. to 6:00 p.m. Monday through Saturday

a.m and p.m get flagged as mis-spelt

Same for i.e. and e.g. (just the i.e and e.g part)

I assume it's because spell check assumes that the last period is the end of a sentence

I was going to just add a.m to the dictionary, but got concerned that sometimes there would be an a.m that really was not correct

Any suggestions or ideas?

[kovidgoyal]

yeah that's basically a tokenization issue, the splitting of text into words by ICU doesnt handle these. There's real solution I'm afraid, you just ignore them. Or use AM and PM and that is and for example instead of the abbreviations.

[DNSB]

You are going to need to decide which one is more important to you. Personally, I tend to add a.m., p.m., etc. to the exceptions since I consider the chance of, for example, having a. followed by a stray m to be fairly low.

[BetterRed]

I accept the limitation, I've seen too many am.s and i.es to try second guessing.

For the UK you could use - "From 8:00am to 6:00pm Monday through Saturday"

IIRC CMS allows closed small caps AM and PM.

[Quoth]

Style guides variously have all of the possible options!
AM PM
AM PM in small caps
am pm
A.M. P.M.
A.M. P.M. in small caps
a.m. p.m.

AD CE BC BCE have less variations (small caps is common, the . usage rare and lower case very rare)

The am and pm is pretty common in British English, usually with a space (perhaps a small space in print).

Consistency is important.

[phossler]

@Everyone

Thanks for the information. I had thought about a RegEx to replace 'a.m.' with just 'am' etc. but then i thought that if the 'a.m.' was at the end of a sentence, I'd only make things worse

[theducks]

There are cases, like you mention, that you just never use Replace all.

Search, <eyeball:yes> replace & find
<eyeball:no> Search (a skip)

Chances are that pattern will only appear a dozen or so times. A whole minute to do this way

[phossler]

In my currently being edited book, chapter 1 had 20+ "a.m."s

Some in middle of sentence: " at 9 a.m. they ..."
Some at end of sentence: " at 9 a.m. Next they ..."
Some at end of clause: " at 9 a.m., and then they ..."

Replace All would really foul that up

[Quoth]

You'd need to detect a.m.<space> <Capital> and a.m.<end of paragraph> at least.

Some at end of clause: " at 9 a.m., and then they

That's

9 am, and then they

It's same as a middle of a sentence.

I'd search and then manual replace as I'd not trust myself to to think of all of the combinations and then do a correct regex.

I had an ebook with messed up chapter headings (up to 13, but there was no 5) and the final edit needed a search and manual edit. As well as crazy spans the title of the chapter was on the line above "Chapter <n>" which IMO is the wrong way round. Also no CSS at all, no system ToC and entire ebook was one file. I added a CSS file, replaced styles with classes and then let a Calibre convert to get file per chapter. It also used multiple spaces (deleted all and did indents etc with CSS) and multiple empty paragraphs for layout (added CSS to new class for headings).

[Tex2002ans]

Originally Posted by phossler

Thanks for the information. I had thought about a RegEx to replace 'a.m.' with just 'am' etc. but then i thought that if the 'a.m.' was at the end of a sentence, I'd only make things worse

All the variants of periods/no-periods + lowercase/uppercase AM/PM (and acronyms) were discussed in these topics:

• 2023: "A.M. and P.M."
• 2021: "RegEx: Insert nbs between initials, etc."
• 2020: "Alternate glyph support (font-variant-alternates)"
  
  • Posts #34+.
    

They discussed lots of helpful regexes + edge-cases + tips on how to catch/fix/normalize these types of issues.

- - -

Side Note: And Kovid is right. It's a hard problem with no real solution.

Just Right-Click > Ignore the red squigglies in these few cases + come up with a few regexes to check for the common edge-cases. Like:

Search: \b[APap]\.[Mm],
Search: \b[APap][Mm]\.,

which would check for a.m. + p.m. missing a period followed by a comma.

If you make use of Saved Searches this can be as simple as a single run of a Group.

- - -

Side Note #2: If you want extreme details on "sentence-ending periods" and why you don't want to enable spellchecking periods at end of words... see my discussion in:

• 2022: "Spellcheck Sigil 1.91"

Sigil 1.9.10+ made that change, and I was STRONGLY opposed to it. The amount of clutter and mess it introduced into the Spellcheck Lists was immense. Heavily outweighed by the handful of acronyms like "a.m." + "p.m." you'd have to check.

In Post #21, I even showed graphs of "Sentence-Enders" vs. "Acronyms", where 0.2% hits were "corrected", but 99.8% hits were made much worse.