Spell checking...every word is always shown at least once

nabsltd · 08-08-2023, 10:41 AM

Here's a sample of the spell check results on an eBook. This is just a sample, as this happens with all books.

Note that words that are obviously spelled correctly ("as", "at", "be", "bed", etc.) are listed, even though "Show only misspelled words" is checked.

In addition, those words appear much more often than the listed count, but it looks like the spell checker thinks that only one instance of the word is misspelled.

The eBook has the language set to "en" (no qualifiers like "en-US") in both the OPF and each HTML page. In "Manage Dictionaries", "United States" is set as the preferred variant for the English language.

Is there any other config I should look for that might be the culprit?

kovidgoyal · 08-08-2023, 11:40 AM

If you are using the builtin english dictionary and your books actually have language specified as english then it will work. So one of those conditions is not as you think.

theducks · 08-08-2023, 03:22 PM

You may have cross languages set someplace:

The Library view may say English, but the books OPF says something else
<dc:language>en</dc:language>
or the individual HTML

Code:

<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="en" xml:lang="en">

BetterRed · 08-08-2023, 06:17 PM

Look for spurious variants of 'are', 'at', 'been' etc in Tools->Reports->Words, I've occasionally seen something like this due to convoluted markup.

BR

nabsltd · 08-09-2023, 09:41 AM

This is from a book covered by copyright, but I don't think the metadata I'm posting violates the rules...if it does, I'm sorry.

Header on each HTML page:

Code:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
  "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html lang="en" xml:lang="en" xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">

<head>
  <meta content="text/html; charset=utf-8" http-equiv="Content-Type" />
  <title>Chasing the Dime</title>
  <link href="../Styles/stylesheet.css" rel="stylesheet" type="text/css" />
</head>

Header and metadata from content.opf:

Code:

<?xml version="1.0" encoding="utf-8"?>
<package version="2.0" unique-identifier="uid" xmlns="http://www.idpf.org/2007/opf">
  <metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
    <dc:title>Chasing the Dime</dc:title>
    <dc:language>en</dc:language>
    <dc:identifier id="uid">3897963789</dc:identifier>
    <dc:creator>Connelly, Michael</dc:creator>
    <dc:publisher>Little, Brown and Company</dc:publisher>
    <dc:subject>Fiction / Thrillers / General</dc:subject>
    <dc:date opf:event="publication">2002-10-15</dc:date>
    <dc:rights>Copyright © 2002 by Hieronymus, Inc.</dc:rights>
    <meta name="output encoding" content="utf-8"/>
    <meta name="primary-writing-mode" content="horizontal-lr"/>
    <meta name="Sigil version" content="1.9.30"/>
    <dc:date opf:event="modification" xmlns:opf="http://www.idpf.org/2007/opf">2023-08-03</dc:date>
  </metadata>

Here's a paragraph where the word "be" was flagged as spelled wrong...pretty vanilla CSS:

Code:

<p class="para-indent">“Well, it’s occupied at the moment but it might not be for long.”</p>

The "p" tag is directly under "body".

Also note that only the menu item for spell check shows the word as spelled incorrectly. The editor does not purple underline the word.

Karellen · 08-09-2023, 03:42 PM

@nabsltd

In the Calibre Editor, move the cursor through each individual letter of the misspelt word. "be" in your example.

Watch the bottom right corner and see if there are any spurious characters in the word. I know sounds silly, but there are hidden characters that can be added. I was able to show "be" as misspelt by adding a word joiner character. See image below...

nabsltd · 08-10-2023, 09:38 AM

Quote:

Originally Posted by Karellen

In the Calibre Editor, move the cursor through each individual letter of the misspelt word. "be" in your example.

I did not find any special characters, but by using the "change selected word to" and replacing the offending word with itself, the error went away. Unfortunately, this also changed the file in another way.

Before:

Code:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
  "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html lang="en" xml:lang="en" xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">

<head>
  <meta content="text/html; charset=utf-8" http-equiv="Content-Type" />
  <title>Chasing the Dime</title>
  <link href="../Styles/stylesheet.css" rel="stylesheet" type="text/css" />
</head>

After:

Code:

<?xml version='1.0' encoding='utf-8'?>
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="en" xml:lang="en">

<head>
  
  <title>Chasing the Dime</title>
  <link href="../Styles/stylesheet.css" rel="stylesheet" type="text/css"/>
</head>

This change makes the file no longer valid for EPUB 2...it requires the !DOCTYPE directive.

DNSB · 08-10-2023, 01:05 PM

Quote:

Originally Posted by nabsltd

I did not find any special characters, but by using the "change selected word to" and replacing the offending word with itself, the error went away. Unfortunately, this also changed the file in another way.

This change makes the file no longer valid for EPUB 2...it requires the !DOCTYPE directive.

One reason why I use both Sigil and calibre's editor. My last step with editing an epub is to open it in Sigil and use the Mend and Prettify tool (Tools => Reformat HTML => Mend and Prettify All HTML Files). I then run epubcheck and if it says no problems, save and exit.

kovidgoyal · 08-10-2023, 01:13 PM

Quote:

This change makes the file no longer valid for EPUB 2...it requires the !DOCTYPE directive.

No, it doesnt. Only epubcheck thinks it does. No actual epub renderer cares. DOCTYPES are just pointless noise. HTML moved away from them years ago. You should too.

DNSB · 08-10-2023, 01:27 PM

Quote:

Originally Posted by kovidgoyal

No, it doesnt. Only epubcheck thinks it does. No actual epub renderer cares. DOCTYPES are just pointless noise. HTML moved away from them years ago. You should too.

Quite a few publishers still use epubcheck as their gold standard for checking epub files. If it fails epubcheck, they will send it back to you to be fixed. So much easier to add the DOCTYPE vermiform appendix until such time as epubcheck no longer requires it.

When I last looked at the epub2 documentation and dug into the supporting documents, they referenced the XHTML 1.1 documentation which states:

Quote:

There MUST be a DOCTYPE declaration in the document prior to the root element. If present, the PUBLIC identifier included in the DOCTYPE declaration MUST reference the DTD found in Appendix A using its Formal Public Identifier. The SYSTEM identifier MAY be modified as appropriate.

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">

nabsltd · 08-11-2023, 10:55 AM

Quote:

Originally Posted by kovidgoyal

No, it doesnt. Only epubcheck thinks it does. No actual epub renderer cares. DOCTYPES are just pointless noise. HTML moved away from them years ago. You should too.

Regardless of your opinion on !DOCTYPE, it should not have been silently removed from the file during a "replace misspelled word".

Those are my headers, and I did not ask the Calibre editor to change HTML tags. I asked it to replace content within an HTML tag.

kovidgoyal · 08-11-2023, 09:03 PM

Quote:

Originally Posted by nabsltd

Regardless of your opinion on !DOCTYPE, it should not have been silently removed from the file during a "replace misspelled word".

Those are my headers, and I did not ask the Calibre editor to change HTML tags. I asked it to replace content within an HTML tag.

Feel free to not use the editor in that case.

Doitsu · 08-13-2023, 02:07 PM

Quote:

Originally Posted by kovidgoyal

No, it doesnt. Only epubcheck thinks it does. No actual epub renderer cares. DOCTYPES are just pointless noise. HTML moved away from them years ago. You should too.

HTML5 also still requires them:

Quote:

DOCTYPEs are required for legacy reasons. When omitted, browsers tend to use a different rendering mode that is incompatible with some specifications. Including the DOCTYPE in a document ensures that the browser makes a best-effort attempt at following the relevant specifications.

(emphasis mine)

kovidgoyal · 08-13-2023, 11:26 PM

Quote:

Originally Posted by Doitsu

HTML5 also still requires them:

(emphasis mine)

No, it absolutely does not. <!doctype html> prevents HTML documents served up with the text/html MIME type being rendered in quirks mode by *browsers*.
EPUB is not HTML it is XHTML and is not rendered directly by browsers. In XHTML served up with the correct XHTML MIME type, the doctype is not required:

XHTML
If you serve your page as XHTML using the application/xhtml+xml MIME type in the Content-Type HTTP header, you do not need a DOCTYPE to enable standards mode, as such documents always use 'full standards mode'.

https://developer.mozilla.org/en-US/...rds_Mode#xhtml

And even epubcheck agrees with me. It does not warn about missing DOCTYPE in more modern versions of EPUB than EPUB 2.

nabsltd · 08-14-2023, 10:59 AM

Quote:

Originally Posted by kovidgoyal

Feel free to not use the editor in that case.

Do you actually think silently deleting headers during a spell check replace is acceptable behavior?

No other search and replace in the Calibre editor does this...only "Fix HTML" and "Beautify files" make these sort of header changes, and the user would be expecting such changes, and those can be reverted by using "See what changed". I'd argue that neither of these should change valid headers, either, but that's a different issue.

This behavior definitely does not follow the principle of least astonishment.

08-08-2023, 10:41 AM	#1
nabsltd Evangelist Posts: 431 Karma: 7000000 Join Date: Aug 2013 Location: Hamden, CT Device: Kindle Paperwhite (11th gen), Scribe	Spell checking...every word is always shown at least once Here's a sample of the spell check results on an eBook. This is just a sample, as this happens with all books. Note that words that are obviously spelled correctly ("as", "at", "be", "bed", etc.) are listed, even though "Show only misspelled words" is checked. In addition, those words appear much more often than the listed count, but it looks like the spell checker thinks that only one instance of the word is misspelled. The eBook has the language set to "en" (no qualifiers like "en-US") in both the OPF and each HTML page. In "Manage Dictionaries", "United States" is set as the preferred variant for the English language. Is there any other config I should look for that might be the culprit? Attached Thumbnails

08-08-2023, 03:22 PM	#3
theducks Well trained by Cats Posts: 30,170 Karma: 57532200 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	You may have cross languages set someplace: The Library view may say English, but the books OPF says something else <dc:language>en</dc:language> or the individual HTML Code: <html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="en" xml:lang="en">

08-09-2023, 03:42 PM	#6
Karellen Wizard Posts: 1,228 Karma: 5390614 Join Date: Sep 2021 Location: Australia Device: Kobo Libra 2	@nabsltd In the Calibre Editor, move the cursor through each individual letter of the misspelt word. "be" in your example. Watch the bottom right corner and see if there are any spurious characters in the word. I know sounds silly, but there are hidden characters that can be added. I was able to show "be" as misspelt by adding a word joiner character. See image below... Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Spell checking and PageEdit	softfoot	Sigil	3	04-01-2021 09:22 AM
spell checking with two or more languages	mcdummy	Editor	5	07-23-2018 05:43 AM
spell checking	brolny	Sigil	1	09-18-2015 09:38 AM
Multi-lingual spell checking	Stingo	Amazon Kindle	6	11-19-2013 04:58 PM

08-08-2023, 11:40 AM	#2
kovidgoyal creator of calibre Posts: 44,180 Karma: 23000000 Join Date: Oct 2006 Location: Mumbai, India Device: Various	If you are using the builtin english dictionary and your books actually have language specified as english then it will work. So one of those conditions is not as you think.

08-08-2023, 06:17 PM	#4
BetterRed null operator (he/him) Posts: 20,787 Karma: 27405072 Join Date: Mar 2012 Location: Sydney Australia Device: none	Look for spurious variants of 'are', 'at', 'been' etc in Tools->Reports->Words, I've occasionally seen something like this due to convoluted markup. BR

Advert

Advert