Preventing breaks at the end of a line.

[roger64]

Originally Posted by GrannyGrump

I am currently without a reading device, and using ADE on PC, and I see ending punctuation and quotation marks splitting off to the next line very very often. Especially quote marks.


I don't know if this happens on reading devices, but it is something I have come to expect from ADE.

This is a disaster. The use of no-break spaces (or narrow no-break spaces) allows to efficiently avoid this pitfall... at least in French (because it may not be part of English punctuation rules). You just need a plain regex to apply them.

[roger64]

mistake

[JSWolf]

Originally Posted by patrik

When I use "justify" on my Kobo there is usually a space between quote marks and the text within. I would prefer it to not be (but it's not a big deal, just looks strange).

What firmware are you using and are you reading ePub or kepub? I've not see a space between quotes with ePub or kepub. I'm using firmware 4.5.9587. But, I've never seen this on any of the firmware I've used.

[JSWolf]

Originally Posted by GrannyGrump

I am currently without a reading device, and using ADE on PC, and I see ending punctuation and quotation marks splitting off to the next line very very often. Especially quote marks.

I don't know if this happens on reading devices, but it is something I have come to expect from ADE.

I hope your being without a Reader is due not to it being broken.

What version ADE are you using? I don't recall seeing this on my H2O reading ePub.

[nabsltd]

Originally Posted by JSWolf

an em dash is treated differently.

It shouldn't be. It's punctuation just like any other, and should always "glue" to the previous characters.

I've never seen a typeset book (where things like line breaks are manually controlled) where an em-dash immediately preceded by a letter starts a line, with the letter on the previous line.

Originally Posted by GrannyGrump

I am currently without a reading device, and using ADE on PC, and I see ending punctuation and quotation marks splitting off to the next line very very often. Especially quote marks.

If there is no space before the character that drops to the next line, then, yeah, that's horribly broken.

Assuming no automatic hyphenation using a dictionary, white space is the only place a line should break in HTML. If a break has to be forced because there just isn't any white space (or no white space near enough to allow reasonably sized spaces in justified text), then breaking after a punctuation mark and before a non punctuation mark is the only choice that matches the style of nearly 100 years of typesetting.

My problem is that I read with a renderer that follow line break rules strictly, and won't break after punctuation that often has no spaces around it (em- and en-dashes, ellipsis, etc.), so I get some very weird spacing on justified lines with at lot of those characters. My solution is to put a thin space (&thinsp after those kind of marks. It is almost not visible, but allows breaking where it should happen.

[AlexBell]

Originally Posted by RbnJrg

Hi Alex;

Try using the following code:

...

Regards
Rubén

Thanks, Rubén. I'll keep that in mind as an alternative. But & # 8288; works well on my Kindle after changing the epub file to AZW3 with calibre, and works on my other readers too.

[Tex2002ans]

Originally Posted by nabsltd

It shouldn't be. It's punctuation just like any other, and should always "glue" to the previous characters.

I've never seen a typeset book (where things like line breaks are manually controlled) where an em-dash immediately preceded by a letter starts a line, with the letter on the previous line.

If you've never seen em dashes at the beginning of lines, then you need to have looked at more books. It is perfectly valid to have a linebreak occur before an em dash. See Unicode's Line Break rules:

http://www.unicode.org/reports/tr14/#Table1

or more specifically, this section:

http://www.unicode.org/reports/tr14/#B2

Last year, we also discussed a similar situation with linebreaking around thin spaces:

https://www.mobileread.com/forums/sh...=278187&page=4

Originally Posted by nabsltd

Assuming no automatic hyphenation using a dictionary, white space is the only place a line should break in HTML.

Read the Unicode links above. Really interesting stuff.

In reality, you would want language- and locale-specific linebreaking rules + low-level access to tweaking which characters allow linebreaks before/after. Nothing I know currently does this (LaTeX is probably the closest).

For now, you are mostly at the mercy of the people who coded the renderers (and you do get some abominations like GrannyGrump's examples).

[Toxaris]

In some typography a dash at the beginning of the line can also mean start of a dialogue. So, there is more than one use for the dash at the beginning.

[BetterRed]

Originally Posted by Toxaris

In some typography a dash at the beginning of the line can also mean start of a dialogue. So, there is more than one use for the dash at the beginning.

But in that case the emdash would be immediately preceded by a newline - wouldn't it?

Originally Posted by nabsltd

I've never seen a typeset book (where things like line breaks are manually controlled) where an em-dash immediately preceded by a letter starts a line, with the letter on the previous line.

My emphasis. I read that as being in respect of a 'hot metal typeset to paper' book.

I've often seen the emdash at the end of dialogue (to indicate interrupted speech) dropped to a new line, even though its coded as “Blah blah blah—” it appears as

“Blah blah blah
—”

rather than

“Blah blah
blah—”

I've never seen an ellipsis (used to indicate incomplete speech) dropped to a new line in the same way.

I first saw it in DISOSS and PROFS - in pre-historic times. It's long been my belief that just because IBM made a blunder in their code, some others have been doing the same ever since. Contagion or inheritance - take your pick.

BR

[Tex2002ans]

Originally Posted by BetterRed

But in that case the emdash would be immediately preceded by a newline - wouldn't it?

Not necessarily. For example, Polish uses it all over:

https://en.wikipedia.org/wiki/Non-En...Quotation_dash

Originally Posted by BetterRed

I've often seen the emdash at the end of dialogue (to indicate interrupted speech) dropped to a new line, even though its coded as “Blah blah blah—” it appears as

“Blah blah blah
—”

rather than

“Blah blah
blah—”

I've never seen an ellipsis (used to indicate incomplete speech) dropped to a new line in the same way.

I've seen the em dash + closing quote on its own line. You also might run across something like this:

Original: As Mr. F—— said.
Wrong: As Mr. F—
— said.

or this:

Original: And then …… nothing.
Wrong: And then …
… nothing.

I've also seen crappy breaks like:

Original: Example Sentence..."
Wrong: Example Sentence..
."

This is where the weird edge cases come into play. So you can't just implement broad rules (break after an em dash or always break after a period), you have to take into account what kinds of characters are also in/around them.

One of the best discussions I've come across on the em dash line breaking was this post on the TeX Stack Exchange:

https://tex.stackexchange.com/a/60038

[Notjohn]

Originally Posted by nabsltd

It shouldn't be. It's punctuation just like any other, and should always "glue" to the previous characters.

I've never seen a typeset book (where things like line breaks are manually controlled) where an em-dash immediately preceded by a letter starts a line, with the letter on the previous line.


If there is no space before the character that drops to the next line, then, yeah, that's horribly broken.

Assuming no automatic hyphenation using a dictionary, white space is the only place a line should break in HTML. If a break has to be forced because there just isn't any white space (or no white space near enough to allow reasonably sized spaces in justified text), then breaking after a punctuation mark and before a non punctuation mark is the only choice that matches the style of nearly 100 years of typesetting.

My problem is that I read with a renderer that follow line break rules strictly, and won't break after punctuation that often has no spaces around it (em- and en-dashes, ellipsis, etc.), so I get some very weird spacing on justified lines with at lot of those characters. My solution is to put a thin space (&thinsp after those kind of marks. It is almost not visible, but allows breaking where it should happen.

For a couple of years at least, Amazon Kindle will start a new line before or after a hyphen or an en or emdash even when there is no space. In the early days, I got into the habit of avoiding hyphens whenever I could, and adopting British style for dash (that is, space/dash/space) in order to avoid rivers of white. This is no longer necessary, though I continue to do it because I decided I preferred it. Like the Oxford comma, it's one of those flourishes that makes sentences easier to read.

[doubleshuffle]

This has been a very informative thread. Thanks to all.

Very useful for me in particular, because I like making poetry ebooks, and a dash at the end of a verse dropping to the next line looks plain awful. To avoid it, I've started to use the "nowrap" method proposed by Rubén in post #17. Works great.

And I've come up with a regex that puts the nowrap span around the last word plus dash of each verse:

Find (note that there's a space at the beginning):

 ([^\s]*)—</p>

Replace with (again the space...):

 <span class="nowrap">\1—</span></p>

[Notjohn]

Originally Posted by nabsltd

It shouldn't be. It's punctuation just like any other, and should always "glue" to the previous characters.

I've never seen a typeset book (where things like line breaks are manually controlled) where an em-dash immediately preceded by a letter starts a line, with the letter on the previous line.

The space before and after a dash (usually an en dash) is standard in British publications. Perhaps they glue the dash to the spaces, and the spaces to the words, but I doubt it.

I don't know if Donna Leon is much given to dashes, but I will check the book I'm currently reading.

[Notjohn]

Yeah, here's one, page 194 of The Waters of Eternal Youth, published in Canada but printed in the US, third line from the top ends as undefined --

[Divingduck]

I've just seen this thread. Very informative. Thanks to all.