Skeleton CSS

[Hitch]

Originally Posted by snarkophilus

I'm certainly guilty of using headers for choosing font sizes for subtitles, etc. I'm curious to see a discussion about this. At the possible risk of dragging up unpleasant memories, do you have pointer to said discussion? I had a quick look and couldn't find it.

BAD! BAD BAD!

Okay, so...in less strident tones, it was, in fact, someone that I met here, back 9-10 years ago, that "learned me not to do that."

In a PDF, at least, before now, it didn't matter if you used a heading style for text formatting. A PDF really only cared what something LOOKED like, not what it was. (Altho, this is a bad comparison, as PDFs do not spring into existence full-grown, like Pegasus from the head of Medusa; thus, these things did matter, but bear with me.) But it does matter in Word, in all word-processors, html and ebooks. Why?

Because headings are structural, not decorative. They mean something. An h1, for example, is more "important" than an h2. An h3 is less important than that same h2. The use of headings informs the document about the overall structure of the file. Think of an outline. That's the relevance of headings, in a word.

If you want to see the relevance of headings, and why you oughtn't use them for decoration, take a tutorial or class in using Word's built-in Styles and Headings. Let's see...start with this one, I guess: http://www.addbalance.com/usersguide...dingStyles.htm and there are numerous courses online.

But the shorter answer is, they're structural. They have NOTHING to do with how text looks. Nothing at all. You tell a given style how text looks, not its heading style.

To understand all of this, you need to learn about Word's built-in styles and then move to headings. Don't say "I don't use Word," the same concepts apply to all word processors whether they are Atlantis or OO or LO or Word or Wordperfect.

Good luck.

Hitch

[Vroni]

Originally Posted by snarkophilus

I'm certainly guilty of using headers for choosing font sizes for subtitles, etc. I'm curious to see a discussion about this. At the possible risk of dragging up unpleasant memories, do you have pointer to said discussion? I had a quick look and couldn't find it.

<hx> is a semantic tag, it gives the content a special meaning, not a special look or formatting!

[lumpynose]

Originally Posted by Hitch

Unasked-for-comment: if you plan to use your ePUB for Kindle-building, don't use the Blitz CSS. You need a lot more experience under your belt to understand what's happening with all the inheritance functionality in that, and a lot of it will not work for MOBI and it certainly won't for KF7.

Hitch

I do use the Blitz reset.css; I hope that doesn't cause problems. In my main css file I use @import to pull it in from the Styles directory. I've always wondered if it's safe pulling it in that way.

I'm also wondering about the best/safest way to set up what I think of as boxes/blocks that go between paragraphs. I was following the guidelines of standardebooks.org and using blockquotes for everything until tex2002ans pointed out the error of my ways. I couldn't think of a way to make a generic inter-paragraph block with a div since divs can be used in other ways as well so I didn't want to attach any css to a plain div. Then I saw a response on stackexchange about how you can attach several classes to an html tag so now in my css I have div.block for the generic block div, and for the other ones, div.letter (italics), div.notice (sans-serif), etc., and in my html I use 'div class="block letter"'. I think another way to set up the generic div would be to use something like 'div.letter, div.notice, etc. { generic div block stuff here }' and then, as before, div.notice would have stuff specific to it.

An important goal for me is to design the css so that I can use as few classes as possible in the html. So that means combinators, +, >, space, etc. Again, I'm not sure how portable that is.

[snarkophilus]

Originally Posted by Hitch

BAD! BAD BAD!

Thanks, got the message loud and clear .

Originally Posted by Hitch

Okay, so...in less strident tones, it was, in fact, someone that I met here, back 9-10 years ago, that "learned me not to do that."

[ ... ]

Because headings are structural, not decorative. They mean something.

Hitch

Done a little extra reading (thanks for the Word link), and also found this thread about chapter formatting. All making much more sense now.

I have learnt the error of my ways!

[theducks]

I avoid sequential H tags (eg chapter number, chapter title) because the non-wary till end up with a Calibre conversion splitting the file there

I prefer to use a span class= around the title part (allows it to be styled differently) and use a <br />to split the lines
Works nice for header based auto-generated TOC

Hitch probably has a case where this would be frowned upon

[lumpynose]

Originally Posted by snarkophilus

Done a little extra reading (thanks for the Word link), and also found this thread about chapter formatting. All making much more sense now.

I have learnt the error of my ways!

I had read that thread and had started doing the "right thing" but then have backslid and was using h3 tags for subtitle-ish things. I reread that thread after seeing this thread. I now think that Turtle91's method is what I prefer; using a span that's set to block display; no need for a br and I can put top and bottom margins on it. But even that is overkill for what I do; so far the subtitle has been where the short story was originally published so I have it as a paragraph that's formatted like an h3. The h2 and that paragraph are inside/between a header and /header which provides the overall spacing for that header.

One thing that I found out is that instead of having to do the math for the size of the text, as DNSB explains, instead of em I use rem; root em.

[Vroni]

What exasctly do you mean with "sequential" H tags. And does Claibre really breaks a page at every H-tag, even H6? Another reason not to use it.

[BetterRed]

Originally Posted by Vroni

What exasctly do you mean with "sequential" H tags. And does Claibre really breaks a page at every H-tag, even H6? Another reason not to use it.

Only if you configure it that way, the default setting has a common-sense XPath expression Ψ

BR

[Vroni]

Ok, more preciously asked: what is the difference between H-Tags and sequential H-Tags?

[BetterRed]

Originally Posted by Vroni

Ok, more preciously asked: what is the difference between H-Tags and sequential H-Tags?

No idea, maybe the first has an implied 'random' in front of it

I was just setting the record straight regarding what calibre does with H Tags.

BR

[Hitch]

Originally Posted by Vroni

Ok, more preciously asked: what is the difference between H-Tags and sequential H-Tags?

Ducky means doing something like this:

Chapter One (H1)
It Was A Dark And Stormy Night (h2) (or H1)

When you try to format TOCs, automatically, that sort of thing can bollix up the layout AND semantically, it's somewhat wrong. After all, from a structural standpoint, the chapter number-title is a single point; the chapter "head." Right? There isn't outline difference between the two; the title isn't a subhead to the chapter number.

Ducky--I have no problem with how you do that, using spans or whatever. It's "more correct" than using a heading class for it, IMHO. Some people "solve" it by wrapping the entire kit and kaboodle in a heading tag and using a break, but we all know the problems inherent in that.

@Lumpy:

When the goal is to have very short CSS--which is in and of itself, a good goal--you can end up doing what standardebooks.org does, which is using combinators, but there's a huge blind spot there, with the assumption that those will work, no matter what. I'm the first to admit that I'm overly-aware of the older devices, due to my work and constant attention to the older Amazon Kindles, but KF7 devices won't pay a lick of attention to a combined class. I downloaded Sayers' "Whose Body," for example and I see that in their ePUB3 (which was allegedly their ePUB2.0, mind you) they used first-child inside a blockquote coupled with breaks to create two lines of verse.

To me, that's somewhat make-work. Why not use a simple indented paragraph class instead? (Yes, someone could argue that semantically, lines of a poem are more "blockquote" than paragraph, but...c'mon...). They used the same coding for LPW's business card, handed to someone. thus, it's their "go-to" for anything that they feel should be indented and set off with vertical whitespace above and below the cited text.

So...you know, one of the reasons that it's hard for formatting companies to "fix" other businesses' or individuals' formatting is just that--HTML and CSS are very individualized. Sure, a p is a p, and all that, but by the time you figure out what someone else has done, it's cheaper to redo the whole damn book.

Hitch

[JSWolf]

Originally Posted by Hitch

So...you know, one of the reasons that it's hard for formatting companies to "fix" other businesses' or individuals' formatting is just that--HTML and CSS are very individualized. Sure, a p is a p, and all that, but by the time you figure out what someone else has done, it's cheaper to redo the whole damn book.

Hitch

It's astounding how many excess CSS classes there can be in commercial eBooks. I forget which book it was, but it had 142 extra CSS classes. That's ridiculous. I put in my own body and p styles, get rid of all the excess styles, clean up the stupid stuff, go through the XML files to see what's there and when done, I have an eBook that's much better formatted then it was before I got at it.

[lumpynose]

Originally Posted by Hitch

I downloaded Sayers' "Whose Body," for example and I see that in their ePUB3 (which was allegedly their ePUB2.0, mind you) they used first-child inside a blockquote coupled with breaks to create two lines of verse.

To me, that's somewhat make-work. Why not use a simple indented paragraph class instead? (Yes, someone could argue that semantically, lines of a poem are more "blockquote" than paragraph, but...c'mon...). They used the same coding for LPW's business card, handed to someone. thus, it's their "go-to" for anything that they feel should be indented and set off with vertical whitespace above and below the cited text.

Yes, I completely agree. That's what I was talking about when I said that tex2002ans pointed out the error of my ways. I agree that many things could be considered quotes but as you say they're using blockquote for a generic block with whitespace above and below, which I can sympathize with until you instead use the double class trick; e.g., div class="block poem". As far as make-work, if you read their instructions about their whole tool chain, well, my eyes glazed over; made me wonder how many volunteers they really get.

[lumpynose]

Originally Posted by Hitch

When the goal is to have very short CSS

My goal isn't to have short CSS but to have as little as needed in the html, which admittedly does require combinators. So for example, that's why I love the p+p thing for indented paragraphs instead of the usual p class="noindent" for the nonindented ones, or listing all the various places where you don't want it indented. (I learned about p+p here.)

Another goal is to have css that's relatively easy to change for people who don't like my style, which is ragged right and indented and no white space between paragraphs. I now have a comment at the top of the css which directs them to the bottom where they can uncomment a few lines to change those things.

[lumpynose]

Originally Posted by Hitch

KF7 devices won't pay a lick of attention to a combined class.

Do you mean what I'm doing with div class="block poem"?