html formatting question

[Nate the great]

Originally Posted by rogue_ronin

That's true, but if you find that you are working toward two mutually exclusive targets (mobipocket and the old REB1100, say) you might find that Dale had a point.

You probably aren't doing that, so this may seem irrelevant.

I am of a mind that you write "pure" xhtml, then convert it to your target's quirks using the regex/macros of your html/text editor. Your goal of finding good xhtml that mobipocket recognizes is excellent for simplifying any such process, however.

My unnecessary .02,

m a r

Speaking of the REB1100, do you know where I can find out more about the formatting it uses?

[zelda_pinwheel]

Originally Posted by Nate the great

Speaking of the REB1100, do you know where I can find out more about the formatting it uses?

i would ask nick (nrapallo). if he doesn't know directly, he could probably tell you where to look.

[Sweetpea]

Originally Posted by Jellby

No, in mobipocket dialect "height" means the space above (top padding or margin), and "width" means indent for paragraphs.

But you'll have to put it in with every <p>. I'd rather use CSS and my paddings. Works perfectly.

[tompe]

Originally Posted by nrapallo

MobiPerl code strips all ending </p>'s when converting from .mobi to .html. I see the same thing when my Mobi2IMP program converts .mobi/.prc to .imp as the intermediary .html is left exploded with images.

I do not think it does that. Instead it does not add any </p>. So if they are missing in the .mobi file they wil be missing in the exploded html file.

[Jellby]

Originally Posted by Sweetpea

But you'll have to put it in with every <p>. I'd rather use CSS and my paddings. Works perfectly.

Only if you want anything other than "default".

Anyway, my point is that, while it's certainly better to have a "master" version using CSS and all, that's not directly used in the mobi. Mobipocket Creator, or Calibre, or whatever will probably make some translation and be reliable most of the time, but if you want to control what really goes into the mobi, you have to edit the pseudo-HTML anyway.

[nrapallo]

Originally Posted by Nate the great

Speaking of the REB1100, do you know where I can find out more about the formatting it uses?

Scary, how Zelda knows these things...

Here is the original 'HTML Formatting for RocketEditions' guide (also for the REB1100) in .rb and .html within the below .zip file.

It basically supports HTML 3.2 with some limitations and quirks! The current eBook Publisher software or rbmake will create .rb ebooks from the same source that is used to create .imp ebooks.

Have fun!

[rogue_ronin]

Originally Posted by Nate the great

Speaking of the REB1100, do you know where I can find out more about the formatting it uses?

What are you looking for? There's a great tool here called rbmake that has some pages on deconstructing the file format.

It's the tool I use to make RB files.

The RB format is basically compressed HTML with a generated index file. There are a bunch of ugly hacks contained in it to generate a pop-up menu. That page I linked has a lot more about it.

There are some tricks that I've found for the HTML -- the biggest one is that to have the cover show as the first page, you need to wrap it in <center> tags. Yes, <center>.

Another is that the screen size is 2 pixels larger on the REB1100 than on the old Rocket. The hardware can do 4 grays, I think, but the firmware is limited to 1. (You're better off pre-converting any images, too.) Any link that passes from one page turn to another is broken. I'm sure that there are a few more -- feel free to PM me about anything.

If you're serious about creating hand-edited files in RB format, you should definitely use the tools from the rbmake project. It's both Windows and Linux, although the very latest version is not compiled for Windows (I use Linux, so I can't help you if you use Windows.) You can use the related project files to explode them, build dictionaries, etc.

It's web-aware, and does a LOT of amazing things. Once you get a script well-written, it does just about anything to a source. Like the Rocket/REB1100 it was way ahead of it's time.

If you use Linux, and you're looking for an upload/download tool, you need rebcomm, which you can find here..

m a r

ps: a little off topic?

[Sweetpea]

Originally Posted by Jellby

Only if you want anything other than "default".

Anyway, my point is that, while it's certainly better to have a "master" version using CSS and all, that's not directly used in the mobi. Mobipocket Creator, or Calibre, or whatever will probably make some translation and be reliable most of the time, but if you want to control what really goes into the mobi, you have to edit the pseudo-HTML anyway.

I use one master to make my ePub version and mobi version. They contain only basic HTML elements and extensive CSS. (and personally, I don't like the empty line at the end of a paragraph, so, my default is to remove it )

[Nate the great]

Originally Posted by rogue_ronin

What are you looking for? There's a great tool here called rbmake that has some pages on deconstructing the file format.

It's the tool I use to make RB files.

m a r

ps: a little off topic?

Does rbmake convert <div style="page-break-after:always"/> to the <hr size=0> tag that functions as a pagebreak in RB format?

[Nate the great]

Originally Posted by Nate the great

Does rbmake convert <div style="page-break-after:always"/> to the <hr size=0> tag that functions as a pagebreak in RB format?

Never mind. I've looked, and I don't think it does.

[nrapallo]

Originally Posted by Nate the great

Does rbmake convert <div style="page-break-after:always"/> to the <hr size=0> tag that functions as a pagebreak in RB format?

The ebookwise/gemstar devices don't recognize the "page-break-after" CSS property but only the "page-break-before" syntax.

That said, the .rb's support of CSS is more limited than the .imp's acceptance of same, so in this case, no, the CSS property for pagebreaks is NOT respected. Only a '<hr size=0>' in front (or in it's place) will produce the expected pagebreak.

I give you an example, Robin Hood (PG Etext No. 28700) converted to Paul Creswick - Robin Hood.rb via my GuteBook program uses the 'page-break-before' CSS property by default and it doesn't work on the REB1100.

However, after editing the source files (included in the below .zip) to include <hr size=0> in front of the <h2> and <h3> tags, you can produce another Paul Creswick - Robin Hood-hr-size-0.rb using rbmake to magically "activate" those pagebreaks within the .rb. See rbmake-log.txt for rbmake details.

I also tried to place the CSS property 'page-break-before' inline to see if that made a difference, but it didn't with Paul Creswick - Robin Hood-inline-pb-style.rb and neither did the .imp shortcut for page-breaks, <pb>, work for .rb files!

And there you have it!

[Nate the great]

Originally Posted by nrapallo

The ebookwise/gemstar devices don't recognize the "page-break-after" CSS property but only the "page-break-before" syntax.

That said, the .rb's support of CSS is more limited than the .imp's acceptance of same, so in this case, no, the CSS property for pagebreaks is NOT respected. Only a '<hr size=0>' in front (or in it's place) will produce the expected pagebreak.

I didn't ask if the file format included it; I knew it wouldn't. What I asked was whether the conversion software would replace the div tag with the correct supported hr tag.

The Mobipocket file format doesn't support that div tag, _but_ MobiCreator does. It converts that tag to Mobipocket's proprietary tag. I was hoping RBMake would do the same. I now see that it doesn't.

I also see that the code is written in C, so I might add support for converting the div tag. I'll give the style tag a shot, too.

[nrapallo]

Originally Posted by Nate the great

What I asked was whether the conversion software would replace the div tag with the correct supported hr tag.

Using rbburst, you can explode the .rb that was created with the '<hr size=0>' tag and you will then notice that the tag '<HR NEW-PAGE>' is actually used by the .rb format!

The '<hr size=0>' is NOT ignored by calibre when creating .epub, .lrf, etc... so it isn't useful within .html that is used to create multiple ebook formats (it is ignored by eBook Publisher ).

Fixing rbmake to support this may work for that software, but it won't be available to eBook Publisher or the old RocketLibrarian that also create .rb from .html.

[nrapallo]

Originally Posted by nrapallo

Fixing rbmake to support this may work for that software, but it won't be available to eBook Publisher or the old RocketLibrarian that also create .rb from .html.

It appears a novel (tried and true?) solution does exist to support <hr size=0> and it is to split the file there, ala calibre & .epubs.

Per the Excerpts from A Technical Writer's HTML Template for SoftBook Editions:

G. The Problem of Page Breaks
If your RocketEdition uses explicit page breaks (<HR SIZE=0> tags) that you wish to retain in the SoftBook Edition, you have several options for handling them. If you do not change them, they will appear as thin horizontal lines the width of the display. The most-platform independent option is to divide your HTML into a series of separate files, one for each section that starts with a page break. Both SoftBook Professional Publisher and RocketWriter insert a page break at the start of each new input file....

Food for thought...

[rogue_ronin]

rbmake also supports regex search/replace, built in. Just add a rule to your basic rbmake script (or a separate regex file using the s switch) or OPT file to replace <div style="page-break-after:always"/> with either <hr size=0> or <hr new-page>. Or whatever your preferred output is. (Maybe you want to keep the div, just replace the style?)

You could have quite a list of standard replacements to make your work easier. (You could share too! )

rbmake can be told to recognize <hr size=0> or to ignore it, using the z switch.

If you're going to update the program, I think you want the RBhtml libraries. I've emailed with Wayne Davison a few times, he's super-approachable. (I think he might hang out here, too -- don't know why I think that...)

m a r