My biggest peeve with GUIGUTS HTML is the use of <span> tags for indents. These may or may not work in EPUB readers but when you run the EPUBs through Kindlegen they don't work at all in the resulting MOBI files. If I was writing the code I would make it so that an indent of 4 spaces translates to a <blockquote> section. For poetry, (any text where the indenting is uneven) ideally the least indented text would be represented as a blockquote section and further indents would be accomplished using multiple  . Not the easiest thing to code, I know. This is the only way to format poetry that works with the Kindle.

Any line with no ident but with multiple spaces before and two spaces after could be an <h2>.

Single underscores on a line would be converted to <i> and </i>, for odd and even occurrences. That would not be correct every time but it would help. Right now I'm using one underscore to represent <i> and two for </i> and while that works gutcheck reports it as an error.

For the Kindle you might make a paragraph style that has a top margin of 1em. Or you might choose not to use <p> tags at all, but use <div> tags throughout. The Kindle by default makes paragraphs indent on the first line and has no space between paragraphs. If you want to preserve the look of the text file then <div> with a top margin of 1em would work. The Kindle rounds things up to the next em so you don't want to use fractional em's.

You might give us a way to specify dropcaps. For the Kindle and Nook this works:

<span style="font-size:3em">F</span>irst letter.

Kindlegen requires everything to have no more than one class. Second and subsequent classes are ignored.

You might give us a way to indicate we want text surrounded by <pre> tags. This would be helpful for family trees and other ASCII art, code listings, etc.

Bowerbird has been suggesting using light markup in text files so we can derive other formats automatically and reliably. He will doubtless have something to say on the subject.

I've been using guiguts for awhile including the HTML generating feature, but I agree it needs improving.

A file you could use for testing is this one:

It has lots of interestingly formatted poetry.

On Sun, Jan 15, 2012 at 11:37 PM, <hmonroe1@verizon.net> wrote:

I would appreciate some guidance on what adjustments would need to be made to the HTML produced by Guiguts to make it more amenable to conversion to epub format by epubmaker or otherwise. That is what works well in terms of how the table of contents is presented, use of CSS versus styles, page numbering? I have seen some discussion on this list but I did not manage to find this in the list archives. Also, is there an automated tool that could warn of features that will not convert to epub well?

As context I am the lead developer for guiguts although I will be putting less time than I have been. The latest version of Guiguts (1.0.3) is much less buggy and runs out of the box on Windows and Mac. The Windows version has bundled with it epubmaker, the Gnutenberg Press, and the Open Jade SGML parser (onsgmls.exe) to validate HTML and PGTEI. It would be helpful to get pointers to some examples of files that convert well and those that do not. Incidentally I need some assistance in identifying exactly which DTD and other files are needed for onsgmls.exe to validate XHTML 1.1. Perl developers are welcome to help.

Hunter

_______________________________________________
gutvol-d mailing list
gutvol-d@lists.pglaf.org
http://lists.pglaf.org/mailman/listinfo/gutvol-d