On 14/12/11 12:42:38, Nick Dokos wrote: > Alan L Tyree <alanty...@gmail.com> wrote: > > > G'day, > > > > Here are the modifications that I needed to make to the exported > html > > file (using the standard exporter) in order to get a valid ePub > > document. http://threepress.org/document/epub-validate provides an > > on-line validation service. > > > > My document is relatively simple, but does have footnotes, > citations > > > and formatting commands. > > > > I will be happy to add a fuller description to worg if someone will > > point me to some instructions. > > > > > > 1. Make the html file have utf-8 encoding; there are probably > several > > ways to do this, but I used the emacs23 menu MULE options on the > org > > > document. > > > > 2 Change the language statement by removing the "lang=" attribute: > > > > From: <html xmlns="http://www.w3.org/1999/xhtml" lang="en" > > xml:lang="en"> > > > > To: <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en"> > > > > 3 Change anchor descriptions associated with section headings by > > removing "name=" entries, for example, > > > > From: <a name="sec-1" id="sec-1"></a> > > > > To: <a id="sec-1"></a> > > > > NOTE: this can now be done by customising org-export-html-headline- > > anchor-format (Thanks Bastien!!) > > > > 4. Change anchor class="footref" by replacing "name=" with "id=". > For > > example: > > > > from: <a class="footref" name="fnr.2" href="#fn.2"> > > > > to: <a class="footref" id="fnr.2" href="#fn.2"> > > > > 5. Change anchor class="footnum" by replacing "name=" with "id=". > For > > example: > > > > from: <a class="footnum" name="fn.1" href="#fnr.1">1</a> > > > > to: <a class="footnum" id="fn.1" href="#fnr.1">1</a> > > > > Interesting - I downloaded the epub validator, ran Calibre on a silly > little org-produced HTML file and duplicated your "lang" problem. > > AFAICT, all of the problems you mention above are legal HTML4, so > either > the epub spec (which I have not looked at: do you have a pointer?) is > made to trip people up by enforcing restrictions that they dreamed > up, > or the validator is not quite as smart as it should be.
Yes, the org export file validates at w3c as a valid XHTML1.0 file. The ePub spec is here: http://idpf.org/epub but I admit fatigue in trying to wade through it, so I don't know what XHTML1.0 constructions it objects to. It seems like the kind of document that nobody has ever read: sort of like EULAs and express warranties. I just made the changes in response to the validation errors. > > Just for kicks, I ran it against the Calibre Quick Start Guide (the > only > "real" ebook I had access to) and it gave me an error :-) > > ,---- > | Epubcheck Version 3.0b3 > | > | ERROR: Calibre Quick Start Guide - John > Schember.epub/OEBPS/content.opf(5,57): attribute "file-as" not > allowed > here; expected attribute "id", "opf:file-as", "opf:role" or > "xml:lang" > | > | Check finished with warnings or errors! > `---- > > so, presuming that the ebook is "fine", I'm not sure how seriously to > take the validator. There is something to that. Calibre explicitly disclaims any intent to produce an epub file that validates. In fact, the FAQ seems to suggest that a Calibre produced file is *better* than one that validates: http://manual.calibre-ebook.com/faq.html#the-epub-i-produced-with-app-is-not-valid The problem is that some marketing sites require validation, although I am not sure how much that is enforced. > > Nick > > PS In case it's not obvious, all I know about ebooks is what is > contained in this email. Caveat emptor. Ahhh, what about statutory warranties :-). Cheers, Alan > -- Alan L Tyree http://www2.austlii.edu.au/~alan Tel: 04 2748 6206