Re: [NTG-context] epub

On Sat, Feb 18, 2012 at 9:31 AM, luigi scarso wrote:

On Sat, Feb 18, 2012 at 9:25 AM, Hans Hagen wrote:

...

On 18-2-2012 08:17, Wolfgang Schuster wrote:

...

I would really prefer to use only html tags when a epub file is produced, now the files can be only viewed without problems in firefox (with the epubreader extension) which is nonsense. With the epub 3 standard also mathml is supported and when a reader is used which doesn’t understand it that’s no problem because the standard is AFAIK backward compatible and unknown functions are just ignored.

In that case some postprocessing has to be applied to as it makes no sense to add tons of code to context to produce a less rich output (also a pain to do and it will never be ok). We would end up with ugly kludges like misusing <h1> for whatever we like combined with classes that make it titles, section numbers, inline highlighting, etc ...

The export is not so much html but xhtml and afaik css can deal with that. I woudn't be surprised if future ebook devices could do the same as what firefox/chrome do. Also keep in mind that the (now xml) export can be converted while some html/class/span/div hybrid would be painful to postprocess. Going from rich encoding to poor encoding (rendering) is easier than the reverse. I can imagine that we provide a couple of mappings from export xml to whatever html. http://code.google.com/p/sigil/ has a checker, and it shows some errors. Last time sorry... http://code.google.com/p/sigil/ has a checker, and it shows some errors. Last time I was quite busy to fix ids and something else, but it was before the epub 3.0 which claims to support mathml. I'm still studying the xhtml back end, I hope in the meanwhile that a good epub 3.0 checker appears .

-- luigi