Format idea

[Jamp]

Hi all, no this is not yet another thread about scifi uber format made by nonprogrammer to be never specified, nor coded. I will do it on my own (hopefully), but I want to be sure i am not reinventing the wheel.

My idea basically is that there is bad way of todays formating of ebooks.
Everyone tried converting books and everyone had problems, because formats are too specific. I want technically return to pre-css ages, with XML-based format that gives you set of basic tags to format ebook. Customization and f.e. re-flow will be done client-side, using some css-like file. This is where HTML basically was, but due to much bigger monitor flexibility, css files became very powerful, and too specific, so its now problem to convert book to smaller device. Similarity can be found in every other format, starting in post-script (but its purpose was to be specific) and ending anywhere (maybe excluding plain text).

So, my plan is to make XML-based, hand-editable, opensource, open specification very simple format, and some basic converters. I want just to know that there is not something like it, so I could just help it.

Hope someone understood.

[zelda_pinwheel]

hi jamp, i think you've described what quite a lot of people want for ebooks. but it sounds to me very much like the epub format, which is already being defined. have you taken a look at that ? it's based on xhtml (so, xml, basically) and css, and a couple of format-specific files for metadata and such. you can get more info about it on the idpf site.

you might also want to look at sigil, the epub editor being developped by valloric ; it has its own forum here. if you are a developpeur and interested in helping out with that, i think valloric would be really happy to get some plug-ins or something.

[Jamp]

Hi zelda, I have a bit of knowledge about formats, and as far as I know, epub is basically zipped html with specified specifications to contain (I've skimmed throught specs before). I want to make format that is distributed like the html part only, with e-book specific tags (also some nice modularity) and user, when f.e. importing to calibre or sigil, will just choose (or will be automatically chosen) style file, which will affect things like fonts, borders, image placements, math formulas rendering etc.

So user will just download file that contains:
<book>
<chapter name="Blaaa">
<h1>blabla</h1>
<text>
blabla
<image source=loc, placement=semiexact here> </img>
blabla
</text>
</chapter>
<toc style=h1></toc>
</book>

[zelda_pinwheel]

ha, okay, i see what you mean. hm. well it's a very interesting idea. i am not sure that it will appeal to the majority of ebook readers, since i think only very few actually *like* to play around with code and formatting, and would want to add their own css. most people prefer to have an ebook ready to go, already nicely formatted and looking good, that they won't have to fiddle with.

on the other hand, i think for professional editors and such, who are producing ebooks to sell, your idea is a very good way to proceed, working in semantic xml and adding the css to the code in the last step, depending on format, etc.

who is your main audience, for this idea ?

[user_none]

Sounds a lot like FB2 to me.

[Jamp]

I think major advantage of this format is ability to convert itself. Something like it you can convert to literally anything and get good results, and with some fancy algorithms and bit of users help you can also convert complex things like pdf textbooks to usable file.

I didn't think about target user, but anyone that have problems with some incompatible formats may benefit from it.

user_none: thanks, I will take a look

/edit Okay, as usual, I actually was reinventing the wheel, fb2 is almost exactly what I invented, I will take much closer look at it (and get through some russian)

[ahi]

Originally Posted by Jamp

Hi all, no this is not yet another thread about scifi uber format made by nonprogrammer to be never specified, nor coded. I will do it on my own (hopefully), but I want to be sure i am not reinventing the wheel.

My idea basically is that there is bad way of todays formating of ebooks.
Everyone tried converting books and everyone had problems, because formats are too specific. I want technically return to pre-css ages, with XML-based format that gives you set of basic tags to format ebook. Customization and f.e. re-flow will be done client-side, using some css-like file. This is where HTML basically was, but due to much bigger monitor flexibility, css files became very powerful, and too specific, so its now problem to convert book to smaller device. Similarity can be found in every other format, starting in post-script (but its purpose was to be specific) and ending anywhere (maybe excluding plain text).

So, my plan is to make XML-based, hand-editable, opensource, open specification very simple format, and some basic converters. I want just to know that there is not something like it, so I could just help it.

Hope someone understood.

You might find this interesting:

Nationally Adopted XML eBook Format

- Ahi

[Sweetpea]

That looks a lot like how I format my books. Except I use <div> instead of <chapter> and I use <p> instead of linebreaks for my paragraphs. <book> is <html> and <toc> is <p><a>.

You can make your HTML as simple as you want. I only use stylesheets to add some extras, like large first letters of a chapter, or no indents for certain paragraphs.

[pdurrant]

You should take a look at the Text Encoding Initiative (TEI) and especially TEI Lite

http://www.tei-c.org/index.xml
http://www.tei-c.org/Guidelines/Cust...Lite/index.xml

Don't re-invent the XML wheel. I suspect TEI Lite will do all you wanted to do. There's some good documentation:

http://www.tei-c.org/release/doc/tei...ilite.doc.html

For conversion of your documents to other formats, you should be looking at XSLT http://en.wikipedia.org/wiki/XSLT where by use of an XSLT template, you can convert from an XML document into any other kind of document.

Originally Posted by Jamp

Hi all, no this is not yet another thread about scifi uber format made by nonprogrammer to be never specified, nor coded. I will do it on my own (hopefully), but I want to be sure i am not reinventing the wheel.

My idea basically is that there is bad way of todays formating of ebooks.
Everyone tried converting books and everyone had problems, because formats are too specific. I want technically return to pre-css ages, with XML-based format that gives you set of basic tags to format ebook. Customization and f.e. re-flow will be done client-side, using some css-like file. This is where HTML basically was, but due to much bigger monitor flexibility, css files became very powerful, and too specific, so its now problem to convert book to smaller device. Similarity can be found in every other format, starting in post-script (but its purpose was to be specific) and ending anywhere (maybe excluding plain text).

So, my plan is to make XML-based, hand-editable, opensource, open specification very simple format, and some basic converters. I want just to know that there is not something like it, so I could just help it.

Hope someone understood.

[ahi]

Originally Posted by pdurrant

You should take a look at the Text Encoding Initiative (TEI) and especially TEI Lite

http://www.tei-c.org/index.xml
http://www.tei-c.org/Guidelines/Cust...Lite/index.xml

Don't re-invent the XML wheel. I suspect TEI Lite will do all you wanted to do. There's some good documentation:

http://www.tei-c.org/release/doc/tei...ilite.doc.html

For conversion of your documents to other formats, you should be looking at XSLT http://en.wikipedia.org/wiki/XSLT where by use of an XSLT template, you can convert from an XML document into any other kind of document.

The Hungarian XML format I linked to is, unbeknownst to me until now, TEI or slightly modified TEI.

Thanks for these links pdurrant!

- Ahi

[DaleDe]

Originally Posted by Jamp

Hi all, no this is not yet another thread about scifi uber format made by nonprogrammer to be never specified, nor coded. I will do it on my own (hopefully), but I want to be sure i am not reinventing the wheel.

My idea basically is that there is bad way of todays formating of ebooks.
Everyone tried converting books and everyone had problems, because formats are too specific. I want technically return to pre-css ages, with XML-based format that gives you set of basic tags to format ebook. Customization and f.e. re-flow will be done client-side, using some css-like file. This is where HTML basically was, but due to much bigger monitor flexibility, css files became very powerful, and too specific, so its now problem to convert book to smaller device. Similarity can be found in every other format, starting in post-script (but its purpose was to be specific) and ending anywhere (maybe excluding plain text).

So, my plan is to make XML-based, hand-editable, opensource, open specification very simple format, and some basic converters. I want just to know that there is not something like it, so I could just help it.

Hope someone understood.

Actually this sounds exactly like FB2. Check our wiki for more details.

Dale