ePub 2.0.1?

[quillaja]

I know the working group for 2.0.1 finished in April or May. Has their recommendations been accepted? Is ePub now on version 2.0.1? I can't find any information on the IDPF website to indicate or confirm this. Does anyone know?

[JSWolf]

Originally Posted by quillaja

I know the working group for 2.0.1 finished in April or May. Has their recommendations been accepted? Is ePub now on version 2.0.1? I can't find any information on the IDPF website to indicate or confirm this. Does anyone know?

Even if it was approved as the new standard for ePub, it would matter not at all. The ePub standard is only as up-to-date as the earliest version of ADE out in the wild. And that means until Sony updates ADE on the 500, 505, 300, 600, & 700, that's the version of ADE all ePub eBooks have to be compatible with. And I have no idea about other companies and their ADE versions or plans to update.

[quillaja]

Originally Posted by JSWolf

Even if it was approved as the new standard for ePub, it would matter not at all. The ePub standard is only as up-to-date as the earliest version of ADE out in the wild. And that means until Sony updates ADE on the 500, 505, 300, 600, & 700, that's the version of ADE all ePub eBooks have to be compatible with. And I have no idea about other companies and their ADE versions or plans to update.

That's true, but I was interested for other reasons unrelated to ADE or actual devices. Besides, you forgot that the iPad is also now a major player, perhaps more so than ADE.

Anyway, the proposed change to 2.0.1 is just minor things and I doubt would cause any incompatibilities. The planned updates for next year to 2.1 might be another story.

[abdelazer]

The updates to OPF and OPS (two parts of the three that make up the EPUB standard) for 2.0.1 were published by the IDPF on 25 May 2010 at http://www.idpf.org/specs.htm:

* http://www.idpf.org/doc_library/epub....0.1_draft.htm
* http://www.idpf.org/doc_library/epub....0.1_draft.htm

I the final versions will be published after all of the formal review hurdles have been completed, but my understanding is that there haven't been any substantive changes suggested, so they should be about the same.

[charleski]

Originally Posted by abdelazer

The updates to OPF and OPS (two parts of the three that make up the EPUB standard) for 2.0.1 were published by the IDPF on 25 May 2010 at http://www.idpf.org/specs.htm:

* http://www.idpf.org/doc_library/epub....0.1_draft.htm
* http://www.idpf.org/doc_library/epub....0.1_draft.htm

I the final versions will be published after all of the formal review hurdles have been completed, but my understanding is that there haven't been any substantive changes suggested, so they should be about the same.

Executive Summary:
The table of contents is now called a 'global navigation structure'.
And, er, that's it.
Oh, they fixed a few typos and a couple of links.

Any real changes will only happen with revision 2.1, which is due in May 2011.

[pdurrant]

Originally Posted by charleski

Executive Summary:
The table of contents is now called a 'global navigation structure'.
And, er, that's it.
Oh, they fixed a few typos and a couple of links.

Any real changes will only happen with revision 2.1, which is due in May 2011.

There's a significant clarification in http://www.idpf.org/doc_library/epub...htm#Section3.1

New:

This specification assumes a use of selectors that is consistent with the CSS2 Specification. Reading Systems must support all CSS2 selectors, including pseudo-elements and pseudo-classes. However, certain pseudo-classes may not be applicable to all Reading Systems (e.g. hover for devices with a touch screen) and may not result in any rendering on such devices.

Old:

This specification assumes a use of selectors that is consistent with the CSS2 Specification. Reading Systems must support all CSS2 selectors.

(Significant clarification highlighted in red.)

So compliant renderers are required handle :first-letter, :first-line pseudo elements and the :first-child pseudo-class, among others.

This has always been the case, but some implementations have chosen to assume that it wasn't required. This clarification shows that this assumption was a mistake.

[charleski]

Originally Posted by pdurrant

So compliant renderers are required handle :first-letter, :first-line pseudo elements and the :first-child pseudo-class, among others.

They're required to 'support' these (as in, I suppose, not crash when encountering them), but not required to cause them to alter the rendering.

...

[pdurrant]

Originally Posted by charleski

They're required to 'support' these (as in, I suppose, not crash when encountering them), but not required to cause them to alter the rendering.

...

I disagree.

They are required to use CSS2 selectors to determine which of the style definitions in the CSS2 stylesheet to apply to a particular character or characters. They are required to support all CSS2 selectors in making this selection, including pseudo-elements. This means that if the CSS contains

p {font-weight: normal;}
p:first-letter: {font-weight: bold;}

then they are required to use font-weight:bold in the rendering of the first character in a <p> element.

Now, it may be that a particular renderer can only render one weight of font, in which case the first character will not be different from the rest of the element. But if the rendering system can make such a distinction (in particular, if it does it for other CSS2 style defintions), to be a compliant ePub renderer, it must do so in this case.

I raised this issue myself in the discussions, and this clause was put in to remove any doubt that when 'all selectors' meant all.

If you believe that yours is a valid interpretation of the 2.0.1 draft standard, please suggest a wording that would remove the ambiguity that you see.

[charleski]

Originally Posted by pdurrant

If you believe that yours is a valid interpretation of the 2.0.1 draft standard, please suggest a wording that would remove the ambiguity that you see.

Reading Systems must support all CSS2 selectors, including pseudo-elements and pseudo-classes. Dynamic pseudo-classes are the only exception to this and may be disregarded if, and only if, they specify user activity which is not applicable to the interface being used (e.g. hover for devices with a touch screen).

As it stands, the meaning of "not applicable" is undefined and only illustrated by a single example, effectively giving the implementers a lot of lee-way to define it as they wish.