Re: follow up on the discussion in HTML5 about metadata access

On Fri, May 6, 2011 at 10:14 AM, Silvia Pfeiffer
<> wrote:

> As you correctly state: a specialised HW or SW eBook reader is
> required. A vanilla browser cannot display the eBook and requires a
> plugin.

I did not use the term "required".  I gave one example of eBook
readers which embed HTML rendering engines.   There is another
approach, which I neglected to mention.  Using DAISY publications as a
use case, I have just consulted with a developer of a DAISY book
reader, running as a Javascripted application within a plain vanilla
browser. No plugin is required, given the capabilities inherent in
HTML and Javascript.  When asked about a metadata API, the answer was
that it would be very useful to them, and it would be used, if

> Once you are at the plugin level, you can provide all sorts of
> additional functionality such as extracting metadata.

But plugins do not appear to be the trend among the talking book
organizations creating book players for the Web.

> This is not a use case to introduce a metadata API into HTML.

I disagree, and will ask the developers to submit specific requirements.


> On Fri, May 6, 2011 at 5:02 PM, Markku Hakkinen <> wrote:
>> The ePub 3.0 document content model is HTML5.  HTML5 UA's will be used
>> to render ePub publications, embedded in hardware or software eBook
>> readers.  Note that ePub is a container (like ZIP) that contains HTML5
>> content and associated resources, which may include any audio and
>> video referenced in the content.  In the eBook world, content is
>> downloaded and read on a local device.
>> mark
>> On Fri, May 6, 2011 at 4:39 AM, Silvia Pfeiffer
>> <> wrote:
>>> EPUB is not HTML, so it does not get interpreted by a HTML UA and
>>> therefore not exposed through the HTML IDL. Even if there is HTML
>>> somewhere in EPUB, you are not delivering a HTML file to the Web
>>> browser but an EPUB file. If you want to interpret EPUB markup in a
>>> Web browser you need a plugin.
>>> Silvia.
>>> On Fri, May 6, 2011 at 1:32 AM, Leonard Rosenthol <> wrote:
>>>> HTML is a markup language that can be (and is!) used in MANY DIFFERENT areas.  To limit it (and it's design/development) to the "Web" is short-sighted and will only lead to interoperability problems in the future.
>>>> Leonard
>>>> -----Original Message-----
>>>> From: Henri Sivonen []
>>>> Sent: Thursday, May 05, 2011 12:37 AM
>>>> To: Leonard Rosenthol
>>>> Cc: Silvia Pfeiffer;;;
>>>> Subject: RE: follow up on the discussion in HTML5 about metadata access
>>>> On Wed, 2011-05-04 at 16:52 -0700, Leonard Rosenthol wrote:
>>>>> > Right now, all use cases discussed on the HTML WG list were solvable
>>>>> > with server-side APIs.
>>>>> >
>>>>> That is NOT true, Silvia!
>>>>> I raised a number of use cases for non-browser-based UAs - for example
>>>>> EPUB viewers - where server-side was NOT an option.
>>>> Why would an .epub book need to be able to introspect its own metadata
>>>> using a script?
>>>> As for viewers, if the viewer wants to do stuff with metadata, it can
>>>> implement whatever interfaces it wants for its own private use. They
>>>> don't have to be standardized or exposed to scripts provided by the book
>>>> itself.
>>>> (I tend to get skeptical when a Web API is motivated by non-Web uses.
>>>> The W3C has been down that road before. Has it ever been a good road?)
>>>> --
>>>> Henri Sivonen

Received on Friday, 6 May 2011 11:02:29 UTC