On 6/9/2011 4:21 PM, Brett Zamir wrote:
> On 6/9/2011 4:05 AM, stuart yeates wrote:
>> On 09/06/11 07:34, Martin Mueller wrote:
>>> This has been a very helpful and interesting discussion that I think I
>>> sort of understand. But I don't really understand it, and I am
>>> fairly sure
>>> that few of my friends and colleagues in English departments or
>>> libraries understand it either.
>>> Is there somebody out there who can explain layman's language what
>>> is at
>>> stake in using microdata to manage the TEI/HTML5 relationship and
>>> why or
>>> how this would help scholars as they work with textual data in digital
>> The HTML community(-ies) realised that there were existing XML
>> schemas which had lives independent of HTML that HTML was never going
>> to be able to encode. So they created a method for wrapping
>> third-party XML tags in HTML so that the third-party XML can be used
>> in HTML5.
>> This works really well for XML that is effective semantic annotation
>> of running text. Things like <tei:person/> and <tei:w type="verb"/>.
>> These are pretty much the motivating use case for microdata.
> Very well said.
>> How it would work for things like standoff markup and genetic
>> editions, is another question. Certainly you can map the TEI to
>> HTML5. It doesn't, however, solve the question of how you display it
>> to the user.
> This is true, but as with XML, HTML doesn't need to be committal about
> how to display markup (if at all). You can use CSS to hide content,
> meta tags to represent hidden attributes, divs or spans to represent
> the hierarchy of elements.
> For example, to take your examples:
> ...firstly, the namespace would be defined probably on an ancestor
> element, so as opposed to:
> <TEI xmlns="http://www.tei-c.org/ns/1.0">
> ...you might have this:
> <html itemscope="itemscope" itemtype="http://www.tei-c.org/ns/1.0">
> Then to the elements:
> ...might map to:
> <div itemprop="person"/>
> <tei:w type="verb">runs</tei:w>
> ...might map to:
> <span itemprop="w"><meta itemprop="type" content="verb"/>runs</span>
> And where there is no consensus in the community about a logical
> default way to display some functionality such as you mention, it can
> be left to the application to determine. Styles and scripting are well
> supported in browsers with respect to HTML (less so, for XML), so the
> application can have its freedom to do as it likes.
Further to this point, I might add that if XBL 2.0 will be implemented
in browsers like Firefox (at one point, was apparently slated for
Firefox 7, due this year, but the road map does not currently indicate
this), there will be a standard XML-based way to define widgets (using
<xi:include xpointer="..."/> could be made to work for your stand-off
markup, even though XInclude is not supported in browsers, or some
its contents in a particular way, etc.
This would enable reuse of such widgets within the TEI community (and
the convenience of continuing to work with TEI XML) with the ability to
swap out widget implementations.
And yes, theoretically, one might think XBL ought to be able to
represent all of TEI (see http://www.w3.org/TR/xbl/ ), except that the
specification currently (and from my perspective, rather unfortunately)
states, "XBL cannot be used to give a document new semantics. The
meaning of a document is not changed by any bindings that are associated
with it, only its presentation and interactive behavior." Otherwise,
XBL is quite convenient in allowing a transformation to occur behind the
scenes will still preserving the original XML DOM, though it suffers, as
with XSL, in not being widely supported in browsers and would also
probably be prohibited for security reasons on sites like wikis.
> Still, having default mappings of tags like <quote/> to <blockquote/>
> (as in the default stylesheets) and default CSS interpretations is
> helpful even if TEI might not require or officially endorse such
> interpretations (and again, HTML, as with XML can use its own CSS to
> overwrite the defaults).
> Best wishes,