RE: [xml-dev] hypermedia affordances

["Rushforth, Peter" <Peter.Rushforth@NRCan-RNCan.gc.ca>]

Liam,

On Fri, 2012-07-06 at 15:59 +0000, Rushforth, Peter wrote:
>  recognition by the XML community of the cultural fact of the
> recognition of the "style" of HTML affordance that is
> recognized globally would be the first step.

> You have to be careful what you standardise.

I agree.  See plugs / receptacles.  But once you standardize on something
 workable, say AC (MicroXML, json serialization etc), maybe it will enable an exponential advance in civilization similar to the advent of transmittable/transmutable power.  The plugs / receptacles already
exist (uniform interface).  Need to adapt to that fact, and get on with it.

> At one time the "obvious and universal affordance" of a hypertext link
> was that it was coloured green - this came from Microsoft Help.

It's still often underlined. 

> The href attribute is pretty obvious and natural to people who grew up
> with HTML, as is the (artificial and unfortunate) distinction between
> href and src.

No, the semantics are different. 
Here's where Marc Andreesen decided to use that string:
http://1997.webhistory.org/www.lists/www-talk.1993q1/0182.html

> The title and alt elements on "a" and "img" elements in HTML are an
> awfully bad design, and should obviously not be used in other
> vocabularies. In time they will probably be supplemented by other
> standard but more powerful methods, just as the label-for mechanism
> arose for making forms accessible and translatable.

I can't say, really.  But since they don't play a role in the underlying protocol,
it seems safe to leave them out.  For now.

> On the other hand, the (disputed) longdesc on the img element is a
> reminder that sometimes one element participates in more than one
> outgoing link.

The dispute is a good enough reason to not include it in xml:.  Once the
waters have calmed, and a good solid RFC defines the concept, maybe then is a good
time to look at it.

> If you want to add features to XML to make it compete with HTML, 

Not compete, collaborate.  

> Don't start out by saying, "those guys are successful so they got
> everything right and if we copy them we'll be successful too" as this
> rarely works out well in the long term.

Evolution is evolution.

> But I think a CG to look at linking for the Web would most likely end up
> making proposals to add features to HTML, with notes on how they could
> be implemented in browsers today using JavaScript and/or CSS, and clear
> explanation of fallback behaviour, rather than hard-wiring some stuff
> into XML.

I think it is easier to add semantics to a language which is built to allow new
semantics to be defined.

Peter