Re: document node attributes

How about HLink then?

http://www.w3.org/TR/hlink/

----
Stephen D Green


On 5 July 2013 13:16, Rushforth, Peter <Peter.Rushforth@nrcan-rncan.gc.ca>wrote:

> Hi,
>
> > > [Liam wrote:]
> > > > No, which is why it's not going in XML core.
> >
> >
> > I don't speak for the XML Core Working Group. However, a "one
> > eature at a time" approach isn't going to fly.
>
> I can't parse this.  Can you elaborate a bit please.
>
>
> >
> > XML isn't used much on the Web today because of money. The
> > business reasons are not primarily to do with hyperlinking,
> > although I agree that's part of it.
> >
> > The problems I see are:
> >
> > 1. You can't put JavaScript-based ads in XML.
>
> XML describes content, not presentation.  If you can put links in XML,
> you should be able to put links to ads if you like.  Why you
> would put links to ads in data is a bit beyond me, but one might
> find a reason in this day and age.
>
> >
> > 2. Search engines don't know how to make useful result
> > snippets from XML, so XML files get lower ranking in search
> > results.
>
> Current search engines are focused on presenting stuff to humans to read,
> so it's no wonder they pay no attention to xml.
>
> But say there was markup language for spreadsheets, with a media
> type application/spreadsheet+xml. A crawler which is programmed
> for application/spreadsheet+xml will 'know' what to do with it,
> and a search engine will develop a presentation that will become
> familiar to people.
>
>
> >
> > But it would need to have backing of browser programmers,
> > most of whom seem to hate XML with a passion
>
> Yes, that's true.  But maybe helping XML conform to the style of the
> web would help that situation: http://annevankesteren.nl/2005/02/xml-links
>
> >, and search
> > engine vendors, who will act if there's enough content.
>
> Yes!  As you mentioned, more XML may be generated in a single day
> than exists on the Web as a whole.  Put in a little standard linking
> and who knows what could happen.
>
> >
> > My "automatic namespaces" proposal is one way to do
> > architectural forms without using explicit namespaces; i
> > think John Cowan is going to propse another at Balisage this
> > August. What I mean here is the ability to make a simple
> > mapping between elements and attributes in the document and
> > some very simple set of properties. One could use XSLT, but
> > that has unbounded computation, so something simpler to say,
> > "construct a URI by concatenating these two attributes in
> > this way with this string" and, "this element's contents form
> > the document title", and, "para elements break paragraphs".
>
> Well simplicity always has an attraction, but I have doubts that
> architectural forms are simple enough.  I believe that you
> can style XML with CSS currently, so if you could link to a
> CSS resource from an XML resource that would help the presentation
> of arbitrary XML.
>
> Cheers,
> Peter
>
>