Re: [xml-dev] the challenges of 'extensible'

From
Simon St.Laurent <>
To
"" <>
Date
2014-04-11T13:16:11Z
ID
<>
Thread
Re: [xml-dev] the challenges of 'extensible'
On 04/11/2014 08:37 AM, Rushforth, Peter wrote:
>   Isn't the DOM the backbone of interoperability in browsers, isn't it the thing that allows
> (many) frameworks to exist?

Yes.  My first (1997) book, winner of the coveted Amazon 1-star award, 
was on its emergence.  It's done everything I hoped and then some.

> If Web Components works, won't it be because of the DOM?  The DOM being what
> can be produced by virtue of the browser understanding the document type,

Yes and no.

Web Components is built on the DOM, though in some ways it's fragmenting 
the DOM.  We now have Shadow DOM, light DOM, and a lot of frameworks 
(most notably React.js) that are creating their own lighter-weight 
object models to avoid the overhead of the DOM.  Famo.us, as noted at 
the end of my piece, is basically trying to dodge the DOM to the extent 
possible while building its apps.

(Also, at this point I think most search engines are still treating HTML 
as annotated text, though I hear rumors of DOM-building.)

The questions are getting harder because the DOM is carrying pretty much 
everything.  The HTML document structure, the CSS annotations, 
JavaScript event handlers, and hooks to many other invisible parts 
inside of the browser.  Exposing those invisible parts likely means 
extending the DOM.

Manipulating the DOM and further extending what it can carry is getting 
more challenging as that load increases.  Spend some time in the 
debugger of a browser working with a complex web app, and you'll get a 
sense of it.

The DOM has other fundamental problems.  Some of those were visible 
early, and are among the reasons XML folks quickly developed 
alternatives.  Others have only emerged as things like multiprocessing 
and ever-growing expectations for animation rendering speed push the 
boundaries of classic models for doing things.

> per
> REST / media types.  In fact, isn't the DOM the original "REST API"?

There's nothing obligating a browser (or other tool) to use a DOM to 
process a media type, and I don't think it ties directly to REST.  I've 
certainly seen RESTful projects using SAX and similar APIs, XSLT, or XQuery.

Thanks,
Simon