[whatwg] Client-side includes proposal

[Ian Hickson]

On Tue, 19 Aug 2008, Elliotte Harold wrote:
> >
> > In the case of non-Web content, the use of HTML is an academic point, 
> > since any format would work as well.
> 
> Really? Why? and how? That's certainly not self-evident.

When you control the software used to read the data, it doesn't matter 
what the data format is.


> Aside from embedded links, which can point into the file system and are 
> usually relative anyway, there's very little web-specific about HTML. 
> It's just one format that can be served over HTTP or read from a disk, 
> just like PDF or text/plain or OpenDocument.

Exactly.


> HTML has some nice characteristics like resolution independence, direct 
> editability as text, and automatic reflow; but these are in no way 
> limited to network transfers.

It has charactersitics that are independent of its use on the Web, yes.


> For many use cases, especially cross-platform ones, HTML is the 
> formatted text format of choice.

HTML isn't a formatted text format...


> A properly designed HTML spec should not require, prohibit, or 
> preference a document being read from the network or from a local file 
> system or via any other protocol. HTML 1 through 4 and XHTML 1 and 2 had 
> this important characteristic. I hope HTML 5 does as well.

I imagine it will. But it's not a requirement. If something comes up that 
would make HTML better for the Web and as a side-effect makes it not work 
for non-Web cases, then we should do it, not because of the side-effect, 
but because it makes HTML better for the Web. That HTML5 works for non-Web 
cases can be a happy accident, but it's not an intentional characteristic.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'