Re: 4.13: URI decomposition - non-standard terminology

On Fri, 11 Jul 2008, Frank Ellermann wrote:
> Ian Hickson wrote:
> >> For example, it uses the term "protocol"; RFC 3986 uses the term 
> >> "scheme."
> >>  
> >> The HTML 5 specification should use the terminology defined by the 
> >> current standard for URIs.
> >
> > This section is just specifying (for the first time) an API that
> > has been implemented in browsers for a decade and a half at least.
> 
> An old browser happily treated http host:80 as different from host,
> this certainly needs a clarification (= old browser got this wrong).
> 
> But there is no <hostport> in STD 66 outside of appendix D.2 about
> obsolete terminology.

Please follow the link given in that section to find the definition of 
<hostport>. I'm certainly open to other ways to specify this, but we don't 
really have any choice as to what the API's semantics are.


> > only the attribute names use these old terms
> 
> <hostport> is no attribute name, it's simple to avoid it.  If you are 
> hunting obscure syntax details in STD 66 tackle the question of port = 
> *DIGIT, WTH is an empty port introduced by a colon ?

I don't understand this paragraph.


> More interesting, what's an empty fragment introduced by "#" ? Various 
> browsers interpret this as "top of file" for text/html.

I don't really understand what you're asking here. Does section 5.9.8 
Navigating to a fragment identifier answer what you were asking?

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

Received on Friday, 11 July 2008 22:07:42 UTC