On Mon, Oct 22, 2012 at 2:46 PM, Ian Hickson <ian@xxxxxxxx> wrote:
> On Mon, 22 Oct 2012, Roy T. Fielding wrote:
>>
>> What you are insisting on defining as a "URL" is the input to the
>> process of making a hypertext reference (the arbitrary string typed into
>> a dialog or placed inside an href/src attribute)
>
> Or placed on a command line to wget(1), or put in an RDFa triple store, or
> in transmitted in an HTTP Location: header, or...
It seems reasonable that someone should write rules for dealing with the kinds of errors that are observed to occur in links as embedded in resource representations AKA HTML pages. It also seems reasonable that WHATWG, who speak (if I understand correctly) for browser builders, can write those rules.
The notion that curl, or an HTTP cache manager, or an XML namespace processor, is going to be routing around errors, strikes me on the face of it as being wrong. One of the main uses I put curl to is making sure I have the URL exactly right before I drop it into chat or whatever. -T