- From: Noah Mendelsohn <nrm@arcanedomain.com>
- Date: Wed, 18 May 2011 17:05:31 -0400
- To: Karl Dubost <karld@opera.com>
- CC: ashok.malhotra@oracle.com, Syd Lawrence <sydlawrence@googlemail.com>, www-tag@w3.org
On 5/18/2011 3:34 PM, Karl Dubost wrote: > Hmm. I wonder what is happening when the link containing #! is clicked from a text mail client, IRC client, etc. Well, it depends how that client handles links. If it runs a Javascript enabled user agent, then the "right thing" happens, the URI without the fragment is dereferenced, Javascript comes down and runs, and the intended content is shown to the user. If the reference is, for whatever reason, handled by a non-Javascript enabled user agent, then you tend to get a rendering of the HTML that was trying to, but could not, load the Javascript. In particular, because resources: http://example.com#!one http://example.com#!two result in the same URI being sent to the server, the same HTML tends to come back, and the user necessarily (in the non-Javascript case) gets the same display for both. Typically that's some kind of base "you lose" page, or other useless result. That's among the reasons #! is so broken and so disruptive to Web arch. (Another is that, to get around this, many sites such as twitter do the equivalent of handing out both of the following for resource one, depending on circumstance: http://example.com#one http://example.com#!one People who pass on links need to convert from one to the other depending on circumstance. (I just hit this the other day passing on a link to a tweet). A real mess. Noah
Received on Wednesday, 18 May 2011 21:05:59 UTC