W3C home > Mailing lists > Public > www-style@w3.org > October 2006

Re: :loading :loaded :missing :corrupt

From: Ted Shaneyfelt <tvs@hawaii.edu>
Date: Mon, 30 Oct 2006 12:29:46 -1000
Cc: www-style@w3.org
Message-id: <45467CDA.8050902@hawaii.edu>

David Woolley wrote:

>>Maybe this exists or was discussed but I missed it. Excuse me if so. 
>>Pseudo-class selectors I wish were available are for images relate to 
>>their status in loading, so maybe a remote image could be hidden when 
>>the network becomes isolated, etc. 
>>    
>>
>
>These are almost all only really useful for user style sheets, for which
>it is not essential to define an over the wire format.
>
>  
>
The scenario that led up to this suggestion is a web authoring program 
inserts its logo at the bottom of the page, but when that site is down, 
the logo missing makes my page ugly and it's not my fault. I could 
dynamically handle it, as I've done in the past, with visibility:hidden 
and onload event changing it to visible, but as people abuse the web to 
use it for secure transactions instead of open information distribution, 
and as that leads to fear of any executable code because some code 
compromises security of such systems, browsers often disable JavaScript 
by default or at least make annoying steps that the visitor has to go 
through in order to make such a simple effect happen.  In another 
scenario, we mirror some content behind a firewall, and it is useful to 
determine whether we're inside or outside (by the existance or not of a 
particular file with a relative URL)

So I would disagree.

>>Distinct states that would be useful to recognize include:
>>
>>:loading,
>>    
>>
>
>The style sheet itself could be loading, so one can't really rely on
>anything except the default (out of line style sheets are generally
>considered desirable).
>  
>
See above scenario.

>  
>
>>:missing,
>>:corrupt
>>    
>>
>
>  
>
See above scenario. I don't maintain the image, but I need to reference it.

>These two should not occur in a properly maintained site.
>
>I think you are also missing two cases:
>
>- blocked by policy (e.g. same origin policy);
>
>- images turned off globally.
>
>  
>
Those would be fine as well, though in practice a method to 
differentiate between something being successfully loaded and not 
successfully loaded is the primary concern. Although one could use 
proprietary pseudo-classes, I try to avoid them myself.

_-Ted
Received on Monday, 30 October 2006 22:30:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:47 GMT