Re: UA norm for redirects (both META and http)

Hi Thomas,

On Jun 1, 2008, at 12:54 PM, Thomas Broyer wrote:

>
> On Fri, May 30, 2008 at 11:24 AM, Robert J Burns wrote:
>>
>> Thanks for the input on this. I changed the wiki page[1] so that it  
>> only
>> concerns 301 (permanently moved) redirects[2]. The motivation  
>> behind the
>> proposal is to advise UAs to treat the 301s and the META redirect
>> consistently and in an interoperable manner that authors (and  
>> users) can
>> count on.
>
> Why would you want <meta> redirects be treated as "permanently moved"?
> How about pages such as:
> <!DOCTYPE html>
> <meta http-equiv=refresh
> content=5;http://web.archive.org/web/*/http://example.net>
> <h1>Down for maintenance</h1>
> <p>We're down for maintenance. In the mean time, you can browse the
> site using the <a
> href=http://web.archive.org/web/*/http://example.net>Internet Archive
> Wayback Machine</a>. You should be automatically redirected to the
> archive in about 5 seconds.

That's a good example. Yes, Boris raised the issue that non-0  
redirects need to be treated differently. My main concern is in  
handling 0 timeout redirects so that authors can actually make use of  
them in the future with concern for usability and accessibility  
problems.

Your example also raises a valid concern even in the 0 timeout case in  
that authors might do the same thing as in your example, but with a 0  
timeout. That's a problem i had not considered, though I'm not sure  
how much existing content this would break. The possibility does exist  
to interactively involve the user to make sure updating the bookmark  
is indeed what the user wants to do.


> Or are you proposing such a treatment only for "immediate" redirects?
>
>> On the issue of authoring misuse of 301s, is there some other  
>> litmus test we
>> can apply (such as consistent response headers) that could help  
>> identify
>> these misused 301s?
>
> Actually the problem is more with misused 302s. CGI "containers" (such
> as Apache) for instance issue a 302 if the program outputs a Location
> response header and no other Status has been set. PHP has mimic'd this
> behavior for consistency between the CGI and "module" versions.
>
>
> However, my opinion is that <meta> redirects should be considered  
> 302s/307s.

Even in the case of a 0 timeout? That may be right. Perhaps we could  
add similar keywords to the meta@http-equiv to let authors express  
permanently moved redirects from temporary: like refresh-temp and  
refresh-perm (or even drawing on the other distinctions http makes).

Actually looking at the latest HTML5 draft I now see that the refresh  
language has been removed. I'm not sure why that would be. It's not  
like removing from the spec will mean it's not available and used.  
After all it's not in HTML4 either. It's also a valuable tool for  
authors who do not have configuration access to their http servers, do  
not have the know-how to configure those servers, or are not using  
http at all.

Take care,
Rob

Received on Sunday, 1 June 2008 13:29:46 UTC