[thelist] What is the Best Practice for Single Page Redirects?

Jeff Howden jeff at jeffhowden.com
Thu Jul 10 16:08:30 CDT 2003


ari,

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> From: Aredridel
>
> > that depends on the timeout you have set.  if it's 0,
> > they won't even see the redirect.
>
> Except as the flash of content before the browser
> figures it out.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

the browser "figures it out" when it encounters the <meta> tag in the
source.  if it's up in the head of the document where it belongs the
redirect should be transparent to the user.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> That's why I used "5" in my examples.  It still messes
> with navigation, however.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

again, no, it does not.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> > > It should not be used with a timeout of zero,
> > > because it makes using the back button on the
> > > browser painful.
> >
> > that simply isn't true.  if there's a timeout of 0,
> > the page with the redirect is replaced in the browser
> > history with the page being redirected to.  it behaves
> > identically to sending a 302 object moved header from
> > the server.
>
> Not really.  Just in newer browsers. . . it's a common
> bug.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

i've been in the business since the days when nn2 was released.  i've yet to
find a browser that doesn't behave as i've suggested.  or, let me put it
another way.  there isn't a browser on the market today that's in the
greater than 1% market share that behaves as you suggest.  perhaps you have
a list of data to counter with?

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> > > The redirect method is better for most other things:
> > > a permanent-redirect code sent to a browser will
> > > allow, say, a browser, to automatically update
> > > bookmarks, [...]
> >
> > if there were any browsers on the market that actually
> > implemented that.
>
> It's being added, and search engines do it already.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

what browsers are adding it?  what percentage of market share do they have?

sadly, not all the major search engines update their indexes when they
encounter a redirect.

i'm not saying a server-side redirect isn't the better option.  my point is
that some of your reasons for saying it's better don't really have any
substance.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> > either way is "proper".
>
> Except that it's metadata more than content, and belongs
> with the header, not the body.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

from a purely theoretical standpoint, i'd agree with you.  however, the
<meta> exists for those users that simply don't have the server-side methods
available to them.  in their case, there's absolutely nothing wrong or
improperly about using the <meta> tag approach.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> > be warned that many browsers don't correctly interpret
> > a 301 status code.  even the search engines recommend
> > that you use a 302 if for all redirects.
>
> Yeah -- I browser-sniff and only send when I'm sure.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

yuck, sounds like a maintenance nightmare to me.  in your opinion, exactly
what percentage of the market share actually supports a 301 status code?

.jeff

——————————————————————————————————————————————————————
Jeff Howden - Web Application Specialist
Résumé - http://jeffhowden.com/about/resume/
Code Library - http://evolt.jeffhowden.com/jeff/code/




More information about the thelist mailing list