RE: Frames

Server-side scripting to support bookmarked framesets would seem to be
beneficial absent other alternatives. However, it does load the entire
frameset which nullifies some of the benefits already discussed. It also
implies a new server component/CGI script to build these "custom-populated"
framesets. 

What about a client-side URL syntax that would define the correct population
of the frameset? The browser knows the contents of each frame at the time
the bookmark request is made. It could create a URL that defines an
"overlay" set of pages that should be used to (re-)populate the frameset
when delivered from the server. I'm thinking of something like:

http://foo.com/frameset.html##title_page.html#main_page.html

The browser, when requested to load this URL, would strip off everything
following '##', request http://foo.com/frameset.html and then use a
(pre-defined) ordered traversal of the frameset to decide which frame gets
title_page.html and which page gets main_page.html. 

Does this seem like a reasonable approach?

-J



> -----Original Message-----
> From: David Perrell [mailto:davidp@earthlink.net]
> Sent: Wednesday, May 03, 2000 11:11 AM
> To: www-talk@w3c.org
> Subject: Re: Frames
> 
> 
> From: "Bob Stayton" <bobs@sco.COM>:
> 
> > The biggest complaint that I get about frames
> > is that users cannot bookmark what they see in
> > their frames after they have navigated from
> > the starting frame content.  I never understood
> > why not, though. Certainly the browser has
> > knowledge of the URLs in each frame.  I suspect
> > the bookmark data format didn't support anything
> > other than single URLs, and since multiple
> > frame content cannot be expressed as a single
> > URL ...
> 
> It is certainly possible to create bookmarkable framesets (see
> <http://hpaa/test/topdoc.asp>), but this loses the 
> persistence of javascript
> variables and functions declared in the top frame.
> 
> It is also possible to bookmark a framed document in NS 
> Navigator and MSIE
> by right-clicking in the frame, but not many users seem to be 
> aware of this.
> 
> David Perrell
> 
>

Received on Thursday, 4 May 2000 13:29:32 UTC