Re: Frames and Documents (fwd)

Sunil Mishra wrote:
> If you are going to put in indexing, I would much rather see indexing
on
> names than on numbers.

Kind of unwieldy in the case of multiple frames with the same name,
isn't it?

> This looks suspciously like arguments to a function.

Nope. It's indexing of frame arrays, as per NetScape's documentation.

> ...
> In this case, I would think you can fully specify the page if you
nested
> the indices. How do you propose to keep track of what indices
belonged to
> which frameset? Is url_in_frame4.html the fourth frame of
> url_in_frame2.html or of main.frame.html?

The idea here is that name references in a framesetting document's URL
can be indexes into frame arrays, followed by the URL. Each frameset is
referenced through its parent. Here's a better example:

A framesetting document http://main/frame.html creates three frames.
These frames are indexed relative to the window as frames[0],
frames[1], and frames[2]. At this point the URL for this page is simply
http://main/frame.html.

A link is selected in frames[0] that loads frame2.html into frames[2].
The bookmarking URL has now become

    http://main/frame.html#[2]frame2.html

since to reproduce the current state of the document, the UA simply
loads main/frame.html normally and then loads frame2.html into
frames[2].

Let's say frame2.html is also a frameset document that splits frames[2]
into two subframes, which can be considered as frames[2][0] and
frames[2][1]. A link is selected in frames[2][0] that loads
frame2-1.html into frames[2][1]. The bookmarking URL now becomes

   http://main/frame.html#[2]frame2.html#[2][1]frame2-1.html

And so on. It doesn't strike me as that difficult to keep track of
which URLs are necessary to reconstruct the current state of the
window. And certainly the index notation is far more compact than
trying to track frame names that may be duplicated.

I see the problem when the URL already has a name reference attached -
particularly if the named reference starts with '[', but this isn't a
formal proposal, just a suggestion that will not break current
browsers.

David Perrell

Received on Monday, 9 September 1996 19:09:04 UTC