Re: Remarks on Checkpoints

> If only one of NOFRAMES or TITLE is strongly required, then I suggest
> NOFRAMES. In a non-trivial example, it may be desirable to provide more
> significant information in NOFRAMES than simply a list of frames.

I would only suggest NOFRAMES in the case where a simple list of
frames (as provided by the titles) is not sufficient (e.g. a
non-trivial example where the frameset is telling things like "look to
the right frame here, what do you see" kind of junk).
> My standard 'how to use NOFRAMES' for an index and content frame (with
> optional pointless banner) goes like this:
> In each content page, include enough navigation links that it can stand
> on its own if the rest of the Frameset is missing.

I disagree, this is calling for alternate web site, which runs
contrary to my motto "author-once-render-multiple". Also, they are
already many documents organized along the principle of a table of
content page which one has to go back to to orient herself in the
document, so this is not unusual.

> This provides for a mechanism which allows voice systems to cope with
> navigation without having to explain a frame based site, or requiring the
> 'frame-navigation two-step' (back, select new index item) which Lynx uses
> to cope with a poor system.

As I said, this two-step navigation is not usual, and is just a
natural "gracecul degradation" of a visually oriented system, much
like table linearization or getting ALT for image.

> It is access, but it is second rate. Where
> there is a choice, we should be pushing for the first class
> solution. 

Not if it calls for duplicating the web site, one for visual user, one
for non-visual user, one for non-hearing user, etc.

Received on Friday, 8 January 1999 04:40:54 UTC