Proposal: 2.4 rewording

I took an action item to reword 2.4 to deal with the limitations of the
current wording with respect to how much time to allocate, when it is
appropriate to allow timeouts to occur, etc.

We were finding trouble setting a number to determine the limitation in
seconds of an interaction event. I think it's tilting at windmills to try to
arrive at a fixed figure for interaction given the domain in which we're
working, so I tried to produce something that overcomes this problem.

Here's what I came up with:
2.4 (original) Give users control over how long they can spend reading or
interacting with content.
2.4. Give users _as much time as possible_ to read or interact with content.

Techniques can include (and these are just conceptual samples, not
comprehensive proposed wording):
- Avoid setting any timeout or interval that doesn't technically need to be
- Where timeouts are technically required and cannot be overridden, allow as
much time as is technically allowable.
- Avoid fixed-speed scrolling. Offer control over objects that scroll or
flash messages in sequence (title credits, end credit scrolls, etc.).

"As much as possible" in this context is not necessarily as lax as it would
appear. For a vast majority of the content out there, "as much as possible"
is forever. There is no need for most web pages to refresh, reload, scroll,
flash, or time out. And if a site is made to prove that it adhered to the
guidelines, it is provable via code review or spec documents that the site
did or did not comply.

On a tangent: for the "meta refresh" problem, I think that a technique needs
to be placed in 2.4 to allow meta refreshes in HTML for site redirection
("this site has moved to..."). It is unfortunately necessary to do these
redirects on many sites, as the designers do not control the operation of
the server.

I think the refresh time should be 0, and the page content should say only
that the site has been moved and the updated URL. Ideally, there would be no
need for that content, but if the redirect fails, users will be stuck on a
blank page. As long as the only content is that pointer, and the page
provides no content on its own, I think this needs to be an allowable hack.


Received on Sunday, 10 June 2001 15:08:29 UTC