Hi Alex,

Here is a demo of what I was thinking of:
<http://folk.uio.no/jkleiser/pico/experiment/timeoutdemo.html>

If you think it could be useful, you can pick the parts you like.

Another detail:
When one logs into the wiki, and then later logs out, the timeout info should be removed, and if possible there should no longer be a session at the server that can time out and give the user the blue Timeout page.

/Jon

On 4/21/10 10:54 AM, Alexander Burger wrote:
Hi Jon,

I still don't understand what it tries to tell me. If it means there's
some 10 minutes till timeout, why doesn't these minutes and seconds count
down?

These are not minutes and seconds, but the time when the timeout will
occur. It changes whenever an event happens, i.e. a reload of the page,
a click on a link, or a JavaScript action. So it always shows the
correct moment when the current session will terminate, as each event or
action resets the timeout in the global '*Run' (see documentation of
'timeout').

Hmm, I see indeed one problem: This time is the local time of the
server, Central European Time in this case. This might not be very
useful in case of the wiki which is different from a local application.
I should at least append the string "CET".


If there shall be something telling the user how much time is left before
timeout, I think it should look less alarming, maybe somewhat more "out of
the way" (not so central), and it should be updated every 5 secs. or so.
When there is only 1 or 2 minutes left, a more visible message could
appear, and at timeout a final message could tell so.

This is technically not possible, because the server wakes up only when
the user clicks something. And only then the timeout changes.

So if the display says the session will time out at half past 12, the
session will terminate then unless the user clicks somewhere.

What would be possible, though, is using an '+Auto' button, in
combination with '*Throbber' or similar. The '+Auto' button "presses
itself", to refresh a page periodically, e.g. every two seconds, and
shows a continuously updated display (there is not example code released
for that). But this is quite some overhead, and then of course the
timeout display is meaningless, because it will never timeout, and the
timeout will always be now plus 1 hour.

Cheers,
- Alex

--
UNSUBSCRIBE: mailto:picol...@software-lab.de?subject=unsubscribe

Reply via email to