Different implementation for "Do not suspend timers when inactive"

Hi,

For some time, Workrave has had a configuration option, where you could
disable the "suspend timers when inactive" feature, which is turned on by
default.

If I recall correctly, this option was added to make workrave more useful
to people reading the screen: Breaks would still come, regardless of
whether workrave had detected any activity.

Although I am not using this feature myself, I do not think it is working
really well in practice. The only time it will work, is if you are reading
the screen without being interrupted by anything else. Interruptions
(phone call, coffee break, visit the restroom) all count as activity,
regardless of how long the interruption takes (In my case, fetching coffee
definitely takes longer than a restbreak). Additionally, there have been
quite a few bugs found in the implementation. The last one is bug 719,
submitted only last month.

Recently, Michael Levin added a comment to bug 479 (which requests
implementation of a "reading mode"), that made me rethink the whole
suspend-timer-when-inactive thing. What Michael proposes, in short, is
that there be a new mode. In this mode, if the user becomes inactive, and
then active again before a natural break is complete, then the user is in
retrospect considered active anyway. Michaels goal was to read from the
screen while occasionally waving the mouse (or scrolling the document).
With this new mode, he'd be considered active the whole time, and get his
breaks in time.

My feeling is that Michaels suggestion will work much better than the
current "disable suspend-timer-when-inactive" approach, because it still
allows a "natural break" to take place if the user is interrupted long
enough. This way, when you return to the computer, you'll have the maximum
amount of time available, instead of finding that you have returned just
one minute before your break is due. Additionally, I would suspect that
this feature has less impact on workrave's design than the current
"disable suspend-timer-when-inactive", and hence less bugs are to be
expected.

I am tempted to conclude that it might be a good idea to completely scrap
the "disable suspend-timer-when-inactive" thingy, in favor of Michaels
proposal. I'm not sure what to call it yet. Maybe something like "Ignore
incomplete natural breaks" or something.

Any thoughts? Anyone still disabeling "suspend-timers-when-inactive"?

Thanks!

Groetjes,

Kees-Jan


-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace