- From: Michael Baynton <mike@wazee.org>
- Date: Tue, 22 Nov 2011 23:01:01 -0600
- To: <public-web-perf@w3.org>
Hello, I would like to propose an interface enabling site developers to detect when there is a state of overall user idleness on a system (defined as no activity on a system's keyboard/mouse/other direct human interface devices) for a period of time defined by the web developer. I am posting it here because I think the Page Visibility specification's goals are most closely aligned with this proposal. I see there was a related discussion on this list (http://lists.w3.org/Archives/Public/public-web-perf/2011May/0136.html seems to be its conclusion; it didn't get very far) but its scope was limited to detection of inputs directed at the page in question - something that can be done pretty accurately already, and the old proposal seemed to suggest replacing the visibility interface entirely with an interface for reporting idleness. The main difference I am proposing is for the determination of idleness to be based on an absence of any user inputs to the system rather than an absence of inputs to the particular webpage, and for it to serve as a supplement to the page visibility interface rather than a replacement. I do not believe that this would breach the privacy or security restrictions one would reasonably expect a web browser to enforce on the pages running within it - my idleness interface would not allow you to monitor anything specific about a user's activities outside your application, but would merely provide a single notification to your application when no user inputs had occurred for some length of time, a single notification when inputs resumed, and probably a mechanism for the page to poll for a Boolean current state of idleness as well. When expanded to consider all user inputs to the system, I think my proposal becomes different enough from the previous discussion to warrant some further consideration. I think that idleness alone is an incomplete measure of whether computational or battery-intensive tasks should be ramped down, but I think visibility alone is incomplete as well. Besides the possibilities for allowing developers to make more intelligent performance decisions, this proposal would also enable web applications to incorporate a more accurate estimation of whether a user was present into business logic than is currently possible. Consider for example the growing trend towards real-time interactivity through long-running XMLHTTPRequests or webSockets. In these applications, humans often interact directly with each other through the web application in real-time. I think there will be a growing desire for such applications to be able to report whether a given user is likely to be responsive. Thank you for your consideration, I look forward to hearing others' thoughts. Mike Baynton
Received on Wednesday, 23 November 2011 09:14:10 UTC