site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com --Amanda _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... On Jan 30, 2008, at 11:16 AM, Uli Kusterer wrote: On 31.12.2007, at 19:53, Michael Smith wrote: If the user does not want the system to sleep, the user *should*not*tell*the*system*to*sleep*. Users make mistakes. One of Jef Raskin's basic ideas in the design of the Macintosh was "forgiveness", to let the user undo dangerous actions, or to warn them before they do them. Neither is possible if the user is about to lose data because an application is interrupted in communication with the network or another external entity. Why is sleep an interruption? One of the things that really annoys me is putting my laptop to sleep momentarily (or via energy saver), waking back up, and all of my network connections have been closed. Nothing requires this. TCP is perfectly happy to work across arbitrarily long pauses. As a data point, my Sony running FreeBSD does not have this problem. I can sleep it overnight, wake it back up, and all of my ssh sessions, etc. are still up. I've worked on a product which was essentially a background program that was doing live device communication. Often, users chose "Sleep" from the menu bar and accidentally interrupted the communication, causing the device to suddenly stop and files to be truncated prematurely. Why would this truncate data, and not just pause and pick right back up later? Any number of things can cause pauses or delays, after all, not just sleep. This email sent to site_archiver@lists.apple.com