On Wed, 24 Sep 2003, Eugene Lee wrote: > : and that they need to instruct > : the system (or "entire computer", or "root", or however they perceive it) > : to perform such an act if it's really what they want to have happen. I > : can't really imagine anyone who understands that concept arguing against > : it. > > So we have two possible courses of action: > > Choice #1: Mail.app tells Internet Connection.app to establish > a dialup connection. Then Mail.app retrieves email. Finally > Mail.app tells Internet Connect.app to close the dialup > connection. > > Choice #2: An AppleScript tells Internet Connection.app to > establish a dialup connection. Then the AppleScript tells > Mail.app to retrieve email. Finally the AppleScript tells > Internet Connect.app to close the dialup connection. > > Some users want Choice #1. Other users (in the name of properness, > appropriateness, correctness, etc.) want Choice #2. But both choices > accomplish the same task, yet both raises the same problems of closing > the connection while other apps might need it. So what's the point of > arguing that one choice is better than the other, when they both do the > same thing? First, I'm not arguing. I don't care where you put the @#$#% button to run the Applescript. Put it on the Mail button bar if you like. My point is that it needs to not be a default/builtin action, that there needs to be an overt acknowledgement by the user that they are operating outside the confines of the Mail application, and affecting the rest of the system. The difference that I'm trying to articulate is between the user who is aware that their actions are global in nature, and accept those consequences, and the user who is oblivious to the fact that their actions are global in nature and may be affecting any number of other users, programs, or processes. This is equivalent to the authorization step of entering an admin password to install an application. > It just doesn't seem so obvious that, in this case, it is utterly > importantly to recognize the difference between the OS and the app. I don't and didn't insist on that; it *is* important to recognize the different roles and responsibilities implicit in the request. A single user of a single application doesn't, and shouldn't, have the ability to affect the entire system. That's the answer to why the application doesn't have a button built into it. If you want to say "I accept responsibility for everything else that may be going on" (which is quite reasonable if you know that nothing else IS going on), then put on your admin hat and say "Make it so.". You can identify that distinction by prompting for an admin login each time, which is cumbersome in this case, or by assigning the responsibility to the system, which is the way it's built into OS X as discussed, or by taking the overt step of writing a script or adding a button, or customizing Mail in some other way that gives it abilities and powers far beyond those of mortal applications. KeS