Hi everybody! Many great people provided great comments and ideas to my thought and the picture gets bigger. Another twist is added by the fact that it's holiday time here in Italy, and I have eaten and drunk things that no English words to my knowledge can describe :)
You made me discover that my thoughts were limited in may ways; among others I considered ifupdown the main center for profile detection and reconfiguration, and I supposed that all the possible system configurations could be represented by many distinct profiles. The problem gets bigger and involves many (still independent) parts of Debian: apmd, acpid, network tools, pppd, pcmcia-cs, hotplug are some examples of parts of Debian that cause or detect changes on the system and provide hooks to react to them, and more can be expected to be added in the future. Reacting to IrDA discoveries comes to mind, as well as deploying a tool to enable some services based on user requests (take a look to http://reshare.sf.net for an example of a tool to manage such requests. Ready to be added to Debian if needed, btw.). Right now there are utilities to scratch many of the itches related to this, but Debian seems not to have a masterplan to define its behaviour for a system that should react to a changing configuration. Maybe I should retarget my thoughts and start addressing this whole issue? In fact, right now we already have some problems: pppd scripts might need to be filled with `if's checking if other network connectivity is present, and start some network services only depending on if the system is running on AC power or if a given PCMCIA hard drive is present. A DNS or a network route might need to be reconfigured by pppd, pcmcia-cs and network-detection tools. At the current state of things this things, if pushed, can easily lead to a mess. Maybe it's time to design a tool that gathers all notification of the changes that happen in the system and provides the user with a single point to program the logic to react to them, and standardize it in such a way that other parts of Debian can interact with it? With it we could have tools specialized for detection, tools specialized for reconfiguration and a central point that can do the coordination to make them work together. I can easily see that the issue will become more important in the future, as outplugging devices will become more popular and Linux systems will be deployed in environments that require more flexibility. Am I still under the effects of my wonder-lunch (anyone here is invided should you pass in the nearby) or is there some logic in all this? If there is, I'm ready to get to work. Bye, Enrico P.S. In the meantime, we might still like to think about merging efforts to share functionalities with all the network detection tools, and it would still be an interesting work. Unluckily I won't have access to whereami until tuesday so I can't make studies right now, besides proposing to extract the link-beat detection code from laptop-net in a daemon that launches a script when the link goes up or down on a given interface and let users hook their preferred tool to it. -- GPG key: 1024D/797EBFAB 2000-12-05 Enrico Zini <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
