The debate about Debian init system replacement continues and many developers have given their opinions.

However, what do the users want? What type of features will make it easier for users to adopt and promote Debian and free software in general? Could this even help bring improvements to other distributions too?

Some of the possibilities I've previously raised for discussion (and they are not specific to Debian by any means):

  • Better integration with tools for service monitoring (for example, so Nagios can automatically detect all the services on a system and the dependency relationship between them, avoiding any manual configuration of the monitoring system)
  • Better integration with tools for service management (for example, so Nagios can more easily restart a service without having to provide some action script developed by hand)
  • Better support for dependencies in clustering and distributed environments (e.g. Apache on host A should only start after MySQL is running on host B or host C)
  • Support for multiple instances of a service, each with separate configuration files on the same host

Some of the systems up for discussion go some of the way to addressing these types of requirements - none of them go all the way.

What other requirements do people have? Should Debian wait a little longer and let the different solutions compete to offer more of these things before any decision is made?

My own feeling is that Debian should not commit to something that will involve any loss in current features or support for ports and we should use this opportunity to find the best possible set of features for the future.