Hi, I'm following up on the tool API discussion since I think we need to get to a point here very soon now. I'll thus try to summarize my opinion on plug-in tools once more. Plug-in tools are essentially a great thing and would be nice to have. A well-defined tools API would allow other developers to contribute to The GIMP without the need for a deep knowledge of GIMP internals and it would allow for continued development even after a stable release is out. Please note however that these advantages entirely rely on the availability of a well-defined and documented API. Unfortunately we can not offer such an API at the moment. In the past two years we managed to split the core up into several subsystems and started to develop some sort of internal GIMP API with defined relationships. This development has however been a slow migration from the 1.2 code hell towards a cleaner structure. At the end of this road stands a GIMP application which is only few code that ties together the functionality from a bunch of libraries. All tools would be modules (and might as well be plug-ins, i.e. run in a different process). This is however only the vision at the end of the road. At the moment we are only halfway there but we want to provide a new stable release to our users so that they can benefit from the development that has taken place. In my opinion we simply are not at the point where we can provide a tools API for external use. The internals simply are not yet prepared to be exposed. Releasing GIMP with the ability to have tool plug-ins as you propose it, would effectively expose the complete internals of the GIMP core. This would almost certainly introduce a maintainance nightmare. People will start to use this API thus making it impossible to do any source or even binary incompatible changes to the GIMP core. We will however need to make those changes in order to fix bugs. I expect that we will even need to do so after the stable release. The only sane way to support plug-in tools would be to put another year of development into the core. After that time, I expect the internal API to be at a point where we can make large parts of it available to plug-ins as well as plug-in tools. I don't think however that people want to wait another year and that's why I propose to remove plug-in tools for the pending stable release and postpone this feature to the next development cycle. Salut, Sven
