There are software engineering tooling problems for which the solution benefits from being encapsulated as a plug-in. Among these problems, to ensure higher leverage, there are categories for which is important that their solution is extensible. However, extending a plug-in in practice often takes a long time, requires expertise, involves hacks and produces low quality code. In this paper, we advocate that assuring early in the design that a plug-in is extensible, by providing the necessary extension points, increases its re-usability, improves its evolution, and ultimately reduces the development time of the extender plug-in. We identify categories of software engineering problems whose solutions benefit from being extensible plug-ins, and review existing approaches to extending plug-ins. Finally, we report on our experience, with some of these approaches, in extending an Eclipse plug-in for a domain specific modeling language graphical editor.