Derek made a guest appearance tonight at the weekly WiX toolset coding night. It was fun having him around again. We all caught up on the things that have happened in the last year or so. At the end of the night (about 1 AM), I dropped him off at his hotel downtown and drove back home. On my way home, I remembered a recent comment that was directed at Derek and I:
Derek and Rob, wouldn't it be better to first document all the current 3.0 features (SecureObj, Torch, Pyro, Setupbld etc) then [sic] adding support for a beta?
My simple answer is, "No." I think it is better to ensure that the WiX toolset supports as many of the available scenarios correctly as quickly possible. My primary goal is to make sure that setup developers can get their job done well and that the tools are not the obstacle blocking them.
That said, I recognize that documentation can greatly improve the speed at which developers can learn the tools. You will notice that with many of the new features going into the core toolset that there is a great deal more documentation than what has been provided in the past. We obviously have a lot of "documentation debt" that still needs filling in.
Ultimately though every user of the WiX toolset has access to the deepest level of documentation, the source code. As long as the WiX toolset supports the full set of features available from the Windows Installer then any developer can learn how to use them.
PS: No, I am not trying to argue that documentation is irrelevant. I'm just saying that with the finite number of resources that I have to put toward the project, making the impossible possible currently seems more valuable than going back and writing more documentation for that which is already possible.