dwt made it sound like the contracts funded by Haiku inc on package management where somehow wasted effort. For the record, during these contracts, Oliver Tappe and Ingo Weinhold worked on setting up the package manager (pkgman) and everything needed to go along with it: defining the hpkg file format, setting up repositories to update Haiku, and reworking haikuports to generate packages using the new format (which led to rewriting most of the existing recipes).
The initial goal for all of this work was to allow updating Haiku from a version to another. The idea being that we can then release an imperfect beta1, but then keep fixing bugs in it and letting users update to the most recent version easily. Without this, the Haiku team would never dare to release a beta with bugs (they have very high quality standards).
However, the initial thoughts on just a system updater evolved onto a much larger project, solving the larger problem of porting software to Haiku and then keeping it up to date. As a result of this, instead of people porting and compiling things manually and in ways that are hard to reproduce, we now have a repository of "recipes" at haikuports, which are essentially little scripts with the instructions on how to build each ported piece of software.
Thanks to these recipes, we can accumulate knowledge on how to build things, and automate large parts of the process. With these tools, we were able to port many libraries and tools and easily keep them up to date. This includes all Qt dependencies, then Qt itself, then the KDE frameworks, then the tools that needs the framework.
There are still some efforts needed on fully automating everything, but we are already seeing the result of this work in the simple fact (but great achievement) that it is possible to get KDE frameworks and big aps such as Krita or Calligra to run smoothly on Haiku.
This is not to dismiss the work that 3deyes has done on getting this to run, but a reminder that previous efforts by other people were also needed for getting there.