With Canonical's announcement that they are abandoning the Unity desktop environment and reverting back to a GNOME desktop, a second missed opportunity is unfolding. KDE's Plasma desktop is an obvious replacement for Unity, as much of the functionality present in Unity is already built into the latest Plasma version by default. For a concrete demonstration of the parallels between the two desktops, watch the following pair of videos produced by Michael Tunnell of TuxDigital:

• Why Ubuntu 18.04 Should Use KDE Plasma Instead of GNOME
• Ubuntu 18.04 Should Use KDE Plasma - Part 2
• Reddit discussion thread on the video above (part 1)

Plasma gives Linux distribution packagers and maintainers finer-grained control over the GUI experience they can present to users out of the box, and more reliable expectations from one development iteration to the next (whereas GNOME developers historically are more likely to break compatibility with desktop extensions).  

In addtion to Plasma making more sense for Canonical in its vanilla state, any further customizations Canonical may make to Plasma to make it more Unity-like can then be available for other users of Plasma on non-Ubuntu distributions -- users who always loved Unity 7 but for whatever reason switched to a different Linux distribution. Whether Canonical's changes are accepted into upstream Plasma, or exist as extensions, really does not matter as long as they are packaged in a redistributable manner like all other modern Linux desktop environments (and unlike the experience with Unity). An example of polish which Canonical could add would be to:

• fine-tune the existing optional global menu that is available now with Plasma
• integrate the functionality of the third-party Active Window Control plasmoid with this, so that the global menu and maximized window title do not eat up separate parcels of real estate on a top panel (by implementing the state change on pointer hover which Unity used)
• add the locally integrated menus (LIM) option which Unity eventually acquired, to work in concert with the global menu, transcending the limiting menu bar paradigms of both Windows and OS X for the ultimate flexibility on any screen size
• make all of this compatible with commonly used Linux applications such as Firefox, Thunderbird and LibreOffice (even if it means creating custom patches for these applications, as Canonical had to do with Unity)

Canonical's revolutionary implementation of a global menu in Unity was and still is the most useful and space-saving implementation yet of such in the Linux ecosystem. That innovative work should not die with the abandonment of Unity.

Canonical has already accumulated some experience working with Qt, and anecdotal evidence seems to suggest that, once exposed to it, developers prefer working with Qt rather than GTK (note that the LXDE desktop is switching to Qt for example). It would be reasonable for Canonical to continue building on their experience with Qt. Given the weight Canonical has in the Linux world, this could also have the effect of marginalizing GTK, and increasing the appeal for developers outside of the Linux ecosystem to start supporting more mainstream applications on Linux.

The KDE community itself also seems like a community more open to symbiotic collaboration than the GNOME community, both today and in the past; well-documented conflicts with the GNOME community are, after all, what led Canonical to drop GNOME and create its own desktop environment (Unity) in the first place, while in contrast years ago the KDE team did reach out to Canonical (per a brief private exchange that I had with Aaron Siego on Google+).

Some testing benchmarks have shown Plasma to be more performant than GNOME Shell. Plasma performance could only improve if similar resources were applied to Plasma development as to GNOME development. Moreover, while for the time being Canonical seems to be putting desktop Ubuntu on the back burner in favor of focusing more on their server and IoT offerings, if they do truly ever wish to be an alternative to Red Hat in the enterprise desktop space as well, it would stand to reason that they should choose a desktop environment different from what Red Hat offers -- but without fragmenting their resources by developing a completely new desktop in-house (i.e., Unity).

Historically much FUD has already been spread about how complicated KDE is, but Canonical has always demonstrated an ability to polish what end users see out of the box (this has always been one of Canconical's main strengths), and could lay these distortions to rest, quietly showing how Plasma is the superior desktop environment for novices and advanced users alike. Plasma and Canonical would both win with more support from Canonical applied to Plasma development, and Plasma would truly differentiate the desktop version of Ubuntu from its Red Hat counterpart, becoming a competitive advantage. Beyond the enterprise space, a popular, stable Linux distribution like Ubuntu, coupled with a powerful, customizable desktop GUI like Plasma (presented in a newcomer-friendly package) enhances the competitiveness of desktop Linux in general with Windows and OS X.