Ideas on modernizing the open-source desktop

Commenters debated how to modernize the open-source desktop, with strong disagreement about change versus stability. Some, like bregma, insisted on keeping classic control and the CLI intact, and Telaneo argued that disruptive UI changes that break workflows are user-hostile and should be opt-in until proven. pythonRon highlighted concrete technical limits (Wayland not preserving window size/position) that he sees as blocking user control, while WCSTombs warned against introducing AI-driven unpredictability into basic file interaction. Others pointed out that innovation has happened: benrutter praised the wide variety of experiments (tiling WMs, custom launchers, Elementary OS), noelwelsh prefers automatic/tiling window management, and utopiah noted practical wins like KDE Connect and VR work.

Several commenters offered specific modernization ideas and recalled past attempts that fizzled. probably_wrong proposed a Desktop Manager to treat the desktop as switchable project real estate; tancop advocated replacing the start menu with a Raycast-style semantic launcher and richer file metadata or relational-storage to improve search; maufl and ezst recalled ambitious but unfinished efforts (Ubuntu One sync, plasmoids, compiz) as cautionary tales. Opinion divides mainly on whether to pursue radical new paradigms that must show substantial gains over WIMP, or to focus on incremental, stable improvements that respect existing workflows and control.