Here's some notes from the 5.9 Kickoff meeting
https://notes.kde.org/p/PlasmaKickoff5.9
I'll post about the breakout topics in following e-mails
Sebas will write a blog of this up today
Plasma 5.9 Kickoff Meeting
present: Riddell, sebas, d_ed, jens, mgraesslin, bshah, llucas,
notmart, alex-l, romangg, sho, others
5.8 Evaluation
- Went well, bugfixing discipline is good
- Very well received by both users and distros
- Attracts new contributors
- cycle timing works well
Plasma's General Direction
- professional tool, reliable workhorse and elegant at getting the job done
- new LTS release in summer 2018 (proposed by sebas)
- stabilize and improve Wayland session, but don't make it default
before it's 101% good, perhaps summer 2017?
- more web services support (?), e.g. file storage cloud services
(own/nextcloud, gdrive, etc.); Freedom services first; good UX for it
- I want plasma to be a thing that doesn't get in your way of doing things
- I want plasma both a solid platform/framework and a no-nonsense
desktop system by default that users would want to chose over others
for being qualitatively better/faster/getting less in the way
- Make Plasma a unified whole, a tool to be used, used easily and
pleasantly by people who actually need it. (Scientific, Students and
Professional Settings)
- make something completely new (not just convergent design like
Discover but convergant workflows and innovations that make sense)
Biggest Pain Points
* QLocale regressions vs Klocale ( https://bugs.kde.org/show_bug.cgi?id=340982 )
* Way too little KDE developers test their apps on Wayland
* System Settings: newbies don't like it (too unwieldy, messy, little
consistency across modules)
* I think our biggest is still quality - Plasma regularly fails the
"10 minute test", I see people trying it and something doesn't work
right in the first 10 minutes
* good text input support and localization for Asian languages and scripts
* regressions between feature releases
* hybrid systems/convertables badly supported, no touch screen
gestures, no auto rotate, etc.
* multiscreen behavior: How should the system exactly behave?
* getting new developers
* still stability issues (a lot unfortunately still related to Qt/Qml
* startup performance/memory usage: we improved massively (we are
already a bit less than an empty gnome session, ohyeah, but we can
still improve
* things that are the same; look and behave differently and things
that should stick out, are hidden and behave the same. Which makes the
user distrust the DE and gets us questions about what should be fairly
easy actions
* we are lackluster in applications that tie into the DE as a whole -
there is also a big empty spot where convergence should be where we
can
Individual Plans
sebas:
* making kscreen wayland-complete,
* kscreen UI revamp ("experimental" for 5.9 perhaps installing
alongside the current one, but hidden (for testing))
* windowmetadata advancements
mgraesslin:
* multi-screen rendering on Wayland (some effects already fixed,
blur still broken)
* improve input stack (relative pointer, pointer confinement,
touchpad gestures, wacom tablet, clipboard manager aka Klipper)
* X-free kwin project (at least startup without hard X dependency
would be nice)
alex-l:
* I plan to work on Breeze and Oxygen themes+icons. Theme for Firefox is wip
Sho:
* grow the KDE community in Asia, and the biggest impediment for
that is good text input support and localization - stuff like the
keyboard KCM I mentioned above
* Work toward Wayland completeness, to free up dev resources for
things like that again
* as mentioned, really interested in somehow working with sysadmin
to get infra for performance measurements (We'd need a dedicated
benchmark runner system)
* We want to grow the widget ecosystem on the store, so I'm
working on putting Netrunner's Simple Menu there - this needs more
upstreaming of Kicker backend mods, and probably moving the backend to
p-w 5.8 I mean
* Virtual desktop API in kwayland, and making the TM+Pager applets use it
* A (lightweight) widget gallery in Dashboard, for the same reason
as why apps in Dashboard - sometimes it's nicer fullscreen
* A rewrite of the Yakuake UI, with a new theming system
bshah:
* plasma mobile stuff.. I want to stabilize it and make it perfect.
* I want to work on some kind of integration testing / regression
testing for plasma and plasma mobile stuff combined
d_ed:
* I want to get back into SDDM
* I'd quite like to do the wayland high DPI
* plus the backlog of other SDDM tasks
* rootless Xorg,
* multi seat,
* accountsservice
* and of course ... bugs
* startup time improvements
notmart:
* ~6 months: integration with QQuickControls2, dependencies in look
and feel packages, probably some tweaks in breeze qwidget style
* 12 months: beside above pojnts, more work in performance
department, try to load by default even less
* 2 years: aggressively go on last mile points, not much new
features, but quality and quality
That means specifically:
- kirigami still needs quite some work: related as we seen with
discover kirigami apps are well-received
- QtQuickControls2 themes: Plasma and Desktop
- dependencies in look and feel packages
- probably try to tweak scrollbars in breeze qstyle as well
romangg:
* ~6 months, 12 months, 2 years: something I would like to see is
a unified framework for tabs. and a KWallet replacement
"tabs"? - framework for tabs (dolphin, kate, konsole, browser
maybe, ... ):
* chrome, firefox as good examples
* recover closed tabs
* interchange them and open new window
kbroulik:
* get global menu back \o/ that means:
* add that to plasma-integration (just uploaded patch to phab),
* resurrect/fix/cleanup kappmenu kded thingie (acts as global menu
registrar, WIP),
* bring back menu button in kwin (WIP), write a global menu applet (TODO),
* and then later (i dont think I can make it for 5.9) is wayland global menu
Schedule
5.8 LTS
5.9 31 Jan 2017
5.10 May 2017
5.11 September 2017
5.12 December 2017
5.13 April 2018
5.14 August 2018, next LTS
Jonathan advised 4 month cycle releasing at the end of every January,
May and September each year
Breakouts to be planned
* system settings direction and improvements (alex-l, ...)
* UI and user vision for cloud services (kbroulik, Sho, sebas, bgupta, ...)
* regression / integration testing (possibly with openQA)
(Riddell, bshah, einar77, ...)
* multiscreen behaviour: how should Plasma exactly behave in
different scenarios?
* how to attract new contributors breakout
* store and dependencies: to both plasma and other apps versions
(for i.e. simplemenu, Dolphin service menus) and to other store items
(i.e. look&feel packages)
* what Qt version will 5.9 require? 5.6 still or 5.7?
* what can go in 5.8 LTS? only bugs or some features?
- bugfixes only, small features can sometimes be considered
bugfixes if e.g. they're missing from kde4
- ask release team for exceptions
