Proposal to plan for "Milestone Releases" on the way to KDE4

[Kurt Pfeifle]

Earlier this evening I sent in my proposal to plan already now for
several "milestone releases" on the way to KDE 4.0, as well as one
polished "Polished KDE Application Release 2006" that concentrates
on featuring the ongoing development and polishing of our best KDE3 
applications. [This mail is not I've not yet received my own mail
from the list -- I've been on travel, and wrote this mail from the
backseat of a car :)  ]

For the KDE4 milestons, the idea includes of course to *announce and 
publish* such a roadmap, and include rough descriptions of stuff they
will contain.

The "Applications 2006 Release" of course is only possible, if it 
gets support by the respective app developers, and if they find it
useful for themselves. Otherwise they'll probably just continue 
releasing each for themselves at the times their creators think is 
best...

Here are more arguments in favor, and thoughts about the two proposals:

KDE4 Milestones ("3.9.0"):
---------------------------
- would allow a better defined and more structured way of working
  for most of us, including translators, documentation writers and
  beta testers;
- the upcoming TWG should take in the driver seat to lead the 
  discussion about a sane and practical intial roadmap for milestones: 
  How many? In which intervals? With which specific (or only rough) 
  implementation goals? etc.
- Milestones should define when each of the envisiged KDE4 core 
  technologies are expected to be usable and "stable" to what degree: 
  dcop/dbus?, kio slaves?, kparts/containers?, bindings/kjs?, 
  kmm/phonon?, plasma?, solid?, kdeprinting/pdf?, ksvg2/kdom2?, 
  tenor?, ktts?, accessibility? ...
- Each milestone should be *functional* with its layed down targets;
  (this doesnt mean one couldnt completely drop one implementation
  for a different one in the next milestone, based on feedback and
  field experience);
- A milestone would be an own "usable" mini-release not only for
  developers, but even for enthusiast and power users of KDE so they
  can play with it, get accustomed to it and provide feedback. There 
  should be VMware-Player images available for different distros (of 
  course, other virtualisation technologies like Xen would play their 
  role too). We should aim to make KDE4 milestone previews widely 
  and easily available to our own base, as well as other developers 
  and ISVs);
Milestones will certainly help the "fan community" to stay aboard; 
and from that core ground the news about what's cooking in the KDE
oven would easily be spreading to the wider community. 

At the same time they would somehow take advantage of the "release 
early, release often" lemma  :-)

The Grand KDE Application Release 2006 ("3.6.0"?):
--------------------------------------------------
- It would allow all those extragear and playground developers who
  find it too early to start porting to KDE 4 to continue their
  work with apps full speed based on 3.5 kdebase and kdelibs.
- It would give a good leverage for marketing and promoting some of 
  the extremley good "single" KDE applications (who at the current 
  time mostly do their own releases, their own PR and their own 
  scheduling "in isolation" from each other, rather than in a more 
  coordinated way, and without good support from the MWG (yes, I 
  am aware that it also has advantages for app developers if they
  can go ahead in their own speed, and releases independently from
  other subprojects).
- It would also be a good testing ground for the proposal Aaron made
  sometime back (which aimed at differentiating between KDE "the core
  technology platform" and KDE "the applications build on top of it"
  by even giving them different names/branding etc.)

Which apps come to mind? 
------------------------
  * First --- quanta/kommander, kmail/kontact, kdeedu, kate...; 
  * then -- amarok, digikam, gwenview, kphotoalbum/kimdaba, koffice, 
    krita, k3b, kmplayer, kaffeine, ktorrent, kile, datakiosk, kst, 
    guidance, kdissert, kalzium, kstars, kig, ktouch, kalva, kopete, 
    guidance -- man!, *do* we have a lot of kick-ass applications! 
    And they are often underrated and/or unknown to many users. They 
    would benefit lots from participating in a joint release effort
    happening once (or twice?) over the next 12-18 months and being
    guided by the MWG (who would gain lots of "practicing" for
    the 4.0 releases and its milestones too :)  [Yes, and I am 
    aware that Krita is part of and deeply embedded into KOffice;
    but need it stay so? Is a Photoshop-like app part of an Office
    Suite?)
  * and last -- there are lots of goodies developed completely 
    outside of KDE SVN: kat, smb4k, luma, tellico, basket, mateedit, 
    ktechlab, kmymoney, kbarcode, taskjuggler, kalgebra, kchmviewer, 
    noteedit, rosegarden, eric3, kdiff3, twinkle, wengophone, 
    texmaker, apollon, labplot, klamav,...
  * I've surely forgotten some -- please forgive me and add yourself
    to the list....
 
The questions are up for all of you: 
------------------------------------
  * How do you plan to handle the time until KDE4 has become
    a platform to base your projects on? 
  * Wait and see? 
  * Do you intend to switch your activities to directly contribute 
    to KDE4/kdelibs/core technologies development, and leave your
    app resting for the time being? (Sure, this would be *great*
    because the current group busy already in KDE4 porting feel
    they can use many more heads and brains. But I've also heard
    some of you just do not like it, or don't feel competent in
    that field...)
  * Do you expect to be porting your own app to KDE4/Qt4 already 
    within the next 4 weeks? 
  * Or do you rather plan for continuing several more months 
    development on the KDE3 platform? If so, do you feel you could
    benefit bing part of a joint "KDE 2006 Application Release" 
    marketing campaign?
Please give your feedback either way -- if you don't, this 
discussion could have an outcome less desired by you.

Some more thoughts:
- to me it looks like it could a worth while effort to guide a KDE
  Application Release marketing campaign if even only a sixth or a
  tenth part of the above projects want to actively participate --
  provided they do really create new features for their apps until
  summer...
- of course, the plan for a coordinated (one time only?) "2006 
  Application Release" that focuses on promoting the most polished
  KDE apps doesnt mean the apps could only release *then*, and not 
  more often -- it is just meant as an offer to you... (so if the 
  amarok squad plans to release their 1.5 in July and 1.6 in Oktober, 
  while the "common" release is planned for September -- well, then 
  the common release can ship 1.5 as well);
- and last, I *bet* that all of the KDE-based commercial distros 
  will also *love* to see something like a "KDE 3.6" (or a "Summer 
  2006 Release of KDE Apps") that they could print on their next 
  boxed products, rather than having to wait until mid-2007 before 
  they can again sell a new KDE...  :)


What about aiming at 6-7 months from now (July/August) for a "KDE 
3.6 Applications Release" (TODO: find a sexy code name for it)? 

And then a "First KDE 4.0 Feature Milestone" shortly after that, 
say in 9 months (September/October?) from now, with 3-5 more 
milestones to come every 2 months until 4.0 is ready to ship?
(A "Zero Milestone" can possibly be ready, and also released, much 
more early: it would represent only the fully ported KDE3 libs and 
base, plus the new build system, but most like none of the envisaged
new features/technologies).

Of course the TWG, once established, may come up with a completely 
different and better timescale -- the thing is: we must have one 
*at* *all* ! 

Cheers,
Kurt