[Kde-hci] Suggested release schedule for KDE 4.0

Olaf Schmidt ojschmidt at kde.org
Thu Mar 8 14:51:47 CET 2007


Hi!

Thank you for contacting the HCI workgroup for feedback about the planned 
release schedule. It is a very good sign that the release team is planning to 
give the KDE4 development work more structure, so I welcome the initiative to 
create a clear release plan.

Some adjustments seem necessary to properly deal with the needs of the various 
HCI teams. The following comments are my summary of discussions I had with 
several members of the HCI workgroup (mainly accessibility and usability) 
during the last hours.

> On the release-team mailing list we are currently planning a release
> schedule for KDE 4.0. Basically the current planning means that the
> kdelibs API is frozen in about a month,

How will this freeze be handled?

At aKademy, the HCI workgroup pushed for a close cooperation and coordination 
between all workgroups (marketing, HCI, TWG, sysadmin at the time) when 
planning the KDE 4 release. We had a common meeting of all workgroups, where 
we agreed to have a list of features and API changes that are 
considered "showstoppers" from the perspective  of the 
usability/artwork/localisation/accessibility teams. The TWG was then replaced 
by the release team, and we assumed that the release team would contact us 
and inform us about the location of this list, so that we can add content 
early enough for the release. I am sorry for not contacting the release team 
myself about this point. I had no idea that you were planning to announce the 
API freeze with only one month of warning.

The plan to freeze the API is OK IMHO if features listed in the "showstopper" 
list can be implemented later. We can try to document all areas with necessary 
API changes within one month, but we will need additional time to pursuade 
developers to implement the needed changes. This additional time can be 
bigger or shorter, depending on how much the release team is willing to 
support us in these finding developers.

Some important topics that need to be part of this list are:

* Reworked colour theme system (I posted a document draft several times on 
kde-core-devel; we can try to complete the colour theme document within one 
week)

* Qt 4.3 contains some layout system changes that are needed for usability 
improvements.

* Qt Accessibility Framework: In middle-term perspective, all widgets in 
kde*libs and kdebase should provide the Qt Accessibility Framework with the 
necessary information to allow their use by users with disabilities (screen 
readers etc). The Qt Accessibility Framework changes in Qt 4.3 are done by 
Harald Fernengel and other Trolltech developers, so we depend on the Qt 4.3 
schedule. Harald, would it be possible for someone at Trolltech to complete a 
kde*libs/kdebase widget API review for accessibility needs within one month?

* Consistent and complete keyboard navigation of all of KDE (we still need to 
evaluate whether this has any influence on the API)

* There might be a number of other (smaller) areas where API changes could be 
necessary for usability. The usability team does not consist of developers, 
so it will take some time and cooperation to complete this review.

> a feature freeze in early May,

Does the feature freeze include changes as a result of usability reviews?
Or would it be possible to explicitly allocate time in the schedule to fix 
usability and accessibility problems (rather than "hard" bugs only)?

> and a release around October

A quick stable release for kdelibs and kdebase/runtime seems to make sense, 
but an end-user release in October seems to be unrealistic. The user 
interface of important parts of KDE4 (e.g. Plasma, Nepomuk) does not exist 
yet and the usability team needs time to conduct a proper usability review of 
the new interfaces.

KDE 4.0 will be reviewed as a Vista competitor, and it would make a disastrous 
impression if core elements of the user interface are either not present yet 
or have an extremely bad usability because of a rushed schedule. It would 
also make a disastrous impression if important desktop applications (e.g. 
kdepim) are missing from the KDE 4.0 end user release.

Would it be possible to have a stable developer API release in October, but to 
release "KDE 4.0" as an end user desktop only after all important tasks are 
completed? (Is there any feedback from the marketing team on this?)

Olaf


More information about the release-team mailing list