KDE PIM/Meetings/Osnabrueck 10
The annual KDE PIM Meeting Osnabrück 10 took place from 10.2.2012 to 12.2.2012 at the Intevation offices in Osnabrück. See the dot story KDE PIM Sprint 10: ACCOMPLISHED! for a report about the sprint.
- 1 Topics
- 1.1 Projects
- 1.2 Discussions
- 1.2.1 Future Development
- 1.2.2 Marketing
- 1.2.3 libkolabxml
- 1.2.4 Server Side Akonadi
- 1.2.5 KDE PIM on Mobile
- 1.2.6 Akonadi bugs
- 1.2.7 Resource scheduler
- 1.2.8 Maildir flags
- 1.2.9 Storing calendar colors for Kolab calendars
- 1.2.10 Porting KTimeTracker away from KResources
- 1.2.11 Saving searches & results in cross-client scenarios
- 1.2.12 Contact Aggregation
- 1.3 Presentations
- 2 Agenda
- 3 Meeting Notes
- 4 Blogs
- 5 Organization
Projects we'll work on in small groups, mostly coding or creating other concrete results.
- Nepomuk integration in KMail (and possibly other PIM applications) in respect to the feeder rewrite.
- Groupwise calendar resource for Akonadi
- Akonadi importer for Polka
- Porting KTimeTracker away from KResources
- KMail migration
- KAlarm Akonadi resource
- Stabilizing Akonadi IMAP resource
- Writing: Explaining Akonadi
- Writing: Report on status of Akonadi
- Document Kolab's use of Akonadi
- Java library to access Kolab IMAP
- GSoC 2012
- Revive the Akoandi part table optimization branch (Volker: I need SQL help for this)
- Retrieval during transaction safety issue in the Akonadi server
Discussions about topics, which are relevant to all or a sub group of people. Please state audience and desired result of the discussion.
Audience: all, Desired Results: Plan regarding future 4.x releases and port to KF5
- Update on the KF5 efforts
- First release of KF5, i.e 5.0, after release of Qt 5.1 which means 2013
- KDE SC 4.9 and 4.10 will happen
- What changes do we want to do in e.g. kdepimlibs, kdepim, etc., for KF5?
- Port everything away from Qt3Support, KDE3Support, KResources, old KCal
- Drop applications we cannot port
- Where: in master
- When: until 4.10
- What does KF5 mean for KDE PIM? When are we going to port to KF5?
- Switch to KF5 at Osnabrück 11
Audience: all, Desired Results: Input on timeline, facts on good things on Akonadi, agreement on communication, group hug
We have three user groups for Kontact:
- pim developers
- kde 'family' (fellow devs, community members, power users/fans)
- wider community
Upon experiencing a serious bug, a PIM developer would (attempt to) fix it; a family member would restart Akonadi (work around the bug) and send in a bug report; while an user from the wider community would switch to Thunderbird.
What is the issue?
We're not ready yet for the wider community; but since 4.7 we are for our "family". We need them to test and help out! Unfortunately, we lost them because after 3 years of "we're almost done, it will get better", KDEPIM 4.7 was disappointing, especially the .0.
What we did was not wrong - OK, it took us a while to get Akonadi to this point, longer than we thought. But 4.7.0 was not released that early - we're only 6 months further and the vast majority of the issues is fixed. This is Free Software - release early, release often.
It was our communication during the 3 years before which was a bit too optimistic and which made people leave when 4.7, our "first" release, was seen as "again a disappointment". You know your marketeers, we are sometimes a tad too enthusiastic. I humbly apologize for that...
But that is history. Now the challenge is: get our KDE friends back.
How do we win our friends back?
We thought about saying sorry. But that just makes us all feel bad and invites a blame game. So let's just move on: we're all unhappy with the last 3 years but the good news is that we're on the way back to awesomeness. This is what we have to communicate now! Part of that is to distance us from what happened. As the terms Akonadi and Nepomuk are quite tainted, we think it's best to start to avoid them. So:
- Get rid of the old: change our communication about Akonadi and Nepomuk. For end users, they are irrelevant and their names are tainted. We need to start talking about Kontact (Desktop) with Kontact Mail, Kontact Calendar etc. Akonadi and Nepomuk are implementation details - we just have the new Kontact PIM backend and the Kontact UI's.
- And welcome the new: make clear to our community that we're on the way up now and that we need their help to find existing issues. This requires some 'proof' of what's been done and where we are and of course continued repetition of that message (status updates).
- Tell the big story to show things have changed: The bold task of refactoring the solution is now fully completed. We started from a mixture of single applications, each with GUI and data handling aspects, pressed into a thin shell. The aim was to create an groupware software product that is more integrated, but technically als more modular with the technical border drawn at the right places. Now it all came to gether for the first time we have an integrated Kontact experience. Technical there is a non-gui pim backend and several graphical views on it. TODO: Add graphic to show this transition.
- Change the wiki pages to reflect 1 - Done
- Communicate 1 to the marketing team. (we won't do a big announcement or anything, just mention it in the sprint report & start using it) Jos
- Find a good time to do 2 - in 1-2 weeks (ask for help with cleaning up bug list, not sending bugs yet)
- Gather the needed information for 2 (list of bugs fixed/improvements over last months/year/etc, maybe comparison to other solutions etc) (Sunday task)
- Start to address some myths and write 'pull' article(s)/blog(s) about the real, refactored Kontact
- Write an article for 2 (Jos, in progress)
Audience: people interested in Kolab, Desired results: Decide about exposed payload formats
State & implementation of the new XSD based Kolab library, see
- Which payload format could/should be exposed through ASAP for Server Side Akonadi clients?
Server Side Akonadi
Audience: people interested in Kolab, Desired results: design decisions
Points to discuss:
- Dependencies, in particular kcalcore, kabc
- Formats, see above
- D-Bus related concerns for server architecture
KDE PIM on Mobile
Audience: mobile people, Desired result: Plan for future developments
How is Kontact and other parts of KDE PIM doing on the N9, on Android, on other mobile platforms?
What are our goals for future development, is the current variety of target platforms and features sustainable? Who is working on it?
Kontact Touch should be available through the community app store, rather than hand-b0rked installation, see
How to make that happen?
Audience: Akonadi developers, Desired results: Update list of bugs
Getting input on a number of Akonadi bugs
Audience: Akonadi developers, Desired results: Design for resource scheduler
- ResourceBase scheduler is designed for the simple, common case of serial access resources.
- IMAP wants to run tasks in parallel, so that'll need a different scheduler
- IMAP also wants a specialized 2-phase sync
- Custom scheduler and custom sync would mean by-passing most of ResourceBase
- David will implement a ResourceBase base-class that provides the resource interface but not a built-in scheduler and syncer
- IMAP will be re-based to that new base class and get a custom scheduler and syncer
Audience: KMail developers, Desired result: Decide how to handle maildir flags
Discussing handling of maildir flags
Andras: Maildir should write out correctly the message flags into the file name since 4.7.2.
Storing calendar colors for Kolab calendars
Audience: Korganizer people / people interested in Kolab, Desired result: Implementation (plan) for storage of calendar colors according to KEP 12.
See KEP 12: Color configuration storage for resources and categories for the specification.
Porting KTimeTracker away from KResources
Audience: Calendaring people, Desired result: Port KTimeTracker away from KResources
KTimeTracker won't be ported to Akonadi because it only uses ical storage as an implementation detail. The ical file isn't exposed to the user.
Sérgio is currently porting KTimeTracker from KResources to KCalCore. DONE
Saving searches & results in cross-client scenarios
Audience: people interested in Kolab, Desired result: Design parameters for system that would work for Akonadi as well as server side clients, e.g. Roundcube & Co.
This is a subject that has been started as a discussion in the Kolab community, but never been fully thought through or finalized as a design. For input, see KEP 15: Saved searches and sharing searches across all clients
Audience: People interested in contact management and/or Nepomuk, Desired result: Coordination with KTP team
- conf call with KTP team to coordinate current efforts Sun 13:00
- plan on integrating that into kdepim mainline
Presentations of things interesting to the KDE PIM community. Please state targeted audience.
- Bachelor thesis: REST interface for Kolab (Thomas Koch) (Audience: people interested in Kolab)
- Server side Akonadi (Georg Greve) (Audience: people interested in Akonadi and Kolab)
- libkolabxml (Christian Mollekopf) (Audience: people interested in Kolab)
- Plasma workspace integration (Kevin Krammer) (Audience: people interested in KDE desktop)
16:00 Start meeting, fill agenda, get organized
~19:30 Dinner at Chow's Garden (Japaneese / Asian Food)
10:00 Review first batch of work, get organized
10:30-12:00 Presentations, discussions targeted at whole group
13:30 Group photo
14:00 Stopp-ACTA Demo
17:00 Review second batch of work, get organized
17:30-18:30 Presentations, discussions targeted at whole group
20:00 Dinner at Arabesque (Arabian Food http://www.arabesque-osnabrueck.de/)
10:00 Review third batch of work, get organized
10:30-11:30 Presentations, discussions targeted at whole group
13:00 Conf call with KTP team
14:00 Close meeting, collect next steps
Store in Akonadi::EntityDisplayAttribute, propagate to/from mailbox annotation in Kolab resource. Christian will implement it.
Input from both our needs in the IMAP resource and our changes to IMAP for ASAP are relevant for the IMAP 5 efforts. Some specific points mentioned were:
- remove concept of currently selected mailbox
- ability to receive change set since last sync
- change notifications for more than a single mailbox
Georg will find time for Kevin O. and Volker to participate in the process.
KAlarm and Akonadi
A major bug in KAlarm whereby it failed to create default Akonadi resources on first initialisation and therefore became unusable, has finally been fixed. It turned out to be due to a combination of bugs in KAlarm and in Akonadi. This would have been very difficult to fix without face to face discussion and debugging. Thanks to Volker for his help.
Sergio and Thorsten agree that ktimetracker should be ported from KResources to KCalCore.
Sergio ported ktimetracker from KResources to KCalCore.
Part Table Optimization
SQL upgrade code done (thanks to Bernhard Herzog), however the process takes ~10 minutes to complete.
Various options on how to communicate that to users have been discussed. A notification UI process (ie. a simplified Qt-only kdialog shipped with Akonadi) will be launched by the Akonadi server for this. Volker will implement it.
Other still pending tasks:
- Test database schema upgrade with SQLite
Akonadi Git Resource
Sergio fixed remaining important bugs and moved the project to playground.
Desktop-wide progress/status notification for agents
Use KNotification from inside AgentBase/ResourceBase for progress and error/warning signals. Other options discussed included a kded module or akonaditray, but they add to much infrastructure complexity for no real gain.
- Osnabrück KDE PIM Sprint 2012 (Kevin Ottens)
- A very productive KDEPIM meeting in Osnabrück (David Faure)
- Releasing Polka 0.9 (Cornelius Schumacher)
See the Osnabrück 10 organization page for organizational details.