Re: [Opensync-users] Getting into KDE-PIM for KDE 4.2

[Daniel G. <dg...@su...>]

On Friday 28 November 2008 10:16:11 Dotan Cohen wrote:
> Apparently the deadline for getting OpenSync into KDE-PIM for KDE 4.2
> is approaching fast:
> http://www.nabble.com/KitchenSync-%2B-OpenSync-Deadline-to20719171.html
>
> Can OpenSync try to work more closely with the KDE team to ensure it's
> inclusion in KDE 4.2? KDE-PIM is very important for OpenSync exposure
> and to actually have users using the software.


http://opensync.org/wiki/OpenSyncDeveloperMeeting2008Berlin

The last OpenSync developers 4 of 10 attendees were KDE-PIM hackers attending.
So this is already pretty close - isn't it?

It's not about the missing communication between the KDE team or the OpenSync 
Team.

It's just about that OpenSync is not ready!
(This got also mentioned in your linked thread two times)

The user expection for 0.40 a way too high - we will not meet them since 
they're quite "utopistic".
The 0.40 release will not be a "everything works out of the box"-release. It's 
a: "we surivied a development-branch"-release.

If we would ship 0.40 today - and quickly adapt KitchenSync to have something 
in KDE 4.2, then people would get pretty annoyed about 
KDE/KitchenSync/OpenSync and freedesktop in general, due to the lack of 
quality of synchronization. And will never look back to KitchenSync or 
OpenSync since the very first impression is what counts. If the very first 
contact with OpenSync/KitchenSync ends-up in:
- that your mobile gets unusable
- your calendar/addressbook is full of duplicates

Please understand that this kind of implementation needs pretty strong QA and 
only a complete implementation is worth to release as "stable".

Right now there only fews plugins using "capabilities" to avoid duplicates. 
It's totally dangerous to sync anything without _full_ capabilities support - 
dataloss or duplicates are guaranteed.

Since we don't have yet much real-life testing of the capabilities 
implementation, just unit-tests, we might need to change the plugin and the 
frontend API. So that's why the frontend development and the plugin 
development gets regularly starved.

I hope this gives pretty much a picture whats going on. Since we slipped 
already several month we decided to only concentrate on high-request plugins. 
KDEPIM3 and 4 (aka. akondai-sync) is among those - as evolution and SyncML is 
(and some more i would have to look up). Right now we try to draft the final 
API and freeze it. The only changes i plan to allow after 0.39 (the last 
release before 0.40 - the stable release) must be very critical. This could 
be changes to fit the capabilities/discovery for frontend and plugins.

The SyncML plugin is the hottest capabilities/discovery candidate - for this 
we might need more changes in libsyncml to get everything out of DevInf, to 
avoid duplicate entries. There will be no 0.40 before this is done (due to 
needed API changes)

Your help is more then appreciated - we have tons of task to do till 0.40. 
Like reviewing and documenting the public API and collecting 0.2x 
configuration and write XSLT conversion stylesheet for smooth OpenSync update 
from 0.2x to 0.40. Last but not least testing, testing and testing.

best regards,
Daniel