Re: [SyncEvolution] [Opensync-devel] OpenSync: fragmentation is harmful
Hi,
just let me add some random thoughts to this topic:
Am Dienstag, den 04.01.2011, 20:04 +0100 schrieb Patrick Ohly:
On Di, 2011-01-04 at 17:12 +0100, Michael Bell wrote:
> Hi Patrick,
>
> I think an important foot note is the question which goals the projects
> have. I am personally (as IT manager of a university) see three general
> needs in terms of synchronization:
>
> 1. a groupware server (with Active Sync support)
> 2. a sync solution on mobile devices (for Linux mobiles)
> 3. a local desktop solution to backup devices
[4. bridge mode: device to desktop to server]
Agreed.
I don’t know where you would include this use case, but I personally see
the primary use of SyncEvolution in cross-desktop sync. I have a desktop
PC at work, a laptop at home, and a netbook for travelling. I use a
SyncML capable PIM provider (currently Memotoo), which you might count
under #1, except for SyncML support instead of ActiveSync.
I don’t know where the technical benefits of each lie, but I would be
happy to see OpenSource groupwares support SyncML properly instead of
focussing on ActiveSync for proprietary mobile systems. I think the
focus on ActiveSync for OpenSource groupwares stems from the fact that
they see their market niche as Exchange drop-ins on Linux servers in an
otherwise proprietary environment (Outlook and Windows Mobile as
clients). There are less OpenSource groupwares with OpenSource clients
using open protocols. (eGroupWare supports SyncML, as does Horde, though
I’m not sure about the completeness of their implementations. Kolab has
found its own IMAP-based solution, which is open, but probably less
standard.)
What I see the most realistic alternative for my current use case is
Ubuntu One, given they had complete support for contacts (currently
there), notes (they have notes support via Tomboy), calendar and to-do
(still missing). Ubuntu One (via CouchDB) doesn’t have the always on
logic of e.g. Google Calendar, but stores offline and syncs when you go
online. It has the advantage of being completely silent in the
background and syncing when needed (which is currently missing in
SyncEvolution, although periodic syncs are one step in that direction,
and AFAIR, change-triggered sync was at least planned).
But in contrast to Ubuntu One (or OpenSync,for that matter),
SyncEvolution (with full PIM support) is there, works, and is stable.
And I think this was one of Patrick’s central points.
So from my POV, I can say that SyncEvolution does a great job for
cross-desktop sync as a client (#1 for the client part, given a SyncML
capable groupware, #2), it looks promising as a SyncML server, even with
bluetooth support (#3, #4), so what is really is missing is a good
OpenSource groupware server with proper SyncML support. (On the other
hand, a Kolab backend for SyncEvolution would be a nice step, too ;-) )
Cheers,
Frederik