Hi Lisa,

Your argument against putting much energy into peer to peer sync with
mobile devices is pretty convincing.  It would take a huge amount of
work to have OSAF support multiple devices on multiple platforms in a
polished way, and the return on investment wouldn't be good, because the
devices would be outdated by the time we finished polishing.

I was almost ready to say, OK, I'll just wait 1-5 years until I have a
device that does CalDAV over a cellular data connection, until then my
Chandler sync urges will have to be plowed into something else.  But on
reflection, I wonder if this might be a place where a little effort
might go a long way.  Could there be a middle path between no device
sync and committing to a professional vision of sync?

What if we put a modest amount of staff time into:

A) Leveraging our existing sharing framework to provide a basic UI for
syncing calendars to mobile devices, ending in stubs with solid APIs for
processing inbound and outbound sync operations

B) Documenting how those stubs could be built out with non-OSAF parcels
on specific platforms for specific devices

The idea would be to have the UI already set up to choose a preferred
device, a preferred collection to sync to, whether recurrence is
supported, whether timezones are supported, whether the sync should be
inbound, outbound, or both, and what to do with conflicts.  Authors of
specific device/platform parcels would then only have to define a set of
capabilities and implement the transport.

I think doing this much would go a long way towards encouraging lots of
people who want minimal export-to-device functionality to write
something that could be shared.  In general, I think there's a lot of
value for an open source project in creating stubs and documenting an
API in cases like this.  It makes for a much smaller bite to help out
with the project, and I think this is exactly the kind of itch
developers like to scratch for themselves if the bar is low enough.

I know I really want this, if we had the UI and APIs in place already,
I'd write the export code for my phone on some weekend, now that I've
figured out how to manipulate the calendar over Bluetooth.  OK, maybe
I'll just go ahead and write the whole thing on a weekend, but I'm not
making that promise. :)

For myself, creating events manually on my phone is enough of a pain
that I'd happily accept just export functionality, I don't need full
bidirectional sync.

Sincerely,
Jeffrey
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/dev

Reply via email to