Alexey Feldgendler schrieb:
On Wed, 30 Apr 2008 02:54:08 +0200, Shawn Rutledge
<[EMAIL PROTECTED]> wrote:
my list starts with:
*backing up phone settings, including list of installed packages (in
case i need to re-install everything)
*backing up e-mails/text messages
*storing gps coords in case the phone's stolen/lost
*backup/storage of any other arbitrary files
You would want it to be a commercial service like .mac, with a
subscription fee and guaranteed reliable service, and use GPRS to
access it? Or you would want to just do that stuff with your own PC?
The difference between an open technology and a closed one is that
while Apple is the only one who runs a .Mac server, anyone can run a
server in an open technology. Someone can install software based on
open standards and run a server, and even charge for its use if they
like. Don't like paying for it? Set up your own!
I believe that an open standards based server solution for assisting
mobile phone features should be created. The most important thing here
is that the data formats and protocols are open, so that anyone can
implement a client or a server. There are several proprietary services
with their closed data formats out there, but no really open one, as
far as I know. Developing a consistent and comprehensive specification
is an immensely important step.
I would define the scope of the technology as “sycnhronization, backup
and remote control for mobile devices”. Here is a very vague, draft
list of requirements:
* Allow backup and incremental synchronization of typical data items
stored on mobile devices:
* SMS/IM histories
* last dialled/received calls
* contact lists
* browser bookmarks, histories and cookies
* calendar events
* various account information (GPRS, HTTP proxy, email, IM)
* stored certificates
* global preferences like language, time zone, visual theme
* software-specific settings
* user documents
* information about installed downloadable addons
* As much as possible should be synchronizable across devices and
software platforms.
* Leverage existing open standards wherever possible, e.g. use vCard
for contact information.
* Provide automatic conflict resolution.
* Be bandwidth-efficient.
* Survive long periods without connectivity and synchronize reasonably
afterwards.
* Don't insist on complete synchronization before parts of data become
usable.
* Be secure so that even passwords and certificates can be trusted to it.
* Do not require server-initiated activity because it's often
impossible, i.e. rely on pull only.
* Allow for interchangeable transport layers to make use of various
connectivity options available on devices:
* HTTP
* various serial port-like connections like IR and BT
* maybe offline synchronization via memory card
* Allow automatic offloading of data that's typically moved from
device to PC:
* pictures, video and audio recordings
* various logs
* Probably allow sharing or publishing parts of user's data.
* Provide remote control and reporting infrastructure that allows:
* upgrades of device software
* activation of special modes on lost/stolen devices and obtaining
information from them
* Transparent expansion of device's storage, e.g. the user can
continue taking pictures without thinking about where they go, and the
older ones are pushed off to the server.
This is an excellent starting point. I also like very much the ambition
to reach out beyond Openmoko (even if it is the natural platform to
start such a projec).
I would very much like to participate at such a project (or even kick it
off ftm), as my the plans for my PhD work include creating a service
that would fulfill an important part of this: The transparent storage of
data in the [Web|Cloud|Grid|<newest_buzzword_for_distributed_networks>].
Who's in?
regards
Christoph
_______________________________________________
Openmoko community mailing list
[email protected]
http://lists.openmoko.org/mailman/listinfo/community
