Notes re. Viability of SyncML as an option for Chandler-to-handheld interoperability.

Notes from a vendor phone call Oct 15, 2003

Timeframe Target: 18-36 months.

Background: Currently two methods to synchronize information between desktops and handhelds; cradle sync (ActiveSync-Microsoft, and Hot Sync - Palm) and wireless (for devices that are integrated with a cell phone. While newer integrated devices are popular there are for example about 30 million PalmOS devices that will ONLY cradle sync. [While the preponderance of non-wireless devices is notable, in 2005 and going forward, it is likely that the "info-centric, early adopter" of Chandler is probably more likely to have a wireless enabled PDA than a cradle sync. only version.

SyncML History: Palm, Nokia, and Lotus/IBM started the SyncML initiative about 4 years ago. This method has been gathering momentum, but will probably become important in a big way 4Q2004. It is currently being driven by the needs of the cellular carriers for provisioning handsets (SyncMLDM) as opposed to data synchronization. The SyncML initiative was adopted by the Open Mobile Alliance (OMA) in Nov. 2002. The OMA currently owns the protocol along with WAP and a few others. There are currently about 30 "sponsor" members of OMA including most of the cell phone carriers, PDA vendors, some PC vendors, and Mircosoft. There are currently 129 SyncML products shipping including server side and clients for Palm, Symbian, and PocketPC^? handhelds.

From the perspective of a mobile device OS vendor the SyncML standard is OK from a technical perspective -- it open, flexible, and robust enough to do what needs to be done.

PalmSource^? for example does not currently provide a SyncML client in the PalmOS^? 5.x version, and it probably won't be in the 6.x version expected to be introduced early in 2004, but is likely to be in subsequent versions later in the year or the first half of 2005. That being said, there are several 3rd party clients currently available for PalmOS^? versions going back to 4.x as well as for other handled platforms. (Critical Path for example supplies such a client.)

If Chandler wants to sync with older PalmOS^? devices OSAF could provide an older style Hot Sync conduit which is a dll type of module that plugs in the the Palm Hot Sync system (LOE would be in the range of 6-8 person months) or it could provide a 3rd-party SyncML client.

SyncML as a Data Exchange Format: Yes, it can be used for this. The handheld markets are also exploring non-traditional alternative for data acquisition and exchange as well. An example is Danger's Hiptop that uses web services to get updated data. Not this is by no means mainstream. Data partitioning which is application specific is another "non-sync" method, by which an application may go to multiple sources to aggregate information, e.g. a calendar app getting work, home, and sport team events from different sources (I suspect that RSS aggregators would be another good example of this--ph)

Licensing / IP issues: From the statements on the OMA website, the goal of OMA initiatives including SyncML are to foster development in the mobile market through open, diverse, platform-agnostic solutions. For example, from the OMA FAQ:

• Q: What are the principles of the Open Mobile Alliance?
• A: The principles of the Open Mobile Alliance are:
  • Products and services are based on open, global standards, protocols and interfaces and are not locked to proprietary technologies
  • The applications layer is bearer agnostic (examples: GSM, GPRS, EDGE, CDMA, UMTS)
  • The architecture framework and service enablers are independent of Operating Systems (OS)
  • Applications and platforms are interoperable, providing seamless geographic and inter-generational roaming

As a matter of fact the OMA members are expected to identify patents they hold that may impact OMA initiatives and provide authorization for that IP to be incorporated within OMA protocols and derivative products. There is currently 12 open source SyncML projects on Sourceforge. One in particular Sync4j was mentioned as a good quality, stable, production open source SyncML server and framework released under a BSD license.

Alternatives: Apple's iSync was based on SyncML.

Comments on PalmOS datastore formats: PalmOS^? uses a proprietary native datastore that each application draws upon through the PalmOS^? "exchange manager" layer. Syncing or importing external data, for example from Chandler would be done via standard data interchange formats rather than PalmOS^? native datastore format.

Conclusions:

• There is nothing we need to do architecturally/development-wise NOW to make our use of SyncML in 6-9 months time less difficult.
• SyncML spec is relatively stable, the risks are mainly in who adopts the standard and when.

-- PieterHartsook - 16 Oct 2003