>>more/better controls, better user support, better interaction, graphic tricks,
>>in general more hand-coded functionality in C and added user exp in not code-free V
>>which I agree is still MUCH easier to implement in .Net compared to fwk-enhanced JS.
>>
>>The second way is the possibility of a not too eye-candy heavy application,
>>where still every field has oodles of DD-based help/user info/hints/client side "standard" validation,
>>local lookup values and so on working on once-loaded client data.
>>Speed IS a feature, caching a part of it.
>>Also when only a phone connection is available, having an option to work offline and resync later
>>when WLAN or cable-LAN is available again for some road warrior scenarios is heaven sent.
>>

>At which point you can offer a WPF/WCF app using the Sync Framework with SQL server. Fast. Pretty. SQL based (reliable, secure and scalable), recognizable current technology (read MARKETABLE), and supportable by any of hundreds of thousands of .NET developers.

Ok, did not specify that point, but I think of myself neither as an automatic MS-basher nor a blue-eyed MS fan boy.
The weak point for that in my eyes is support for xPad and even perhaps large screen xPhones
(5.x' is perhaps usable for some people, 7' ***for me personally*** is a great mesh of portability and readability).
SQLite is probably the most often installed data base engine (Firefox, Thunderbird, Symbian, iPhone...)
which makes it a great contender for the local data engine as well:
you have to work with it IAC for reading data on ARM-devices, so get used to it's shortcomings.

If Silverlight vision/promise had been kept alive on HW running Symbian or at least the more tasking OS '
with >=20% market saturation (meaning OSX plus android) I would have been very happy coding Ironpython in it.
They developed a very good implementation before it became "less strategic" and I was impressed.

regards

thomas