At some basic level, Windows CE's GPSID and Location Framework are doing the same thing. They're giving you the position of your device. I've had people ask me if with Location Framework if we were going to retire/deprecate/etc... GPSID.
First, Location Framework depends on GPSID in order to retrieve GPS lat/long for it. Second, we care about back-compat. There are enough apps out there using GPSID that we're not going to go break them just because there's some cool new technology.
Most interestingly, there are still scenarios where GPSID makes more sense than using Location Framework. To call out scenarios where they're interesting
Location Framework is better when you need abstraction. For instance, you don't know the type of underlying location hardware your platform will have or may potentially add in the future. This is especially true for Windows Mobile type devices. It is also better when you need data other than lat/long (like city/street/address) or where you know you will not be dealing with a GPS device.
GPSID is better in those cases where the abstraction of Location Framework is unneeded or even gets in the way. For instance, if you are an OEM doing a navigation system for a car and the car has GPS built into it but no other underlying location sensing technology and no means to add new technologies by the user, then you could probably just use GPSID. (Note that in the long-term your care may have some non-GPS technology, so even here Location Framework may not be a bad call.) Also Location Framework ends up removing information that is overly specific to GPS that it returns to the application. GPSID lets you get information about satellites in orbit, more precise information about the type of accuracy, etc. Precisely because Location Framework has to work with any lat/long producer in the world, it can't return applications such low level details that are so specific to a particular technology.
So I think long-term Location Framework is the clear winner for Windows Mobile, and it has benefits for future proofing general embedded devices.
[Author: John Spaith]