Very few digital cameras incorporate GPS technology, and those that do tend to offer limited and difficult-to-use capabilities. The advantage of a GPS is obvious: a picture can be "geotagged," or marked with the geographical coordinates at which you snapped a shot. To support this capability, iPhoto '09 added its Places feature, which reads geotagging metadata stored in a picture to stick a photo on a map (see "iPhoto '09 Adds Faces and Places," 6 January 2009).
There are ways to make GPS work while taking pictures, such as add-on products from camera makers and third parties that constantly record coordinates. Later, you use software like HoudahGeo and Geophoto to extract a timestamp from each photo and match that with coordinates captured at that moment by the GPS recorder. Also, the Eye-Fi Geo Wi-Fi camera card uses Wi-Fi positioning to grab Wi-Fi snapshots when a picture is taken, and then retrieves approximate location data (when available) when a photo is uploaded using Eye-Fi's software (see "Eye-Fi's Geo Targets Apple for Wireless Photo Transfers," 29 July 2009).
The iPhone geotags images using the best data it has from Location Services, which means Wi-Fi positioning and cellular tower trilateration (an alternative to triangulation that relies solely on distance measurements) if the GPS system hasn't yet locked onto satellites and computed a good location.
I've tried cameras that integrate GPS. The controls are wonky, the GPS lock is slow, and it's often unclear whether or not coordinates are being captured. Cameras aren't good at GPS because they typically lack the capability to use Assisted GPS (AGPS), in which tables of computed satellite location data are preloaded into GPS receivers. Those satellite locations are critical, because a receiver has to trilaterate its current location based on the satellite's distance. You have to know where the satellite is, obviously, to produce that calculation.
AGPS enables much faster GPS locks by predicting a satellite's location at a future point of time, instead of the 12.5-minute minimum required for a "cold boot" - a startup in a different location or after a few days - for GPS devices sold a few years ago, and for some very cheap devices still offered today.
To support AGPS, camera makers would have to offer software that downloaded updated satellite tables (the "ephemerides") into a USB-connected camera from a computer. Cameras with Wi-Fi - still very much in the minority - could more easily retrieve updates from Internet-based servers. (For more on AGPS and modern usage in cameras, read my Ars Technica story from early this year, "Inside assisted GPS: helping GPS help you.")
A new module from CSR, a chipmaker that acquired leading GPS system builder Sirf, may help camera makers and other gadget designers achieve better results. Instead of producing a GPS fix on demand, when a camera is powered up or takes a picture, the chip produces a continuous series of coordinates based on the best-available fix at the moment. This is quite like Location Services in the iPhone OS.
The CSR module also has a number of features designed for mobile devices to remove interference and deal with bouncing (while holding a camera or jogging with a music player, for instance). Plus, it embeds the capability to store a week's worth of satellite location data - computed in advance by CSR - as long as the device can manage a connection.
Once Wi-Fi becomes a regular camera feature and these GPS chips (and similar future chips from other firms) become routine additions, we'll finally be able to forget entirely where we took a picture, and rely on more-or-less accurate - but always available - coordinates captured and stored in the picture's metadata. Until then, not so much.