What Convergence?

I have spent a fair bit of my time navigating. I've climbed mountains and hiked through wilderness for weeks with nothing but a compass and a map. I even teach navigation to ROTC cadets periodically.

When I finally succumbed to the GPS, I did some research but was overall pleased that – at least for the top brands – they abide by navigational standards. So they integrate nicely into the whole system. Which, with their propensity for minor failures, is a good thing.

For close to ten years I've been waiting for convergence to bring me a GPS in a phone. And for much of that time, after peripherally working on some of the systems myself, I've been instead waiting on a useful GPS in a phone. The N95, being a great device and noted as being good for GPS, made me excited about this again.

We've griped before about how location is not just GPS and maps (you can use towers, and use the data to make other services more contextual). But what I still find is that a GPS receiver still does not make a useful GPS unit in a phone. Sure, you get maps, but there are a stack of things that are just wrong with it. At least for someone like me.

Depending on how broken down I am, I run a couple times a week. And I use my Garmin GPS60 to track the time and distance, and tell where I am if I get lost. So this has been a baseline "useful GPS" test for some time. I did finally determine Nokia offers their Sports Tracker (although it's beta, so is not officially supported on many devices) which seemed to give speed, altitude, distance counters and so on so it might replicate the functions. So I took the phone for a run and tried it.

And it was pretty painful. Most of the experience was significantly sub-optimal, but let me try a few examples.

Where's the path?

Sports Tracker has no map. I mean, it has a path chart, sorta like older non-mapping GPS units, but there's no map data overlaid there. Maybe I was missing something, but it sure was easy to miss then.

So I tried Nokia Maps and Google Maps for Mobile. These all do, thankfully, run at the same time, so I was able to switch around between them to see what was happening. They still didn't help. I could tell where I was, but there was no track. The track on a GPS unit is a little dotted (usually) line showing where you've been. I like to keep them for years, so I can tell where the turnoff is to an obscure scenic outlook, and so on. For a local run, they have the same use. I can see old tracks and tell where the good (or bad) routes are, and I can backtrack if I went the wrong way.

This is a key feature for a GPS navigator, which is essentially entirely missing.

You are heading "right"

Standalone GPS units, and I guess mobile phone mapping software when route tracking, orient the map to your direction of travel. This is taken from real life, where a good way to walk around in the woods is to orient the map to your direction of travel; things ahead of you in the world are ahead of you in the map.

Orienting maps to the real world helps with mental models. It's called "mapping" even when no maps are involved; light switches should be "mapped" to their actual locations by relative location and orientation on the panel, and so on.

Disregarding Sports Tracker, which has no useful map (and if it did, it is still tiny) the two map nav programs I used orient north-as-up. Always. Again, I tried a lot of settings, but could not get them to switch over unless (sometimes) I was in route-following mode. And that was useless for two reasons: 1) route mode seems to insist on displaying in a cute 3D display, so the map is not really readable, and 2) I don't want to run a pre-planned route. I'd rather use paper.

Advantage: Garmin. They do indeed orient the maps north-as-up when you zoom out far enough (around the time the state takes up the whole screen). Below that, they automatically orient up as direction of travel. And there's no way to change this, because it's the right thing to do.

Waypoints, units and accuracy

There are three other features I use all the time on a GPS. The first is location. I very often want to know my actual location. Not as a point on a map, but as coordinates. Now, Sports Tracker does have this, but... not very usefully. For a lot of reasons I don't like lat/long (I use MGRS or UTM, and if you ask I'll tell you why) but even within lat/long, there are several ways to communicate it. DD.ddddd is only one of them; not being able to change it means I cannot really communicate with other nav devices.

An absolutely critical feature of every GPS is setting waypoints. You can do it before the trip to plan it, or do it on the fly. "Hey, a place that sells honey!" (this is real), then you mark it, label it at your convenience, etc. Absolutely no such feature seems to exist on any software I can find for any mobile. Yes, I should mention that Garmin and some others do offer software for mobiles, but it's so expensive that you can buy a cheap piece of actual hardware, so I don't know anyone who has it. I also just got to poke around with Sprint's navigator solution on a new Blackberry, and it has a sort of waypoint feature, but assigns a street address to each saved point; a sort of odd solution, that also misses the point of the coordinate systems above.

And lastly, there is another issue that is dear to my heart, precision and accuracy. Here, everyone is at some fault. GPS, like most electronic devices, implies great accuracy due to its display of data very precisely. But this is not true; only so much accuracy is available, and sometimes this is shown, but not robustly enough to me.

The Garmin unit, and some of the mobile phone software I have used has a precision circle, but all of it puts a very precise-looking center point, and (when available) displays coordinate location to many decimal places. I have been driving while others navigated incorrectly because a GPS (or more often, a phone) said we were a block, or a quarter mile, to the left. These devices need to work on ways to better visually imply they only have a certain amount of accuracy, and help avoid such use errors.

Read another blog post where I gripe about many of these same issues.

Datum and trust

I ran with both the GPS60 and the N95 running Sports Tracker on a couple of runs. At the end of the run I encountered one of my biggest concerns about this: they disagreed. See the image at the top of the posting. How is this possible? Well, I have some ideas, but they relate some to the issues with data above. There is no way to set (or even see) which datum is being used. This can mess with position indication, and position calculation. Sure, most users won't want to mess with it, but if the data is wrong, shouldn't there be some settings to let you try to fix it?

I want to be clear that I'm not hating on Nokia, or Google, or any specific product. In fact, these are finally good enough they are worth this sort of use and review. I don't even want to talk about trying a couple years ago to get a Nextel GPS to display useful data on the device. But this seems to be typical of far too many products that get "improved," or integrated (say, into mobiles). Core functions get lost or modified just a little, but enough to make them a lot less useful.

And I'd maybe gripe about how this could be fixed, but I do have hope it might get better on it's own. Camera phones used to be universally comically bad, but I might have bought my last point-and-shoot camera. Our current devices all have cameras as good as, and with more storage and easier transfer, than most standalone devices. Maybe I just need to wait another generation or two to see GPS – and other convergence features – that work in really useful ways. Or maybe they will always stay like this, as consumer driving navigators, and like my DSLR is the serious camera, standalone GPS will always be the serious navigator.

My father said, in email (note I am third generation engineer):

Since I use a GPS on my bicycle and ride in areas where cell phone coverage can be spotty I would want any GPS capability to be able to function independent of the phone network. That would include some method of storing map data on the device before I lose my phone coverage. It might be acceptable for me to download data before entering a no coverage area.
WAAS is a capability that sends error data to GPS receivers. My first GPS device in 2001 had this but it was slow to process that data. With WAAS the location uncertainty could be under 10 feet without it the uncertainty could be 3, 4 or more times greater than that. I once tried to use a GPS without WAAS enabled to record a run on a one mile loop. The data points were all over the place so the distance information was useless. Do phone based GPS programs take advantage of WAAS?

The DATUM is probably only an issue when transferring information between devices. That is if you were putting in coordinates that were measured using one Datum you would want to have the receiving device set up to accept them using the same Datum. This is not a major issue if you’re dropping a large atomic bomb but it can make a difference in several feet in some places. If you are flying using GPS it’s nice to know where the ground actually is.

Orientation is an issue. North up is good if you are comparing your display to a map. Instantaneous direction of travel is good if the route is relatively smooth (not twisty). I’d like to see a “next waypoint up” mode. That might be nice on some of these twisting roads since it could give me a better idea of what’s ahead without the display constantly turning. When stationary (and held horizontal) the display on a hand held unit should switch to a north is north mode. (It would probably need an internal compass for that which would require more power.)

Maps are out of date and have errors. I could spend $120 to get a new map database for my Garmin GPS and have no assurance that the local problems have been corrected or updated. At least Google has corrected Old Bible Creek Rd. to Old Bridle Creek Rd and it didn’t cost me.

I’d like to be able to mount my weather proof cell phone on my bicycle handlebars and get rid of the GPS. (Note the key words in that sentence.) Comment #1 is a major roadblock to that. It seems that the cell phone companies see their market as urban/suburban texters who want to find the next coffee shop.