the tiger data and Kim's nav system is very helpful as a supplement to written instructions. It is great to know what street you are on, what address range you are in, what direction you are heading, and be able to use the moving map as a reference if you sort of know the area.

You know what? This statement got me thinking...

I would still pay for that functionality alone. I mean, even if it didn't even attempt to route-plan for me... If I simply had a moving map on one part of the screen, and a arrow with range/bearing to my destination on the other part of the screen, I would love to have that on the empeg display. I could still get anywhere I wanted to go with that kind of data.

I know that there are more sophisticated nav systems already available which can do proper on-the-fly route planning. But those require room in my dash for the large screen, something I no longer have now that I've got the empeg. So until I can buy an integrated 1.75-din NAV+MP3 device, I'm going to be content with whatever I can get on the empeg.

Now here's something else that just occurred to me. The Tiger data doesn't have enough detail to do proper route-planning. But if I log into MapQuest.com, it will route-plan and will give me proper directions (well, most of the time. I've seen it glitch before).

Of course, we can't get our hands on the dataset that MapQuest.com uses, so it's no good to us.

Or is it...

Let's think for a second. Before we leave on a trip, we usually know our destination, correct? Well, what if we punched it in to MapQuest, have our (hypothetical) Windows client software screen-scrape the HTML that MapQuest spits at us and extract the route-planning instructions. Click a button on this client software and it automatically squirts the route into our empeg. That route is fed to Kim's software which simply parrots the MapQuest instructions while using its internal Tiger data to verify we're on the right road and display our progress.

The only thing we'd lose is the ability to accurately route-plan "on-the-fly". But if we can punch in an address and we've still got the moving map and the range/bearing data, that's still useful for on-the-fly calculations.

Of course, where the Tiger data and MapQuest data differ, there will be glitches. Okay. And it's a bit of a rube-goldberg solution for something which should be much slicker. But until we get our hands on better map data, it's at least something, right?

Hell, I'll even volunteer to write the screen-scraper, that's pretty simple. I've already got software in VB that does this sort of thing.

Surely I can't be the first person to have thought of this?