Let’s go back to Google Earth and run some more tests, in which we apply some of these oracles. I spent another five minutes with the application and found the following problems. I’ll share with you how I found each problem, and why I think it’s a problem (using our test oracles).
Negative Values for Altitude
A quick blink test allowed me to find a negative value for eye altitude. It would seem that if you zoom all the way out, tilt all the way down, and then just hold down the Zoom In button, you can get Google Earth to show negative altitudes (see Figure 3).
Figure 3 Negative value for eye altitude.
An odd little combinatoric problem (I can’t get this error unless I start all the way out and I have the tilt all the way down), this result illustrates the following behaviors:
- Inconsistent Within the Product. I can’t zoom to a negative altitude if I don’t apply a tilt.
- Inconsistent with Purpose. The purpose of zooming and tilting isn’t to get negative altitudes—it’s to get a better view of the landscape.
- Inconsistent with User Expectations. I certainly didn’t expect to see a negative value.
Rendering Lines When Selecting Layers
Another simple test is to select multiple layers and then simply navigate around the globe, zooming in and out when it takes your fancy. When I select more than one or two layers, a series of white lines becomes visible when rendering. If you select many layers, the problem becomes very visible (see Figure 4).
Figure 4 Rendering with multiple layers selected.
This result demonstrates a couple of problems:
- Inconsistent with Purpose. The lines obscure my ability to see the actual map.
- Inconsistent with Image. This is a fairly basic problem that I discovered the first time I ever used Google Earth (months ago). It looks bad, and since I like Google’s software so much, I expect more from Google.
Different Travel Times for Directions
While trying to think of an error for an inconsistency with a similar product, I thought I might try comparing Google Earth with Google Local. Sure enough, the very first test I ran showed a difference. I ran a quick search for directions to my dad’s house and found that the suggested drive times differed by 24 minutes with the same distance in miles and the same suggested route. Figure 5 shows the distance in Google Local; compare the distance in Google Earth shown in Figure 6. This test illustrates two problems:
- Inconsistent with Product. Both Google products should get the same result.
- Inconsistent with Image. Google makes both products; their inconsistent behaviors aren’t good for Google’s image.
Figure 5 Distance in Google Local.
Figure 6 Distance in Google Earth.
An interesting bit of follow-up on this bug found that if you click the Printable View link in Google Earth, it opens a browser window with the Google Local directions, which displays the difference in time. So there go any arguments that Google Local might not be a comparable product (for this feature, at least)!
This leads us to our next find. The Google Earth help file (beta, last updated January 18, 2006) claims, "The Google Maps web page appears on the bottom half of the screen with the directions highlighted on the page along with directions" when you click the Printable View link. It even shows an image of the two applications working together (see Figure 7).
Figure 7 Image taken from Google Earth documentation.
Rather than appearing at the bottom of the page, when I click the link the web page opens in my Firefox web browser. This is Inconsistent with Claims. On the other hand, I don’t feel that it’s Inconsistent with User Expectations or Inconsistent with Image because the functionality as it worked met my expectations as a user, and I think it’s good for Google’s image to show that their application works with other products such as Firefox.
Okay, enough beating up on Google Earth. It’s in beta, after all! Sadly, I didn’t get a chance to show you an Inconsistent with History example, but that’s what happens when your sample application is in beta. However, I’m sure you get the idea.