RE: Difference in 1.4.10 Testing Approach

This illustrates what I have always argued, that 1.4.10 is about zooming to
400%, so that is how we should test it. Testing on a  window artificially
reduced to the required pixel size is effectively just using an "emulator",
it is not testing the real thing! If the developer does something really
clever or unusual, perhaps using JavaScript, you run the risk of not
finding issues (or oppositely finding too many) that the real life users
using real life zoom actually see.