I understand what Hoppy's saying. Basically, it can be quite difficult to match a particular shade of a color to a color chart when you're holding a test tube in one hand and the chart in another. Now imagine that tube being inside a tank full of water, etc.
Secondly, the accuracy of that particular test (the kh/pH relationship) has been shown to be quite variable depending on the presence of any other acids in your tank. Many folks have thought they had much higher or lower amounts of CO2 than what was in actuality due to these errors.
Thirdly, the method is accurate in the sense that it is based on a simple acid/base colorimetry reaction where a pH indicator will change colors depending on the pH present.
Clear as mud now?
Personally, I prefer the degassed tank water measurement followed by increasing the gas and observing your fish.