Why Big-Oh is not a good scale for comparing algorithms?

Where did you read that? What do you think about this?

PS. I think you crossposted this across the web. What are folk telling you?

I left you hanging because this topic is fairly well done but maybe you are in your first year or so of compsci so let's read a back and forth like

All algorithms can be better or worse depending on how you measure. Sorting is one of the fun ones since results vary with how you measure.
Time: Sure but is that the only criteria? How about...
Space: As in how much space a routine consumes. Or "Sort the stars in the known Universe."
Comprehension: "What does that do?"

I'm sure there are more ways to question an algorithm but today, right now, with average datasets and a rocket in your pocket computers (smart phones) my vote goes to comprehension over speed unless speed is so bad you are asking "What is it doing?" over "What does that do?" Even a large dataset like the names of everyone in the USA isn't so large that we couldn't put that in a smart phone today.