Who but a monomaniac programmer would care how the HTML looks? It doesn't affect appearance of the page view one bit.

The only justification I can think of is that the interpreter may convert lower case to upper before interpreting the tag, but it seems to me the code is going to run through the motions regardless, because first determining if it is upper case and converting if it isn't would take a lot more CPU cycles.

I type HTML and content as a single stream, and upper case tags both take up too much space on the line, and make it more difficult to read and correct the content, which is the actual object of the exercise, not "pretty code".

Perhaps a person who just adds HTML tags to someone else's content would look at this differently, giving more visibility to the tags.