Great post - I can get pretty OCD about the way code is written, but I have a hard time complaining about things like this without feeling like a dick since as

Great post - I can get pretty OCD about the way code is written, but I have a hard time complaining about things like this without feeling like a dick since as you pointed out it's not particularly inefficient (even if there was an O(1) Clear() method, how many TabPages are we really working with that it would matter?) But you've reassured me that it's a reasonable thing to do, especially if I know beforehand that it's a piece of code that will probably be used for a long time. That said, I've learned that with an early stage startup where you're trying to iterate as quickly as possible in a desperate attempt to get somebody to care about your product, you often have to pick your battles. Just yesterday I came across this: category_count = [] for i in range(10): category_count.append( db.execute("SELECT count(*) FROM table WHERE category = %d" % i) ) For one thing, this iterate separately and then append to list approach in Python annoys me slightly (list comprehensions are so much cooler!). But far worse, it hits the DB 10 times instead of once, and no matter how small your site is you obviously can't be having that. How'd it get there? Who knows. It was written in the Django ORM, where the only way to do this is with a pretty obscure command like Object.values('category').annotate(count=Count('category')). At first we were picking up Django as we went, so at the time whoever wrote it probably had no idea that the values() or annotate() methods even existed, and the way it was written got something up on the page and working so we could decide whether or not we'd be throwing it out the next week. But, whatever, you come across something like this, go throw up, fix it and move on. And finding these kinds of issues puts into perspective smaller ones like using a for loop where you meant to use a while loop. tl;dr - Having the luxury of sexy-ing up your your code as described in the post is strongly dependent on the stage that the project/company is in.