This post is to bring some light to areas that Leetcode needs to work on, and make the experience better.
When submitting TREE related problems, LC just doesn't make it easy to recreate the tree for us to understand. When there's an error, the input variable is jibberish (an array with a lot of values, NULL values, and no way to decipher the tree) and we're unable to recreate the tree, and try optimizing/correcting our solutions. This has been really frustrating. The sample test cases are too basic while the submission-test cases have some really nice complex edge cases.
Filtering out poor quality questions. Have you wasted hours trying to solve a question, when you find it's too frustrating and give up after downvoting the question? I have. Turns out, a lot of the questions that are frustrating due to lack of 'quality' have been downvoted by our members (good job btw) but LC hasn't taken them down. To the newbies and the people trying to prepare with limited time, DO NOT try problems if they have a large number of downvotes compared to the upvotes. (There could be a number of issues, from ambiguous descriptions, to poorly elaborated test cases). LC needs to weed them out, and have a clearer benchmark on future questions.
Wasn't there mention of a Leetcode Dark mode? Wouldn't that be cool? It was mentioned couple of months back, haven't heard back yet.
Do you think you've got some pointers to add, make your LC experience better? Add them to the list.
Code on!