Problem tags, which are supposed to help people categorize problems, some of them doesn't make much sense and need revamp.
For example, take a look at 2057E2 - Another Exercise on Graphs (hard version). At the time of this blog, it has tags "binary search" "brute force" "dp" "dsu" "graphs" "shortest paths" "sortings". Some of them doesn't make much sense to me.
"dsu" "graphs" "shortest paths" are fine imo. They either describe the core idea or the main techniques the problem use.
Then "dp" might be a bit controversial, it is probably about the fact that the modified Floyd algorithm is based on DP. It could be reasonable, especially if you didn't recognize that you are using Floyd algorithm when adding edges dynamically. The editorial also agrees with this idea. This trick seems rather classical to me. I don't think this is a big problem.
However the rest of them does not make much sense imo. For "binary search", I guess it's about the binary search approach for E1. That is too slow and cannot pass E2. And for "sortings", the only place to use "sorting" is the process of calculating the MST (you need to sort the edges), which isn't really related to the main idea of this problem. I guess a more suitable place for this tag for example, would be a problem that involves some greedy observation, and the conclusion ends up to be "sorting the elements in some order" as the key part. "brute force" makes absolutely no sense. Again, a better place for this tag might be some problem with a brute-force like approach that turns out to have a good complexity.
There is no fixed criterion to sort problems, and there is much space to argue. But in my opinion, problem tags are for helping people categorize problems (and decide which problems they will practice with, for example). The examples listed above don't serve this purpose well. I suggest revamping some of the tags and writing a brief guide about what each tag means, so people can have a common understanding of them.
If you have doubt regarding any of my opinions above, please comment in comment section and discuss. Thanks for reading.
I think it's because a lot of people are dying to tag problems with the methods they used to solve it. So maybe only the problem authors should get to set tags.
I think if one problem has several different methods to solve it, putting all of them on should be fine. But the examples listed above are clearly not of this situation. They are likely caused by the different understanding of people on the same tag.
no one reads tags
I feel like they're used to search for problems (using bots or manually) that can be solved with a certain topic in order to practice it. That's why tags shouldn't be unrelated/misleading.
no one except greys (actual or in spirit) cares
fair enough lmao
ThemeCPers in shambles after this statement
here is the fix to this issue:
Go to profile, click change settings, turn off showing tags of unsolved problems.
Easy. I really hate this is not the default option of Codeforces...
I think that not turning that off would be even stranger if tags were accurate
UPD : I am sorry, I misunderstood you
That's what he said
I can see it on problemset page
The "binary search" could be refer to binary search on the answer when answering queries online, for a $$$O(q \log n)$$$ solution. I agree that the tag is not necessary; the $$$O(qn)$$$ solution also passes.
how to pronounce your last name
why would you not respond? I would never ignore someone who asked me how my last name was pronouncd
how is your last name prounounced
Koh-vaks
how to pronounce your last name
Are you hungarian?
custom tags should be allowed. make it possible but not too easy to create a new tag. it will solve the problem.