Codeforces and Polygon may be unavailable from December 6, 19:00 (UTC) to December 6, 21:00 (UTC) due to technical maintenance. ×

cry's blog

By cry, 3 months ago, In English

Hello Codeforcers, especially fellow problemsetters.

I'm lazy. I like setting cool problems, but I don't enjoy preparing them. Sometimes, I take longer to prepare a problem than coming up with the idea. Surely I'm not the only one. My laziness has motivated me to make this blog about a feature I think will benefit the future of problemsetting and more frequent contests for codeforces.

I propose that there should be a hub for crowdsourcing useful generators from past problems. Similar to a hub for Minecraft mods, I think there should be something similar for testlib generators. Authors of past rounds should be able to upload generators to that hub. They should be able to specify the tags (number theory, brute force, tree) solutions that their generator targets to TLE, and be able to specify what makes their generator special. For example, there can be generators focused on special kinds of trees required to make strong tests for a certain tree query problem. Future problemsetters should be able search up the types of generators they need for their own problems and examine these posted generators and modify it to their own needs. This should save them a lot time. Sometimes, they can recognize other problems on CF and think "man, it would be nice if I had a generator that this problem also requires".

Here's an example, consider 1985E - Secret Box. The solution to it requires to loop through all possible factors of two of the three integers given in the input. Obviously, inputs that are big but also have a lot of factors should be nice. If only someone made a generator on selecting random numbers with large amount of factors in a range, like a more extensive list of Highly Composite Numbers :weary:

Anyways, TLDR: this would be quite cool though :v

KAN, MikeMirzayanov

  • Vote: I like it
  • +267
  • Vote: I do not like it

»
3 months ago, # |
  Vote: I like it +61 Vote: I do not like it

Actually would +1 on this, personally it'd lower a huge load of overthinking-without-knowing-a-way-out in my head. ;)

»
3 months ago, # |
  Vote: I like it +8 Vote: I do not like it

I would like to take this opportunity to ask how many people using Jngen?

»
3 months ago, # |
  Vote: I like it +33 Vote: I do not like it

Not only does it have value in terms of lowering the preparation overhead, but it would also be an educational resource.

On another note, I agree with oversolver — Jngen is worth looking into if you're doing any serious problemsetting. It is a better alternative to testlib's somewhat outdated (but still decently useful) generators.