If your don't agree with this blog,have a better way dealing with this kind of problems,or think i need to improve my blog skills,just tell me at the comments
Sorry for my poor English skills:(
I cant bear more.I have to write a blog about it.
Eaxmples
1557C-calculating C(a,b)in log(n)'s time
What's wrong on these problems
I started programming on cf because my friends told me that cf problems need coders's thinking more than coding.
But now things started to change.
These problem can just stop coders from solving them because they don't know a certain algorithm,even if they found the correct way of solving them.
What's more,most of these problems' difficulties are on the data sculptures.It's not the purpose of giving problems.
I wonder how many blue or less coders know the and can use algorithms like C(a,b) or segment_tree well.
Even worse
We use our own computer and our own Internet access when coding.When we don't know how to use/type a certain algorithm,most of us thought of searching it online at once.
normal coders want to solve the problem before the contest ends,Experienced coders want to solve them faster.All of them will choose to copy-paste when facing a data-sculpture which need lots of time coding.
So here comes copy-paste forces.Guys who had written a similar algorithm before can debug faster and solve faster,who hadn't will have to find one on the Internet and took more time.
Problems become useless when falling into this hole.
My suggestions and thoughts
These problems start to exist because Div2 problems are normally given by Div1 coders.They can use these algorithms well because they are Div1 coders,but most of us can't.
I hope that these kind of problems are minimized and large data sculptures will only exist on problems for experienced cpers which need more thinking and calculating,not just make some small changes on the original data sculptures.
Why readers are downvoting?Tell me to help me increase my blog writing skills or let me know how you deal with these problems better than me.
Excuse me, how is 1557C - Moamen and XOR not a problem which need coders's thinking more than coding?
It might be my problem.
I used it as an exampe of calculating C(a,b) in log(n)'s time
I think most cpers like me can't write a C(a,b) correctly without the help of internet.
I don't think you need $$$C(a,b)$$$. Can't it simply be solved by adding various exponent values?
Like $$$(2^{n-1} + 1)$$$ possibilities for each column and if you have $$$k$$$ columns then $$${(2^{n-1} + 1)}^k$$$. This if for odd $$$n$$$. For even you need to handle "all bits in a column are $$$1$$$" separately I guess.
And you don't even need to know fast exponentiation for that. Just run a loop at the beginning and precalculate the powers.
What in the world is a Data Sculpture?
People complain for constructive, implementation, maths, datastructure, adhoc... What left?