Comentarii la http://www.infoarena.ro/blog/interviu-two-tap

http://www.infoarena.ro/blog/zece

Petru you are wrong. Think about it some more.

what? sunt 2 oameni acolo jumatate din 2 e 1. e destul de clar ca cel putin e confuz rau textu daca nu si bushit exemplu

edit: ai dreptate, am fost eu confuz

care e treaba cu inghetarea clasamentului? aduce vreun beneficiu?

in exemplul 1, eliminam pe 4, dupaia eliminam pe 3. ramanem cu 1-2 fiecare e influentat exact de jumatate dintre parlamentari, nu mai mult de jumatate. totusi voi ziceti ca rezultatul  1,2 nu e un set ok pt exemplul 1.

din nou in enunt apare "mai mult de jumatate" dar din exemplu si daca un parlamentar e influentat de exact jumatate e caz negativ. cred ca trebuie sa modificati enuntul

in enunt scrie "mai mult de jumatate din colegii lui" nu e "cel putin jumatate din colegii lui"?

http://www.infoarena.ro/blog/transpose

Nu, nu e. Raspunsul e sqrt(3)/3. Ai citit ce s-a discutat mai sus?

Solutia mea de acu multi ani era asa: planele D2A1C1 si A2B1C2 sunt paralele. Distanta intre ele e egala cu distanta intre D2A1 si A2B1. Diagonala B2D1 de lungime sqrt(3) e perpendiculara pe ambele plane si e impartita in 3 de cele doua plane. Deci distanta e sqrt(3)/3

http://www.infoarena.ro/blog/distance

http://www.infoarena.ro/blog/acm-2013-etapa-nationala-partea-ii

http://www.infoarena.ro/blog/acm-2013-etapa-nationala

Nu ai raspuns la intrebare.

Cine e n?

http://www.infoarena.ro/blog/statistics

Yes, you're right. I was meaning going through all permutations.

http://www.infoarena.ro/blog/bafta-la-oni

 The problem already mentions : "(the same element can be used multiple times)". I did that to avoid handling the annoying case you mention .

Good comment Andrew. I've added an example about path finding for GPS applications.

edit: I've just read your article. The idea of having named datasets of different sizes it's pretty smart! But for me those names don't mean too much(as they are UK centric) so other than UK and World I have to look the descriptions up to understand the table.

Yes. Lots of problems have multiple parameters (graph algorithms, knapsack, etc) which I skipped in this article. Also there are subtleties in the constant factors. For example not all n^2 algos are alike. One that iterates through a matrix vs one that does just some math operations are pretty different. Also some optimization may significantly change the average case behavior.

I was thinking it would be fun to tag more execution times, complexities, algorithms and datastructures and then play with this table:

find all algorithms with some execution time
find all algorithms for a given data structure
find all data structures for a given complexity

http://www.infoarena.ro/blog/numbers-everyone-should-know

n^2 log n dar s-a luat 100 cu n^3

@Petcu You can solve 5) better than n/2 queries

@Petcu
For 1) that's the cleanest solution I know. Good job! I like it more than Vlad's improvement over Andrei's solution.

For 2) you're wrong. I can prove you need at least min(n, m) queries to solve the problem. It's not trivial to prove and lots of people fall into the mistake of thinking some di. Can you guys see a proof?

3) You're wrong. Read Voroneanu's solution or my solution to see something that works in n + log n - 2 queries. I doubt one can do better.

4) Andrei has another 3n/2 - 2 solution I think. It's in line with my own.

@Claudiu your celebrity problem solution is the one I had in mind.

So we're left with
2) showing a worse case, and Vlad to explain his solution.
5), 6), 8 ), 9)