Comments
On goryinyich → GCJ 2014 — Round 1A, 12 years ago
0

Thanks! Interestingly, I don't see it in "Recent actions".
On Nickolas → April Fools Day Contest 2014, 12 years ago
+28

F..ck, I sent it 5 times and then gave up! On 6-th they were giving ACs! =(
0

This logic is strange. If you see empty lines of latin letters — could it be interpreted as one string? Yes. So, empty line is valid input for the problem.

Of course, you can interpret it as multiple strings, but 1) this contradicts to the problem statement 2) some way to separate strings from each other should have been given. So, such understanding of empty line doesn't seem to be correct.
-16

I estimate probability of this to be 20%. Otherwise why VK security guards quickly hid suspect inside the building? If this was not Pavel Durov — why not to tell the police who he was and give all the materials (like video from inside)?
+12

Looks like Pavel Durov has more serious problems to resolve now. And this requires a lot of money — hasn't been left for the contest =)

P.S. It was not good idea to hit police officer by car...
+13

There is a lot of idiots on Codeforces who vote agains posts without clear reason. Don't mind.
+14

Not exactly. This is usual DP by profile (not broken), so it probably'll get TLE if you don't use precalc. "By broken profile" is when you store the state of the last 2*m+1 cells, e.g. picture like (where # denotes cells for which you save their state):

------
--####
######
###---

This allows to consider only few transitions from previous states to the given state, and works much faster.
On goryinyich → Codeforces Round #102, 14 years ago
0
Done
+5
Do not listen to Tranvick =) You need to store 1 line only! - this is only 2 Mb of memory, or 250Kb at all if you use bit operations! So, the problem could be solved by DP even in Pascal!
0
Yes. Try to understand why.
+10
I found the way how to skirt this around: after you bilt a candidate for answer (some progression largest by inclusion), try to check as an answer this progression without the first element, and this progression without the last element.
0

Did you guys ACed it to explain solution here?

General idea is right, but I think many failed on, e.g., the following test:

7
1 2 3 8 4 0 -4

So, important thing: after finding the largest progression (by inclusion) for every two elements of the first three, you should check whether we get solution if delete first or last elements of the progression obtained (and, correspondingly, add them to the second progression).
On Seyaua → Codeforces Beta Round #85, 15 years ago
0

Many didn't see (at least three other persons except me), that's why I personally don't like the round.

UPD: Yeah, baby, give me more minuses, people never like truth =)
On amirali1374 → Unknown Language Round #3, 15 years ago
+4
Nice round, thanks to authors!
On goryinyich → Codeforces Beta Round #78, 15 years ago
0
I don't have neccessary rights. It will be rejudged soon.
On goryinyich → Codeforces Beta Round #78, 15 years ago
+14
This was joke for you to remember "first violet round" =)
On KrK → Timus 1035, 15 years ago
+1
You don't need any additional changes to the graph - just think what property every path in this colored graph possesses and how to minimize what you need.
On dalex → Codeforces Beta Round 76, 15 years ago
+5
Yea, you missed good problem "C" =) I VERY like well-masked backtracking =)

But concerning the rest of the contest - I totally agree with you...
0
fixed =)
On MarioYC → 5 x 5 Magic Square, 15 years ago
+3
AC (0.000 s) :D
On MarioYC → 5 x 5 Magic Square, 15 years ago
+3

Just for fun I solved the problem with the abovementioned approach (0.675 sec. for all tests). The only thing you should add to get AC is intelligent traverse order. For example, suppose you're given test like:
0 0 0 X 0
X X X 0 X
0 0 0 X 0
0 0 0 X 0
0 0 0 X 0
Of course, not to get stuck, first of all you should find the number for the cell (2,4), and only then bruteforce all the rest. Good test to test your solution is:
5
1 0 0 0 0
0 0 0 0 0
0 0 2 0 0
0 0 0 0 0
0 0 0 0 13
Possible answer is:
1 9 22 16 17 
14 24 12 4 11 
20 10 2 15 18 
7 19 8 25 6 
23 3 21 5 13 
On MarioYC → 5 x 5 Magic Square, 15 years ago
0

You can google an algo for constructing magic square of any size in O(N^2).

For this problem - since part of the square may be filled, it seems that bruteforce is the only possible way. Try to trim nodes which are definitely not solutions for your problem as early as possible.
For example, you know sum of the numbers in any row or column ("magic number") - use this information to detect whether it is necessary to continue, or it worth to stop at this branch and go back (if you completely filled a row/column and sum of the numbers in it doesn't conicide with the magic number). And don't forget about diagonal sums - this narrows search space further.
Also, for empty squares you can just precalculate the answers.
0
Didn't I write essentially the same?
0

Ok, to prevent future speculations on the topic, "not to take as seriously as others" in the letter definitely means "мы не воспринимаем их так же серьезно, как другие (компании)", but not "мы не воспринимаем их так же серьезно, как другие (категории людей)"
0
Ok, agree, changed "good coders" to "clever guys".
BTW, having clever mind is better than having skills and not having clever mind, because you (usually) can acquire necessary skills being clever, and not vice versa. Especially it is noticeable in financial industry, where investment banks like to hire PhD's, and not practicioners, because so-called "soft skills" is easy to obtain, but "hard skills" (like quant. finance, math and ...) is much harder.
That's what for Yandex and Co. should respect sport programmers, and that's why other normal compalies respect them (or, at least, demonstrate respect - there is little difference, if one think of it)
+11

Hm, sometimes it is better to be silent than speak (c)
Firstly, I thought it's bad joke - such letter from Yandex seems improbable. To me it sounds not just as not having much respect for the one of the most clever people from several countries, but boldly and arrogantly.
If that's real - Yandex just showed its true face and, as Coder noticed, answered the question whether clever guys should take them seriously...
On simp1eton → Page not loading Properly?, 15 years ago
0
Hm, is there any difference in which environment (eng/rus) I post my comments?