Page 1 of 2

Please stop this rejudgement rate

Posted: Sat Dec 13, 2003 7:26 pm
by PdR
You are rejudging problems too fast!

Are we suposed to solve new ones, or fixing the old ones forever?
Also improving the test case just because someone has 0.00 seems
just too strange for me.
Please impose some limit, 8 for month should be adequate.

Yes!!! Stop it.

Posted: Sat Dec 13, 2003 9:26 pm
by razibcse
Yes, I completely agree with PdR. What's the rejudgement hell is this? Last 3 days, 3 of my solved problems are excluded after rejudging. I solved them long ago. Is it reasonable to rejudge the solutions so late? What we are supposed to do? Thinking about new problems or trying to understand what's going on in a program solved years ago?

Just make the judge bearable to all. It's certainly not going to help anyone improve his levels.

Posted: Sat Dec 13, 2003 10:43 pm
by Adrian Kuegel
I think it is ok to increase the input file of a problem and do a rejudge, but in my opinion it doesn't make sense to change the problem requirements (like done in 147), because then everyone has to do the same problem again.

Posted: Sun Dec 14, 2003 8:25 am
by sohel
Instead of correcting the old problems they should set new similar problems with the corrected data.
In this way we don't have to do the same problem again.

Posted: Sun Dec 14, 2003 9:53 am
by Andrey Mokhov
I agree to you all!

It is so strange to solve a new problem and to see that you have less problems solved than before... :roll: :roll:

Yesterday that was the case. I solved problem after problem and couldn't reach the day before yesterday's number of solved problems - they seem to rejudge problems faster than I solve them :evil:

Of course rejudgements are necessary to be sure that solutions are correct. But to change the problem statement!!! Where is the logic? I solved the problem where N was 5000. I didn't want to solve the problem where N is 30000. What is wrong in my solution? Judges are free to make new problem where they will raise N up to any limit they want. By the way when N=30000 (I talk about problem 147) the problem still can be solved very easily. Any sense to go from 5000 to 30000? I just changed size of my array and got AC. If they want to make a difficult problem let N be, say, 1 000 000 000. And make a new problem, say, 147 - horrible remake.


Best regards,
Andrey.

Rejudge

Posted: Sun Dec 14, 2003 8:24 pm
by Md. Sajjad Hossain
i am also tired of this...............it seems that i have to remember every solution and logic behind them throughout my life :(
recently the rejudgement of 147 vexed me too much :(
it wasn't very urgent i think...........
especially changing the problem statement is very disturbing :(

Posted: Mon Dec 15, 2003 9:09 am
by shamim
What is even bothering is, while the rejudgement occurs at a rapid rate, the OJ is not bothered to send a mail to inform, at least it is not doing it as soon as it should.

All i am seeing is my AC number decreasing yet i do not know which one has been removed.

Rejudgment should occur once in a while, and should not become a batch process as it is now.

Posted: Mon Dec 15, 2003 12:06 pm
by CDiMa
shamim wrote:What is even bothering is, while the rejudgement occurs at a rapid rate, the OJ is not bothered to send a mail to inform, at least it is not doing it as soon as it should.
I agree on the subject that the OJ should mail contestants if a rejudgement breaks an already AC program.
I also agree with Adrian Kuegel that the statement of a problem shoudn't change, while bigger or corrected sample input should be rejudged. An incorrept program shoudn't be AC only because it works for a limited/particolar input.

On the other hand I don't agree on the complaints about the difficulties of correcting old problems. This is a programming contest and good programming is also about well documented code. In this case programming in clear stile with intelligent comments gets a reward for the time and effort spent...

This not meaning that my submissions score well regarding documentation ;)

Ciao!!!

Claudio

Posted: Mon Dec 15, 2003 12:15 pm
by anupam
I am not in a mode to reply. But I have something to say..

I have seen many to find the solutions for all the cases in the input, just storing them in a array, then just print them according to the input. Is it what the problemsetters wanted??

They wanted that, all the solvers should solve the problem with a real and good algorithm. In that case, they don't mind if any1's solution time is less than that of the judge. They just wanted some sincere solutions.

About 147, as I was the person to modify it and found me guilty in your court, similar problems can occur. Search for one of my posts about 147. I didn't learn DP then and so solved the problem with a BF search and listed all the solutions for all test cases possible and got ac then. If you wish I can put that solutions. It ran 0.00 s and even faster than that of the judge. I want to solve more problems then. So I was a little bit happy. But now i found that these type of solutions had no real values.

The submit page can upload solution upto 40 kB. So, it was possible to submit such a dumb solutions for the problems(just like 147). But, now the limit has been increased so, now no one can't (even if he precalculates) send the solution(precalculated) as it will be over 40 kB now.

Hope you, the best programmers can easily understand. I have also my statistics down. But I don't mind if i have a real solution instead of five dumb(precalculated) solutions.

--
Have a nice day and of course night.
Anupam

Posted: Mon Dec 15, 2003 12:42 pm
by Observer
I understand what you mean, anupam, but what's most disturbing is that the problem statement actually changed without any notice. It was not until I visited this board that I found out what makes my original code fail. It is definitely unfair to people who do not visit this board at all!

Anyway, just posting a message in the "Fixing Mistake" section of the board is obviously not enough when you want to tell everyone that the problem requirement has changed. They should at least put a remark next to problem name in the Index page, right?

P.S. Yes, I do hate pre-calc. submissions sometimes (as in Q10575, where everyone can find the test cases online), and I really don't mind rejudgement; actually I like the rejudgement of Q652 Eight very much! It's the change in problem statement that's annoying. :wink: It is not great seeing some easy tasks such as Q357 now have their acceptance rates low as 1.7% :-?

hmm

Posted: Mon Dec 15, 2003 3:57 pm
by shahriar_manzoor
Changing the range is not good I guess. For example I have lost many of my previous codes so do I have to write again the solutions that has been rejected for changing problem statement (Not very time consuming but why should I?). I think there should be some guide lines on changing the data. A fraud can always cheat as many solutions are available on the WEB. I hope some common sense will prevail in future. There are problems of coins with large value in the problemset already. So 147 needed no change I think.

Posted: Mon Dec 15, 2003 5:51 pm
by Tahseen Mohammad
Sometimes you just have to let the people decide
what they are going to do be honest || cheat.

There are over 1200 problems to solve now. How many
of them can you send pre-calc solution?

Sometimes smaller data set do causes a problem for me when
I want to compare my algorithms complexity with others
because I'm not sure whether the faster solutions are pre-calc
or they just use very efficient algorithms.

There may be cheaters in this site but I don't really worry about
them much. May be they are up in stats but I can proudly say that I'm honest and being honest has improved my knowledge and me as a person.

As I said earlier those who cheats will always find means to cheat
and those who won't will not cheat even if theres no protection.

Let us all understand "Honesty is the best POLICY"

Posted: Tue Dec 16, 2003 3:26 am
by yahoo
As some people can never be honest why we shouldn't we try to make them honest. Precalculation is a very bad thing to solve a problem and i also solve a lot of VOJ problem by this technique. But this is not at all a good idea. This should be strictly prohibited.
As VOJ site is one of the best programming practice( to my opinion is the best) site it should also increase the creativity of the problem. So data set can be increased and no one should mind it.
But program statement changing if judeges think a good idea why should we stop them? They are not changing all the problem statement. Only few of the problems. :-? :-? :-?

It's all my opinion. No one should take it to heart. :wink: :wink: :wink:

Posted: Tue Dec 16, 2003 3:27 am
by Observer
Well, with the informative Internet and the powerful search engines, you can always cheat if you want to...

Thus I agree with other board members that "instead of correcting the old problems they should set new similar problems with the corrected data. In this way we don't have to do the same problem again." Even if they want to get rid of 0.00.000 s run times, "it doesn't make sense to change the problem requirements" without putting up a proper notice. But "I think it is ok to increase the input file of a problem and do a rejudge."

P.S. At the past I cheated too :P but soon I found out that it's meaningless to just send out stupid pre-calc. solutions... :)

Posted: Tue Dec 16, 2003 11:49 am
by CDiMa
yahoo wrote:As some people can never be honest why we shouldn't we try to make them honest.
Everyone can be honest! :)
I believe that you cannot force someone to be honest. You can only be honest and possibly others will appreciate your honesty and choose to be honest themselves.
yahoo wrote: Precalculation is a very bad thing to solve a problem and i also solve a lot of VOJ problem by this technique. But this is not at all a good idea. This should be strictly prohibited.

Precalculation isn't a bad thing. In real world programming it's often used. In many cases even doing precalculation right needs some effort. The problem statement sets a specification. You must write a program that meets the specification. If a precalculated solution fulfills the specifications then it is a valid solution. Otherwise one could also question that brute force algorithms or some other tecnique shouldn't be allowed... I think that if a program does its job within time and memory limits it's a good one :D .
yahoo wrote:But program statement changing if judeges think a good idea why should we stop them?
I think that a problem statement should be changed only if it's wrong or unclear :( and not to change specifications.

Ciao!!!

Claudio