About new incoming features

[fpnc]

I have seen that 3 of the 6 changes have are already take place.

Yes, #1, #2 and #3 are done. #6 does not depends on me at first, and #5 seems easy for everyone but not for me (what means that if I'm able to, I'll do it, and if not, another of my teammates will ) #4 is relatively easy to be done but first I need to know what information should I provide (because our complete events database can grow up to ~2GB in XML format and not everything in that database is usefull).

I would like that you let me know when are you going to upgrade the compliers. Besides i would like to know exactly which compilers are you installing.

When? I don't know exactly. Not before September, 16 I think, but I cannot give more exact schedule. I don't know either if we're going to wait a few weeks until version 3.x has been found to be really stable. What versions? Our first thought was upgrading to gcc 3.2. BTW, we'll post to this board and also announce it in the problemset page, with new versions and new compile flags if changed.

About the Java Support i also would like to know which classes and packages are supported with the new compiler that were not supported with the old one.

Now in Java is the time to use THE STANDART java provided by SUN freely. I think that if you take another implementation NOT STANDART like gcj the problems that are happening now is going to be repeated in the future.

And we know. We're using it now because:

• It compiles to native code, so timing is more accurate than the VM of Sun Java.
• It uses almost the same compiler options that gpc, g++ and gcc, so it was very easy to implement in our system. This should be a problem no more, as FreePascal uses a compilation scheme quite different and it has been implemented.
• Rankings now includes all submissions in all languages. We're thought in dividing them into languages, because if we put javac the timing is going to be quite different, I think.
• We don't know for sure how well we will be able to measure the memory, the runtime errors...
• And more things I don't remember right now

Personally I would like to use javac 'cause it is the "standard", but right now I cannot promise we will.

Best regards,

[scottaugust]

First, I would like to thank all of you at Valladolid for hosting this site and making it the success that it is!

Second, I do have my own opinions about what I would like this site to be, however I have to realize that I am not hosting the site - I am merely using the site. The way the maintainers choose to update/change their site is all up to them. I am happy that they have set up a board to handle the discussions like this to give us a chance to request upgrades/changes/improvements that we would like to see.

Third, I also have my own opinions about what the priorities of upgrades should be, however - read second statement

Fourth, To all of you at Valladolid that participate in the maintenance of this site, do what you want, make it look like you want it too, make it support what you want it to support, make it work like you want it to work, and make the changes you want to make in the order that you want to make them!

I have to say that I have been offended/annoyed by some of the inconsiderate people out there using this site, however, it is my own responsibility to either continue using the site and ignore the inconsiderateness of some of the participants or discontinue using the site if the offense/annoyance is too great. Valladolid is not responsible in that aspect!

Enough said, carry on with the site!

Scott E August

[ithamar]

Hi guys.

I dont want to bother you and i kow that you have a lot of work to do. But there are things that i think you have to considerate.

* It compiles to native code, so timing is more accurate than the VM of Sun Java.
* It uses almost the same compiler options that gpc, g++ and gcc, so it was very easy to implement in our system. This should be a problem no more, as FreePascal uses a compilation scheme quite different and it has been implemented.
* Rankings now includes all submissions in all languages. We're thought in dividing them into languages, because if we put javac the timing is going to be quite different, I think.
* We don't know for sure how well we will be able to measure the memory, the runtime errors...
* And more things I don't remember right now

* Right now theres a big difference between Java and other languages. Today I sent the problem 10304 with pascal and with Java and i get times of 2 Secs with Pascal an 11 sec with Java. And this code is optimized without optimization the java program give me TL and pascal Program runs in 16 Sec. Maybe freepascal is too good or maybe gcj is too bad but right now there are a difference that gcj does not avoid.

* About runtime errors i dont think that gcj handles this well. I have submitted a lot of programs that have to give run time error. and never i got that response. so this is not a pro for gcj.

I hope that keep this in mind when you take the decission about the compiler.


PS: I'm really sorry if i am bothering you with my suggestions but if this is happening tell me and i wont bother you again. I just want to let you know the things that happen to a standart user of your system (althougt in the 0.5 % Java minority

[fpnc]

...but right now there are a difference that gcj does not avoid.
About runtime errors i dont think that gcj handles this well. I have submitted a lot of programs that have to give run time error. and never i got that response. so this is not a pro for gcj.

I hope that keep this in mind when you take the decission about the compiler.

We'll do. Really I've never written a program in Java (well, only one for Distributed Systems in my University, but it was so... arggggggg!!! ) so I haven't tried our support for this language. So I'll believe what you're telling me is correct, and I'll vote for javac when the moment's come. But, there are some other problems with javac (technical ones) that I don't know if we'll be able to fix them easily... I'll let you know when we tried.

PS: I'm really sorry if i am bothering you with my suggestions but if this is happening tell me and i wont bother you again. I just want to let you know the things that happen to a standart user of your system (althougt in the 0.5 % Java minority

No problem. I want to listen to your opinions. And as scottaugust said: tell me what you want, that we'll do what we want (it's a better sentence in Spanish, anyway) Seriously, we want to take note of your opinnions - if not, this thread would not exists - but we cannot make everyone happy. Sorry.

[Ivor]

Thank you for running such a good site. All the changes are usually done for the best, so I hope everything will improve in time. There's one thing I wouldn't like to see:
don't subtitute names with ID codes -- that would make the ranklists ugly.

But all in all changes are good. And making a separate ranklist for different languages would be a neat idea, because rarely can pascal programs beat c/c++... I don't even want to mention java.

However, I want to remind you that timing still has glitches. I got 'nan' time again today. And that in my opinion overweights all of the few changes you've had this far. It makes me sad that the problem hasn't been fixed for more than two weeks now...

There are few people who care about precise timing and optimization to the max. And I'm one of them... But I don't want a credit I do not desrve.

I still hope the system will become stable and all fake times will be rejudged.

Carry on,
Ivor

[fpnc]

Timing is like a nightmare. I think something gone wrong when the last system crash, and we don't know what!! I think NaN never appeared before, and it is not a constant reply, so it is very difficult to track it.

Please, give me all the information you can about this... in what languages it happens, if there were more submissions in the moment yours were judged, if your program is doing something suspicious, if you submitted it via email or submit-o-matic... everything can be important to find out what's happening. Also I'd like to know what % of submissions are "affected" with this sympthom. Please tell me also the exact place in which you see NaN (if email, what is the exact reply) and what is shown in ranklists and status... everything!

Best regards,

[Ivor]

I'll post it in the other thread.

Ivor

[SMBfromRU]

Hi to ALL!

I'm an old Pascal user and fan ("old" also means age), and hardly can change this language to anything else. My version of Pascal is BP and so far I had very less problems with BP-GPC compatibility. But when I started to submit programs, I had received many "Compile Error" and WA, until learned GPC documentation.

Now I ask: where can I get full Free Pascal Manual or something like this?
I suspect that not only Strings declaration changed against GPC version.

Thanks in advance.

[Shahid]

really it's robotic to see the judge ID's in the online status ....rather the names are really very much lively.....
whatever i don't find the 'T' option to watch the names in the online status.....

and i think it's better to show the names in the status normally and give the option 'T' to them who r not interested to see the names..

[Adrian Kuegel]

you can see the names using this adress:
http://acm.uva.es/cgi-bin/OnlineJudge?S ... d:IDSCMTLN
default is now an 'A' instead of 'T'.

[karl]

I used options T and A, very funny


But, by the way:
In my opinion removing the algorithm field can't be decided to be "only a good" or "only a bad" idea. I agree to the admin that most written there isn't helpfull in any way. I misused it too (telling something 'bout my feelings on my wife )...
Thinking about this... Let me suggest to create an additional option so that algorithm field is shown or not (like options T/A).


Have a nice day...

Karl

[judge]

Now I ask: where can I get full Free Pascal Manual or something like this?
I suspect that not only Strings declaration changed against GPC version.

http://www.freepascal.org/moreinfo.html
http://www.freepascal.org/docs-html/user/user.html

[ithamar]

What happen with the compiler upgrade?

Nobody talks about that in the board anymore. The upgrade is going to take place or just was a idea that can't be implemented.

[fpnc]

It will be implemented. But I cannot tell you when. Sorry.

[Fresh]

Right now everything will be hold as is - Real Time System Status shows numbers unless you use "T" option, and the other ranklists will show the name if they show it now. It may be changed in the future, however.

How?

-novice