hacker.org

Hi,
I don't understand what the different error messages mean on this challenge. I've seen:
<sol> is incorrect
and
failed to flood correctly (or something like that).
Does exceeding the cycle limit come under 'failed to flood' or 'is incorrect'? I have code that on a sample grid I generated runs correctly in 22k cycles, which seems well within the allowed total, and I'm sure my algorithm is right...
EDIT - every single one of the fifty-odd random grids I have tried (some require extensive filling) has been correct and below 45000 cycles. :/ I'll PM adum and see what the problem might be.

laz0r wrote:I'll PM adum

While you're at it, ask him why there is such an abundance of challenges using this silly language.

I spent some minutes thinking about a funny deserved reply for you, but could not find one.
So, just a piece of advice: if you do not have something constructive to post, just don't post anything.
By the way, I do not think there is someone forcing you to solve these challenges, is there?
Finally, something is not necessarily silly just because you can't do it (it seems a 5-year-old child reasoning, in fact).

Sorry, about these words, but what were you expecting with your sarcasm?

*************************************

A not so constructive answer to laz0r:

you can send an infinity loop just to see what the answer is...
The lack of feedback of these challenges were discussed in the forums already (http://www.hacker.org/forum/viewtopic.php?t=808). It is about the subsequent challenge, I think.

Thanks portal, I am starting to worry that perhaps adum's chosen a grid that my algorithm solves really slowly (worst case performance, the empty grid, takes 130k instructions!) - the infinite loop just gives '<sol> is incorrect' so that's probably it...

@portal: What I am expecting is intelligent and interesting challenges. Aren't you too?

Okay, so I can understand that when one has made the effort of writing the implementation for the HVM, that one does not want to use it only one single time. So if there were three or four, or maybe five challenges of this type, that would be okay. But a dozen or more is simply too many. Nothing of value is added by letting there be this many.

This stack programming language is honestly silly, anyone in their right mind agrees to that. You have 32-bit memory cells, but no command to load an (arbitrary) 32-bit value to a cell? Come on. This restriction does not make anything more interesting. At all. Same for jumping to a certain address. Even adum himself in one forum post admitted that this is a pain. Why should it have to be a pain? It needs not and should not be painful, simple as that.

It would certainly be more interesting to have to write or modify code in e.g.:
-C
-C++
-VHDL and/or Verilog
-x86, or Motorola, or MICROCHIP, or Analog Devices, or 8051, or any other real assembler

But there seems to be no challenge which requires you to do that. Why?

There are some challenges that requite you to use existing (virtual)machine code and programming languages such as x86-64, JAVA, COBOL, Ada, Lisp etc. which I guess you haven't gotten to yet, even though the JAVA ones start quite early. For the existing machines and languages there are full sets of tools that exist for working on them such as disassemblers, decompilers, VMs, assemblers, compilers and linkers.

HVM lacks all but a VM implementation and making these tools - especially to cater for the limits imposed by the challenges - is certainly a challenge in itself. You aren't told to make any of these tools but it is fairly clear that you would benefit from making at least some of these after running into a few of the challenges and is just as enjoyable if not more than doing other challenges.

I think given the variety of challenges on offer and the different requirements for each of them, of which a fair few have reusable elements and others which have little to no bearing in the real world. In short I (as I assume portal does) don't see any reason to whine about it.

CodeX wrote:There are some challenges that requite you to use existing (virtual)machine code and programming languages such as x86-64, JAVA, COBOL, Ada, Lisp etc.

There is a fundamental difference here:
In HVM you are not given code to work with, you have to write the code completely from scratch. There is no challenge, as far as I know, which explicitly requires you to write a program from scratch in a specified language (other than the HVM challenges). I could be wrong but having read the descriptions of altogether 188 challenges (when there were descriptions, that is) I rather think this is true.

Oh, and because I liked somone's reply to a certain argument on another forum so much, I shall quote it here since it shows very nicely how silly the argument in case really is. Second quote is from said other forum:

portal wrote: So, just a piece of advice: if you do not have something constructive to post, just don't post anything.
By the way, I do not think there is someone forcing you to solve these challenges, is there?

This is the moronic argument, to which the quoted reply is:

Southwest, on another forum wrote:"If you don't like X, shut up and do something else instead of doing X" has never been a very impressive argument against criticism.

"The steak at Emile's Bistro, however, was slightly overdone, proving dry and a bit leathery."
"If you don't like the steak, then shut up and order something else."

"The character development was weak and left me wanting more."
"If you don't like the book, then shut up and read something else."

"Smith's acting, while spot-on at times, strayed toward hammy and unrealistic in the second half of the show."
"If you don't like the play, then shut up and watch something else."

Pure excellence.

Finally, something is not necessarily silly just because you can't do it (it seems a 5-year-old child reasoning, in fact).

I never said I couldn't do it. I have solved e.g. King Rat, strlen, Mod, Fast Mod, Haven't We Met Before. I said it is a silly language to have to write code in, and if you do have to then a handful of HVM challenges would absolutely be enough. And that is the truth.

... your argument is moot. Your opinion is that HVM is a silly language but does it really matter? It's a hard language to master and adum is giving us challenges that require this language. I don't see the point in having less challenges because frankly we need more altogether.

I doubt you voicing your silly opinions on HVM challenges or whether this forum needs a moderator are really going to help at all, it's just getting annoying seeing every one of your posts complaining about something.

AMindForeverVoyaging wrote:
There is a fundamental difference here:
In HVM you are not given code to work with, you have to write the code completely from scratch. There is no challenge, as far as I know, which explicitly requires you to write a program from scratch in a specified language (other than the HVM challenges). I could be wrong but having read the descriptions of altogether 188 challenges (when there were descriptions, that is) I rather think this is true.

Oh, and because I liked somone's reply to a certain argument on another forum so much, I shall quote it here since it shows very nicely how silly the argument in case really is. Second quote is from said other forum:


This is the moronic argument, to which the quoted reply is:

Pure excellence.

I never said I couldn't do it. I have solved e.g. King Rat, strlen, Mod, Fast Mod, Haven't We Met Before. I said it is a silly language to have to write code in, and if you do have to then a handful of HVM challenges would absolutely be enough. And that is the truth.

I've seen your profile in hacker.org and apparently you haven't written any code to solve the puzzles (Mortal Coil, Modulo, Crossflip, etc). You can see them as challenges, and you will write their code fro; the scratch. There are eight different puzzles in this site where you can write code in any language you wish (C, C++, Java, Python, etc) but it seems you have not exploited this possibility yet.

Though you can solve them in HVM too, of course. heh

DaymItzJack wrote:I don't see the point in having less challenges because frankly we need more altogether.

Agreed, but there should be a certain level of quality to them. We don't need the umpteenth HVM challenge, but better and more interesting ones. Like I said, why not something with VHDL (or Verilog, for that matter). Or digital circuitry. Or astronomy. I do have some ideas in mind. If I create a challenge, I will make sure to ask you to be a beta tester.

AMindForeverVoyaging wrote:There is a fundamental difference here:
In HVM you are not given code to work with, you have to write the code completely from scratch. There is no challenge, as far as I know, which explicitly requires you to write a program from scratch in a specified language (other than the HVM challenges). I could be wrong but having read the descriptions of altogether 188 challenges (when there were descriptions, that is) I rather think this is true.

Well, there are also the superhack challenges.
In itself, it is almost impossible to make a challenge that forces people to use a specific programming language. With just a single result to enter, there is no way to check this, and running server side just adds a lot of complexity to keep your server safe.

I must say for myself, I quite like the HVM/Superhack challenges. Certainly the last one. The graphical aspect makes it intriguing.
In itself, there also aren't that many different challenges of this type. Most of the time, you have a few main challenges, and then some king of the hill variations of these challenges. (smallest length, shortest execution time)
The main reason by looking at the challenges why there are so many HVM challenges is more a historical one. Those are usually the result of someone solving the challenge in a better way then the precious person. Since at the time when the challenges were released, the King of the hill system wasn't yet in place, an extra challenges was added with more challenging requirements.

If you want to use a lot of programming languages, you can offcourse always challenge yourself, with writing some of the programs you need for getting the solution in those languages. Also, as suggested, have a look at the puzzles on the main page.

In case you are interested: http://www.spoj.pl/
The focus is more algorithmic, however...

AMindForeverVoyaging wrote:If I create a challenge, I will make sure to ask you to be a beta tester.

http://www.hacker.org/challenge/chal.php?id=109 Make one then. If or when I get to it, I will try to solve it.

portal wrote:In case you are interested: http://www.spoj.pl/
The focus is more algorithmic, however...

Looks interesting, thank you for the link.

Is anyone still active here?

Wondering how the layout of the image is.
Is it 0-15 is first row, 16-31 is second row and so on?

I have a working solution that runs in the Hacker VM IDE, but is not accepted by the challenge.
It runs <50000 cycles, so that should not be the problem.

Anyone could help me out?