I recently posed a question about attempting to ruin Lepton's attempt to compress a JPEG. I've done other file-format-type questions before, to generate a WAV, defeat LZMA compression, etc, in part because I like to learn how file formats are put together at the byte level.

The question is all about file formats and compression methods, both of which are inherently computational, and rather than try to do some frankenscore of the program size and the compression size with some arcane mixing rule, a la the LZMA question, I just made the scoring all about the JPEG. Maybe if I phrased it "you need to generate the JPEG with code", it would be "about programming", but really, I don't care about saving bytes with esolang XYZ, and there are plenty of other challenges for that.

The notion that it's just about "finding the right data" I find to be a bit of a red herring. Every answer to a "true programming problem" is about finding the right string of code to be run, and using a totally naive approach (i.e. attempting to run random strings to get the correct output) in most languages would have orders of magnitude better luck than trying to do the same for this "restricted-source" question.

The core of programming is one of specifications. Computers do specific things as dictated by an input and rules as to how the input is handled. Programmers and hackers need to understand how things work in order to (mis)use them. To that end, I suppose if the specifications of the site are defined so narrowly, I could jump through them, but that kinda sucks the fun out of it.

  • \$\begingroup\$ Related \$\endgroup\$ – user45941 Jul 15 '16 at 5:30
  • \$\begingroup\$ Also related \$\endgroup\$ – James Jul 15 '16 at 5:32
  • \$\begingroup\$ @DrGreenEggsandIronMan that seems to bolster my argument; and with this there isn't even the ambiguity problems they discuss there. \$\endgroup\$ – Nick T Jul 15 '16 at 5:35
  • \$\begingroup\$ Also very related (thanks Dr Green Eggs and Ham for linking it in chat) \$\endgroup\$ – user45941 Jul 15 '16 at 6:11
  • \$\begingroup\$ Nick, as the person who wrote the answer which @DrGreenEggsandIronMan linked to, I don't see how it bolsters your argument. To be honest, I'm not sure it's particularly relevant to this discussion at all. \$\endgroup\$ – Peter Taylor Jul 15 '16 at 9:38
  • \$\begingroup\$ @PeterTaylor I was referring more to the question as a whole: both answers suggest that the challenge (to craft a file of a specific format) would be on-topic. And while it might be a poor question (i.e. downvote) for other reasons (PDF contents are compressed using some algorithm, ambiguous as to what counts) that's a separate matter versus being on-topic altogether. \$\endgroup\$ – Nick T Jul 15 '16 at 15:33
  • \$\begingroup\$ inherently, not intimately :) \$\endgroup\$ – cat Jul 18 '16 at 15:59

We should allow computational challenges of all sorts

tl;dr: This site is about puzzles based on programming, not puzzles whose solutions are programs. This challenge qualifies.

We currently accept a variety of challenges that are not based directly on code. For instance, we accept algorithm-only challenges. The accepted answer for why we should accept such challenges is that algorithms are part of the programming process. The exact same argument applies here.

Programming doesn't end when the code is written. Testing code is an essential part of the programming process, and finding inputs to a program that have interesting behavior is a key part of testing. In this challenge, we are asked to find an input to a program which has unusual behavior for that program. Sure, the submission isn't a piece of code, but that doesn't mean it's unrelated from programming entirely.

In addition, this is not a challenge that can realistically be solved without writing computer programs. It's possible that someone will come up with a good solution by hand, but the best solutions are going to be found by a computer program, with very high probability.

Moreover, imagine the the challenge was slightly modified from "Find a JPEG that Lepton can't compress" to "Write some code that outputs a JPEG that Lepton can't compress", where the scoring metric, in either case, was the compression ratio of the resultant JPEG. Would that change anything about the merits of the question? Of course not.

Let's not let an arbitrary rule like "Solutions must be code" hold us back from good challenges such as the one in question.

  • 1
    \$\begingroup\$ If the optimal JPEG was likely to be found by hand, would you consider the challenge on-topic? \$\endgroup\$ – Nathan Merrill Jul 20 '16 at 17:29
  • 2
    \$\begingroup\$ @NathanMerrill If it was likely to be found by hand, but still required a computer to test, it'd be closer to borderline. If it was likely to be found by hand and was easy to test by hand, I'd ship it to Puzzling. \$\endgroup\$ – isaacg Jul 20 '16 at 23:49
  • 1
    \$\begingroup\$ I need to stop bothering with the site...a fixed-output Print 1-10 racks up the votes. Dull. Uninspired. \$\endgroup\$ – Nick T Jul 22 '16 at 4:42
  • 1
    \$\begingroup\$ @NickT Honestly I think that happens on a lot of sites. Hard work in question writing is rarely rewarded as much as the question being "accessible". \$\endgroup\$ – Liam Jul 24 '16 at 3:23
  • \$\begingroup\$ I think that regardless of a whether or not your challenge is on-topic as is, it could be nominally changed to an on-topic challenge. To me this makes it seem that it should be on-topic in the first place. Otherwise we would just be making you jump through hoops to pose what is essentially the same challenge. \$\endgroup\$ – Liam Jul 24 '16 at 3:26
  • 2
    \$\begingroup\$ @NickT that question is sitting on +28/-9. I wouldn't say that's particularly well received, especially when you consider most of the downvotes were probably placed by established folks. \$\endgroup\$ – Blue Jul 25 '16 at 12:38

[R]ather than try to do some frankenscore of the program size and the compression size with some arcane mixing rule, a la the LZMA question, I just made the scoring all about the JPEG. Maybe if I phrased it "you need to generate the JPEG with code", it would be "about programming", but really, I don't care about saving bytes with esolang XYZ, and there are plenty of other challenges for that.

Your scoring method is not what jeopardizes the "ontopicness" of your question. We have a lot of challenges that aren't scored by the byte count or some other metric that isn't related to the functionality of the submitted programs. Scoring a program by how well the image it generated can be compressed is perfectly fine.

So why was your question closed in the first place?

If you used some code to carefully craft the JPEG, please post it along with the image.

That implies that we do not actually have to generate the image with a program. I can take a picture with my phone, scan one of my daughter's drawings, use some photo manipulation software to create a random image, browse the Internet for noise images, etc., all of which are completely unrelated to programming.

Granted, it's unlikely that the first two approaches would result in a competitive entry; I'm actually not convinced that any approach would stand a chance against pseudo-randomly selecting the pixels with a uniform distribution. The latter two might stand a chance though, and submissions that consist of nothing more than a swirl of pixels are neither related to programming nor interesting.

So, regardless of what the community ends up deciding, why not circumvent the whole problem and make submitting the code that generated the image mandatory? That can't possibly affect the challenge in a negative way, and nobody will be able to argue that it's off-topic.

  • 1
    \$\begingroup\$ Opponents have argued that having a fixed-output program is invalid because it's not about programming but finding the right data, e.g. the program could just have a binary literal in it that it dumps... \$\endgroup\$ – Nick T Jul 25 '16 at 2:02
  • 2
    \$\begingroup\$ If you actually demand code that generated the image, I would consider such a program a violation of this loophole. \$\endgroup\$ – Dennis Jul 25 '16 at 2:06
  • 2
    \$\begingroup\$ @Dennis Consider also the first/top comment on that loophole: "Often, this vulnerability is an indicator of a dumb question..." Though I would use "poor" rather than "dumb", the point remains - if a challenge is vulnerable to hardcoding, then there's a very good chance that the issue lies in the challenge. This is one of those scenarios. \$\endgroup\$ – user45941 Jul 25 '16 at 5:30

The difference between your LZMA challenge and this challenge is, the LZMA challenge actually required writing a program as a submission. This challenge merely requires outputting a constant output - there's no programming involved in generating the output (aside from probably writing a program to find a valid output).

All solutions to challenges must be code submissions (note that a pseudocode explanation of an algorithm can count as "code" for challenges). If the only point of a program is to output a constant output[1], then it's not much of a programming challenge, and thus it's off topic for this site.

Simply put, this is a site about programming. If there is no programming involved in a challenge, it is off-topic.

[1] challenges are a bit different, but solutions to these type of challenges typically involve programmatically generating the output by exploiting some structure/repetition in the desired output, rather than simply doing print("large string here").

  • 2
    \$\begingroup\$ On the linked question, the answer that explicitly says "you must include code" primarily argues "Without working code, there is no way to test the answer to verify that it works, to determine its score, etc." and other points that devolve from it being untestable/theoretical. This patently isn't. \$\endgroup\$ – Nick T Jul 15 '16 at 5:49
  • \$\begingroup\$ @NickT Look at the accepted answer - it gets to the crux of the problem. Designing an algorithm is (part of) programming. Coming up with a specific, constant sequence of bytes to output is not programming. \$\endgroup\$ – user45941 Jul 15 '16 at 5:51
  • 1
    \$\begingroup\$ No, instead the question requires understanding of two other algorithms in order to implement a good answer. \$\endgroup\$ – Nick T Jul 15 '16 at 5:57
  • 1
    \$\begingroup\$ Understanding one or more algorithms is not the same as designing and implementing an algorithm to solve a problem. There is no algorithm to design and implement - it's literally just "find a sequence of bytes that works". Had you made the challenge "given input data, modify it as little as possible so that LZMA compression fails" or something similar, then it would be on topic. \$\endgroup\$ – user45941 Jul 15 '16 at 5:59
  • \$\begingroup\$ Slight brain slip in my previous comment: substitute "Lepton" for "LZMA". \$\endgroup\$ – user45941 Jul 15 '16 at 6:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .