The new version looks good to me. I don't know if I would have gotten the 6-letter word, but I think I could have correlated the first and last in this form. Maybe I could get the 6-letter after getting the L from the last word... Hard to say now that I know what the substitutions are from brute forcing the original :)

Riven (sequel to Myst) used a similar puzzle in one of its Islands. However, clues around the islands allowed to narrow down the permutations. For example, you would know that one character would be A, and another be X, so even if you didn't have all of the information, based on odds and deduction, it was much easier to get to the end result.
If you intend on using this, I recommend using another aspect of your game to "seed" these clues as rewards.

Hmm... Perhaps. This puzzle is based on one that I encountered in The Fool's Errand--although that one had the player swapping letters of the English alphabet, rather than using a second set of glyphs. I didn't recall that game providing clues, but looking at the website it seems that I may have misremembered.

Nevertheless, others here seem to be managing at least part of the puzzle, which suggests to me that it's not impossible to solve these puzzles without external hints, especially if I build up from simple puzzles to more complex, allowing players to start to learn the glyphs.

I have taken note of your advice, however, and may well look into implementing it.

Yeah, in a general sense a "check" is a second piece of evidence for what letter is in a square. So in a crossword puzzle, it's that letter also being in the crossing word, in a substitution cipher it's the letter occurring in another word.

Ah, I see--thank you for clarifying! ^_^

The new version works pretty well, I think.

Ah, excellent! ^_^

The new version looks good to me. I don't know if I would have gotten the 6-letter word, but I think I could have correlated the first and last in this form. Maybe I could get the 6-letter after getting the L from the last word... Hard to say now that I know what the substitutions are from brute forcing the original :)

Excellent!

I'm glad to read that I may be on the right track here. ^_^

You say that the six-letter word might be a cause of trouble--but not the longer last word?

Looking at the last word, I do note that the first three letters will likely come from other words, and are perhaps somewhat telling, and that will likely have at least one letter on top of that, while the second word gets fewer letters from other words, and is perhaps less obvious.

I might look for a replacement for that six-letter word, then--something that benefits more from the preceding words, or perhaps the last.

Thank you for your help, everyone! ^_^

I haven't actually tried solving it.

The in-game version highlights not only the selected slot, but all others that have matching glyphs, and fills them in automatically when a letter is entered. To what degree--if at all--would that make the puzzle easier, do you think?

It's the first thing I would need. Especially interactive hoovering and highlighting makes it a lot more visible where and how often each glyph is used.

However, being a programmer, I am biased towards expecting such tasks to be performed by computers :P

That also holds for the timings you get here. I don't know what you're public is, but the above people are all programmers used to do sleuthing based on minimal amounts of information, and they know how to turn and twist to get the best results with as least effort as possible, by exploiting computer tricks.
There is nothing wrong with that, but your public may not know these things.

Actually disregard my post, if the game makes it clear that this is a substitution cipher, it should be pretty straightforward (for those that know how to solve them).

The in-game version highlights not only the selected slot, but all others that have matching glyphs, and fills them in automatically when a letter is entered. To what degree--if at all--would that make the puzzle easier, do you think?

It's the first thing I would need. Especially interactive hoovering and highlighting makes it a lot more visible where and how often each glyph is used.

Indeed. Hum... I'm tempted to put out a small test-program with three of these puzzles and ask for responses to that...

That also holds for the timings you get here. I don't know what you're public is, but the above people are all programmers used to do sleuthing based on minimal amounts of information, and they know how to turn and twist to get the best results with as least effort as possible, by exploiting computer tricks.
There is nothing wrong with that, but your public may not know these things.

True, and that was something that I considered, I believe. I did try asking on Twitter, but had no replies at the time. :/

Hum... I could perhaps ask on one other, non-gamedev forum, however... It's worth thinking about, perhaps.

Actually disregard my post, if the game makes it clear that this is a substitution cipher, it should be pretty straightforward (for those that know how to solve them).

Hmm... I had thought that the instructions (a rough draft of which should be in the version of the puzzle that I posted first) might convey that, noting that each glyph corresponds to a letter. Perhaps my final wording should be a little more explicit?

I don't believe this to be the issue. I think the puzzle is clear about this fact, but to someone that has never solved one of these before, it may be a steep learning curve.

I had never heard of this type of cipher before (as weird as that may sound) thus my inability to efficiently understand what was expected of me to solve it. However, I assume the game is puzzle-oriented, and as such, the expected target audience would likely have some experience solving these types of puzzles (I'm geared towards mechanical puzzles for the most part, hence the disconnect).

As for the instructions themselves, from the text, I was led to believe it was a scrambled cipher, hence my inability to understand how letters would map out to each of the characters if their position wasn't fixed.

Perhaps start with a puzzle where some glyphs are given (either in the puzzle or in some other way)?

Biggest problem is perhaps that people expect the same letter to go with the same glyph every time.

I don't believe this to be the issue. I think the puzzle is clear about this fact, but to someone that has never solved one of these before, it may be a steep learning curve.

*nods* Fair enough--I should be careful about my first few puzzles then, I think.

I had never heard of this type of cipher before (as weird as that may sound) thus my inability to efficiently understand what was expected of me to solve it. However, I assume the game is puzzle-oriented, and as such, the expected target audience would likely have some experience solving these types of puzzles (I'm geared towards mechanical puzzles for the most part, hence the disconnect).

Well, the game is somewhat of a mix, with traversal, exploration, puzzle-solving, and sparse combat. There are puzzles (at this point the player has solved a simple button-pushing-and-inventory puzzle and had access to an optional interpretation puzzle), but they've also clambered over some stone blocks and likely faced a large magical construct in combat.

As for the instructions themselves, from the text, I was led to believe it was a scrambled cipher, hence my inability to understand how letters would map out to each of the characters if their position wasn't fixed.

Oh wow, that's unfortunate! I want to revisit my wording anyway, so I'll hopefully keep that potential misunderstanding in mind when I do.

Perhaps start with a puzzle where some glyphs are given (either in the puzzle or in some other way)?

Not a bad idea--thank you. ^_^

Biggest problem is perhaps that people expect the same letter to go with the same glyph every time.

... But, the same letter does go with the same glyph every time...

I think if you want a puzzle like this to get solved by an average player, you're going to have to teach us how to do it, maybe with a few much simpler decryption puzzles. I hear there are books on decryption problems like this that explain tips for how to do it, but it seems incredibly difficult when you've never done one.

You could help us out with clues about the sentence, or maybe a word that's already known (allowing you to fill in some of the letters throughout the problem, giving you a starting point to solve the rest). It might not take many of these before we can do a completely blank one.

I have an interactive version now! It should, hopefully, better demonstrate how the puzzle is intended to work in the game. As suggested just above, I've also added a simple, optional hint system that fills in a few chosen letters.

You should find the new version here.

(The instructions in the main screen mention tweeting to or emailing me, as it was intended for linking to on Twitter; just respond here instead, please. ^^; )

I hear there are books on decryption problems like this that explain tips for how to do it ...

Are you suggesting that I find such a book as a resource for what hints to give to players, or how to teach them the puzzle?

You could help us out with clues about the sentence, or maybe a word that's already known (allowing you to fill in some of the letters throughout the problem, giving you a starting point to solve the rest).

Hmm... I'm now considering implementing a difficulty setting for these "minigame puzzles" (as opposed to puzzles integrated into the levels); in this case the "easy" or "normal" level might include hints as in the interactive test-program mentioned just above. I already have such a setting for combat, so it would make some sense, I think.