I’ve got a little sample of a similar idea in my mind, which I think may be similar enough it’s worth mentioning. I’ll describe it using “is” language instead of “sounds like it is” language below for brevity.
A sazen is a description of a completed process which is not a description of how to start the process; an island of self-supporting concepts which you cannot jump directly to from pre-existing concepts.
Picture a resource-management game (like Dwarf Fortress, Factorio, etc). You might need Resource A to get Resource B and Resource B, like needing electricity to refine uranium and uranium to run a power plant. Saying “You need 2 power plants to every 5 uranium mines” is not useful to a beginner, because they can build all 7 buildings and nothing will happen.
Instead, they will build iron mines and copper mines and coal mines and steam engines and … all until they build those buildings, at which point they will see that 2:5 is the correct ratio.
Someone has already noted that the concept of sazen is related to that of shibboleths, but my above definition suggests that sazen is related to any kind of verify-only knowledge, like a cryptographic hash or any zero-knowledge proof.
Sazen are for verification, not teaching, and I hope explicit labeling of them as such is enough to get the benefits without the costs.
Also koans. You give a pattern which will be significant when understood, but insignificant until then. The student tries random interpretations until the pattern is understood, and know that the random data they have found is valuable. This is done because the data is difficult to describe directly, but easy to hold once achieved.
The problem being, there can often be multiple things which “click” with the pattern!
Or I guess “problem” might be too strong/overstating. Like, if you get value out of the koan then you got value out of the koan, regardless of whether it’s the same value the koan-speaker hoped you would get.
But it’s a problem from a communication standpoint.
I wonder if koans work best under partial supervision. Instead of the master having to check each attempt, they check 1 in 100 attempts, allowing them to teach roughly 100 times as many students at once.
If any teachers out there use koan-likes, do they work well for homework?
I’ve got a little sample of a similar idea in my mind, which I think may be similar enough it’s worth mentioning. I’ll describe it using “is” language instead of “sounds like it is” language below for brevity.
A sazen is a description of a completed process which is not a description of how to start the process; an island of self-supporting concepts which you cannot jump directly to from pre-existing concepts.
Picture a resource-management game (like Dwarf Fortress, Factorio, etc). You might need Resource A to get Resource B and Resource B, like needing electricity to refine uranium and uranium to run a power plant. Saying “You need 2 power plants to every 5 uranium mines” is not useful to a beginner, because they can build all 7 buildings and nothing will happen. Instead, they will build iron mines and copper mines and coal mines and steam engines and … all until they build those buildings, at which point they will see that 2:5 is the correct ratio.
Someone has already noted that the concept of sazen is related to that of shibboleths, but my above definition suggests that sazen is related to any kind of verify-only knowledge, like a cryptographic hash or any zero-knowledge proof.
Sazen are for verification, not teaching, and I hope explicit labeling of them as such is enough to get the benefits without the costs.
Strong upvote and strong agreement vote for the crystallization “Sazen are for verification, not teaching.” Thank you.
Also koans. You give a pattern which will be significant when understood, but insignificant until then. The student tries random interpretations until the pattern is understood, and know that the random data they have found is valuable. This is done because the data is difficult to describe directly, but easy to hold once achieved.
The problem being, there can often be multiple things which “click” with the pattern!
Or I guess “problem” might be too strong/overstating. Like, if you get value out of the koan then you got value out of the koan, regardless of whether it’s the same value the koan-speaker hoped you would get.
But it’s a problem from a communication standpoint.
Yes, it’s an interesting issue.
I wonder if koans work best under partial supervision. Instead of the master having to check each attempt, they check 1 in 100 attempts, allowing them to teach roughly 100 times as many students at once.
If any teachers out there use koan-likes, do they work well for homework?