[HN Gopher] Structuring Competency-Based Courses Through Skill T...
___________________________________________________________________
Structuring Competency-Based Courses Through Skill Trees
Author : PaulHoule
Score : 40 points
Date : 2025-05-05 13:34 UTC (4 days ago)
(HTM) web link (arxiv.org)
(TXT) w3m dump (arxiv.org)
| didgeoridoo wrote:
| Exercism.org takes this approach; I'm not sure if it's super
| pedagogically-driven or more about making it fun and keeping a
| sense of progression (maybe it can be both?)
|
| I wonder which fields of knowledge are resistant to being
| structured like this from a learning perspective.
| Garlef wrote:
| From the requirements - or rather desired properties.
|
| > Subskills cover skills
|
| Isn't this requirement a common fallacy in education?
|
| First, just because you know skill A and skill B does not mean
| you can automatically do both A and B together. (Of course it
| helps!)
|
| Assume you're learning a to play piece of music. If you learn a
| given sequence [A, B] and start with individually learning the
| parts A and B you'd still have to practice the transition.
|
| Second, just because you know a composite skill does not mean you
| can do each part individually.
|
| Again assume you've learned to play a part of a piece of music
| and the notes constituting this part can be decomposed as [A, B].
| Then you'd not directly know how to play B on it's own because
| you've propably learnt to enter B transitioning from A.
|
| Still: I think the overvall approach is interesting. Just the
| details might be a bit off.
| schmidtleonard wrote:
| Sure but don't let the perfect be the enemy of the good. "Big
| bucket of courses, bring your own plan" and "here's your mass
| produced course plan that may or may not be optimized for
| departmental needs rather than yours" are both pretty rough. A
| skill tree is a step up even if it needs a little (or a lot) of
| refinement.
| Swizec wrote:
| > > Subskills cover skills > Isn't this requirement a common
| fallacy in education?
|
| Whole careers have been built on being one of only a few people
| in the world who can perform A and B together. Like that
| physics guy who can also write good sci-fi.
|
| There are 3 skills actually. A, B, and A u B. That intersection
| is often the hardest.
| queuebert wrote:
| > Like that physics guy who can also write good sci-fi.
|
| Stephen Baxter? Or someone else?
| Swizec wrote:
| I was thinking Andy Weir but yeah there's probably several.
| A strong physics background surely helps with writing
| better sci-fi
| antonvs wrote:
| Alastair Reynolds is who I thought of. PhD in astrophysics,
| worked at ESA. But there are several other scifi authors
| with decent physics resumes.
| chrisweekly wrote:
| Yeah, I've heard this referred to as "skill-stacking".
|
| If you're in the top 10-20% at two different things, you
| might be in the top 1% of people who can do both.
| criddell wrote:
| Is your skill tree too sparse? Instead of [A,
| B]
|
| maybe it should be [A, A->B, B]
| yorwba wrote:
| Yes, the whole point of having the "Subskills cover skills"
| requirement is that if this requirement is not fulfilled, the
| skill tree is missing a subskill and needs to be expanded.
| The authors state that in the process of building a skill
| tree for their database course, they noticed that SQL
| triggers were introduced without discussing SQL update
| queries first. Probably because they hadn't tried to
| explicitly list all requirements at such a fine-grained level
| before.
| roenxi wrote:
| This idea is probably more interesting if applied to software
| engineering _management_. There is a bit of a competency crisis
| among software managers where it isn 't possible to tell whether
| they are any good at what they are allegedly doing.
|
| The kneejerk reaction of most businesses seems to be either
| metrics that don't really work or relying on hypercompetent
| individuals in the higher parts of the management tree. Neither
| approach works particularly well. But I haven't seen a skill tree
| style competency training & testing approach and I suspect it'd
| work better than the apparent status quo. At least on paper the
| managers have been trained to do something instead of just
| dumping smart people into the role and hoping that they figure it
| out.
|
| For example someone can get certified in Agile and that is ... in
| some sense a bit silly because then they just spend the rest of
| their life noticing that most people don't use Agile. But a
| skill-based approach to the components of Agile development and
| training them to raise hackles in particular situations where
| people make common bad architectural decisions would cause
| productive cultural shifts. I suppose the sort of training that
| gets done doesn't go far enough.
| PaulHoule wrote:
| If people decomposed agile into little bits they probably
| wouldn't be doing agile anymore. The other day somebody posted
| a question about "how do I push back about management going to
| one week sprints?" My answer, which I didn't post then, was "if
| management really wants to go fast _give up on sprints_ "
|
| That is, Kanban + Continuous Integration >> Scrum by a lot.
| Seen from that viewpoint sprints are not something that speeds
| anything up (seductive name there!) but rather a bunch of phony
| deadlines and needless meanings (you _really_ think people feel
| psychologically safe in a retrospective meeting that was
| scheduled just to have a meeting? is there _really_ something
| worth talking about in an every two week one-on-one in your
| manager which isn 't important enough to knock on their door
| and ask about _right now?_ )
|
| If I was evaluating a manager I'd probably get them to give me
| a list of practices that they say their team is following and
| then check to for conformance against that. I am less bothered
| with do they do code reviews or not but rather "did they tell
| me that they enforce five conventions in the code and looking
| at the code I find they rarely do... Lets sit in on a code
| review"
| accrual wrote:
| This makes a great deal of sense to me. If it's demonstratably
| effective I'd love to see higher education incorporate it.
|
| As a side note, I'm thinking about pursuing an MSc at Georgia
| Tech OMSC. I wonder if the online nature of it takes into account
| this kind of skill tree, or if it's more traditional.
| Jtsummers wrote:
| OMSCS is composed of multiple courses with no hard prerequisite
| requirements (except for one course, and you can skip it if you
| can persuade the professor). Within each course, they're
| conventional. If the material lends itself to this form (not
| all graduate courses do) then they may be in a similar form to
| this just like any other college course that has a clear
| progression of "You need to know X before you can produce/study
| Y".
| accrual wrote:
| Thanks so much for the insight!
| ecshafer wrote:
| OMSCS is basically a normal university structure, except for
| requiring you to do some foundational courses first (which
| essentially act as a competency check for the rest of the
| program).
| johnecheck wrote:
| Modeling mastery as a set of skills to be acquired with
| prerequisites and requirements is clearly valuable. Visualization
| can serve as a map that guides learners.
|
| Universities employ this on a larger scale - degrees require
| classes that require passing grades and prerequisites. I like the
| idea of using it for smaller-scale learning like lessons,
| classes, and job training.
| whiteborb wrote:
| a "Hard Problem" with this approach is that it needs definable,
| enumerable "skills." real-world skills often involve flexibly
| combining tools and tactics in new ways. but not impossible to
| overcome (LLMs might be uniquely suited to help there actually -
| coming up with new scenarios etc.)
| mym1990 wrote:
| I would argue the opposite, that LLMs are not good at creating
| novel scenarios which span multiple domains, or sub-domains.
| That is where creativity plays an important part. Of course an
| LLM will spit out something that might look like a viable way
| to combine tools and skills, whether it is correct is highly
| questionable.
| johnsutor wrote:
| It looks pretty akin to mathacademy.com, which is fantastic for
| brushing up on math concepts.
| rahimnathwani wrote:
| The PDF cites Justin's book about Math Academy.
|
| I like how the paper lays out clearly the steps to create a skill
| tree, and some general guidelines.
|
| The paper is perfect for use as part of an LLM prompt, e.g.
| Using the approach detailed in the attached paper, create a skill
| tree for writing a magical realism novela. Turn this into
| mermaid code. Don't use parentheses in labels.
|
| https://mermaid.live/edit#pako:eNqNWH9v2zgS_SqEFvufW8SWm228w...
| pyinstallwoes wrote:
| Reminds me of SWG skill tree pre CU nerf
| cultofmetatron wrote:
| looks like mathacademy's approach which I've found to be
| extremely effective. I wish more moocs used this model of
| teaching.
| constantcrying wrote:
| >Computer science education has seen two important trends. One
| has been a shift from raw theory towards skills: competency-based
| teaching.
|
| There is no greater skill in computer science, science or
| mathematics than a deep understanding of theory. It is the
| understanding of how distinct things fit together, nothing could
| be as helpful. Yes, even when developing a JS web app.
___________________________________________________________________
(page generated 2025-05-09 23:02 UTC)