...
Three states: published, draft, unpublished changes
State must appear clearly on content in Libraries
Library searches from within the Library include content with all three states - in other words, all Library content.
State must appear clearly on search results in Libraries.
Option to refine search results by state - need approach to reduce confusion; possible solution to see published view from the unpublished state
Authors have three two options to publish content:
At the component levelAt the collection level’
At the library level
From within courses:
...
Name change from Randomized Content Block to Problem Bank
When course authors wish to randomize a pool of content, they select the Problem Bank block
Ways the content within a Problem Bank is determined:
Collections. Authors can choose a collection to use as the source of content.
When authors choose a collection, they can either add all the problems in that collection to the randomization pool, or they can refine by tag queries.1. All the content in a collection via a bulk add
2. Choose individual problems
3. Use a query to get a subset of problems from the collection
Tag queries. Authors can search within a Library with tag queries in order to populate the Problem Bank block. “Give me Give me all the multiple-choice problems that are tagged with “accounting” AND “easy”.
[REMOVE] Option to randomize or not randomize the set
Extend randomization to be able to include any problem type. Don’t limit randomization to only one type of problem. “I just want to create a big bucket of problems of all different types that I can filter by tag, and then use that refined selection as the pool of content to randomize.”
...
Authors have two options for syncing courses with Library updates:
Manually review and accept new updates
Course-level setting to auto-update every time changes are made published in the Library
Version history but no alerts
...
Default setting is that users cannot edit Library content in a course
Authors have the option to override this setting
Authors who do override the setting are given clear notifications when they make local changes to Library content
Notes - 3/13/24
Questions on Collections
What are the expectations of Collections size/number?
1:1 v1 mapping use case is the initial
Small pools of randomized contents
Early adopters like WGU would want larger sets of content, not small randomized content pools.
If they have a set of questions, better to keep that as a bounded thing.
Feedback from Colin: Us prescribing the difference between a problem bank and a collection is overreach. Workflow for problem bank vs. collection seemed like a step too far.
Can we get rid of the option to “not randomize”?
So you can pick components one at a time (add from library → no problem bank involved)
Or bulk select or all select
To avoid the “collections as units” use case.
This is okay for MVP.
Problem Banks as distinct concepts in a Library, what makes them distinct from Collections:
Teaches a specific learning objective, probably tied to an exam
May have ordered relationships - ie, Bank A is a pre-req to Bank B
Settings that apply to all components, default values (could be overridden in a course)
a/b testing, data on effectiveness
What would these ^ look like as specs/requireemnts:
Title captures the learning objective
In the Problem bank block, there are limitations about the types of collections they can pull over - limited to “small” collections
requirement of a limited amount of content in a problem bank (100?) [can we get this data from coursegraph?]
Concept of learning objectives as a first class thing in Libraries
Decisions:
Will go with generalized collections as the only method for creating subsets in Libraries
Still open questions:
Are collections versioned? Logs?
What’s the user story for needing to track changes to collections?