Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
@@ -0,0 +1,89 @@
.. _openedx-learning-adr-0004:

4. Pathways: Split Between Catalog and Content
==============================================

Status
------

Draft

Context
-------

Courses in ``openedx-core`` already separate the *catalog* side (``openedx_catalog``: ``CatalogCourse``,
``CourseRun`` - what learners browse and enroll against) from the *content* side (``openedx_content`` - what is
authored, versioned, and published). Pathways have the same two aspects, and the same reasons to keep them apart:

- **Different change rates.** The display name, description shown in the catalog, and SEO metadata are revised
frequently and casually. The definition of what a learner must do to complete the Pathway changes rarely and
deliberately.
- **Different people doing the editing.** Catalog data is typically maintained by marketing or communications staff;
the Pathway definition is maintained by content authors. Both are visible to learners, so the distinction is about
who edits what, not about who can see it.
- **Different permissions follow from that.** We expect instances to want to let marketing staff update catalog
copy without granting them the ability to change what learners must complete, and vice versa. Keeping the two
apart makes that possible without inventing field-level permissions.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Keeping the two apart makes that possible without inventing field-level permissions.

In our new RBAC world, I would expect that there's not much difference between implementing "field-level" permissions and "model-level" permissions, since we're using Casbin to define whatever roles and permissions make sense for each API. Either way, it's simply permissions and needs to be enforced in the REST API via permission_classes as appropriate.

However, I still think this split makes sense for the other reasons you've stated.

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@bradenmacdonald, ah, that's good to know - I'm not very familiar with the recent RBAC changes. Should we remove this point from the ADR?

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd say we should add a note that this reason may not be as important with the new RBAC system. But as I said, I still think the overall decision makes sense and is a nice way to structure it.

- **Auditability.** Progress and credentials must be judged against the definition that was in effect at the time,
which requires versioning the definition - but versioning catalog copy would be pure overhead.

Decisions
---------

1. A Pathway is split into two parts:

- **Catalog Pathway** - the learner-browsable, enrollable thing. It includes the display name, the description

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you say more about this? It sounds like a credential will tie together a learner, a catalog pathway, and particular content pathway that "implemented" that catalog pathway.

Is that right? Will an enrollment do the same?

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kdmccormick, I added more details about the enrollments in 9218009. I will add the details about the credentials to #764, since it's out of scope for the current ADR (and I wanted to avoid forward refs here).

In short:

  1. Enrollment binds User and Catalog Pathway, because a learner enrolls in a pathway, not its content - i.e., any active authoring (content) changes are visible to the learner. Therefore, I believe there is no real reason to track, at the DB level, the exact state in which the learner was enrolled.
  2. Credential is different - it binds User and a specific version of the pathway content, because a credential is a claim that a particular definition of requirements (that was active at that point in time) was met.

shown in the catalog, SEO metadata, and a **Category**. It is **not versioned**.

- **Pathway content** - the definition of the Pathway: its Items and its completion criteria. The content is
**versioned**, so that we can always tell what the definition was at any given moment. A version of the Pathway
content *implements* a Catalog Pathway.

2. The **Category** is a student-facing label for the kind of Pathway (e.g. "Master's Degree", "Annual Training").
Learners see the Category rather than the word "Pathway". It is always required: rather than falling back to
"Pathway" in code, we ship a default database entry with that name, so the behavior is uniform and operators can
rename or extend the set without a code change.

3. In authoring contexts (Studio, Django admin, code, docs), the terminology is always "Pathway", with the Category
shown explicitly. Relabelling is a learner-facing concern of the catalog side only.

4. **Dependency direction**: ``openedx_content`` knows about ``openedx_catalog``, never the reverse. This has the
following consequences:

- Pathway Items may reference ``CourseRun`` entities directly.
- The link from a Catalog Pathway to the Pathway content that implements it lives on the content side.
- Anything that has to tie the two sides together belongs in ``openedx_content``, or in something downstream of
it, but never in ``openedx_catalog``.

5. **Enrollment** ties a learner to a Catalog Pathway. Progress is evaluated against the currently published
content version, not against a version frozen at enrollment time, so that authoring changes reach learners who
are already enrolled.

Example content of each model:

============================ ===================================
Catalog Pathway Pathway content
============================ ===================================
Display name Pathway Items
Category Completion criteria
Description References to CourseRuns
SEO metadata Link to the related Catalog Pathway
Enrollment
============================ ===================================

.. Run `dot -Tsvg images/pathway-catalog-content.dot > images/pathway-catalog-content.svg` to regenerate the diagram
after making changes to `images/pathway-catalog-content.dot`.

.. image:: images/pathway-catalog-content.svg
:alt: Catalog Pathway vs versioned Pathway content
:width: 100%

Consequences
------------

- Catalog edits never create new content versions; definition edits (Items, criteria) always do.
- Because evaluation follows the published version rather than the enrollment-time version, edits to a Pathway apply
to learners who are already enrolled, which is what we want, but it means edits need care and re-evaluation.
- The unversioned Catalog Pathway can be long-lived even if its content definition is changed significantly over time.
- The dependency direction means a Catalog Pathway cannot, on its own, tell which content implements it. Queries in
that direction start from the content side.
40 changes: 40 additions & 0 deletions docs/openedx_learning/decisions/images/pathway-catalog-content.dot
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
digraph pathway_catalog_content {
rankdir=LR;
fontname="Helvetica";
compound=true;
node [shape=box, style=rounded, fontname="Helvetica", fontsize=11];
edge [fontname="Helvetica", fontsize=10];

subgraph cluster_catalog {
label="openedx_catalog (not versioned)";
fontsize=10;
fontcolor="#4d4d4d";
style=dashed;
color="#b3b3b3";
catalog [label="Catalog Pathway\n\nname, category,\ndescription, SEO"];
courserun [label="CourseRun"];
}

subgraph cluster_content {
label="openedx_content (versioned)";
fontsize=10;
fontcolor="#4d4d4d";
style=dashed;
color="#b3b3b3";
v1 [label="content v1\n\nItems, criteria"];
v2 [label="content v2\n\nItems, criteria\n(currently published)"];
v1 -> v2 [style=dotted, color="#808080", arrowsize=0.7, label="revision"];
}

user [label="User", shape=ellipse];
enrollment [label="Enrollment"];

user -> enrollment [label="learner"];
enrollment -> catalog [label="enrolled in"];
enrollment -> v2 [label="progress evaluated\nagainst the currently\npublished version", style=dashed, color="#808080", fontcolor="#4d4d4d"];

v2 -> catalog [label="implements"];
v2 -> courserun [label="Items reference"];

direction [label="dependency direction:\nopenedx_content knows about\nopenedx_catalog, never the reverse", shape=plaintext, fontsize=10, fontcolor="#4d4d4d"];
}
115 changes: 115 additions & 0 deletions docs/openedx_learning/decisions/images/pathway-catalog-content.svg
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.