SCORM 1.2 vs SCORM 2004: which version should you use?

Every Intle generation gives you two downloads — a SCORM 2004 4th Edition zip and a SCORM 1.2 zip — built from exactly the same content. The question LMS administrators ask me most is which one to actually upload. The honest answer depends less on which standard is 'better' in the abstract and more on the LMS you're uploading to and how much learner state you need to carry between sessions. Below is how the two versions genuinely differ at the data-model level, what your LMS is likely to do with each, and a rule of thumb you can apply in about a minute.

What SCORM actually is

SCORM (Shareable Content Object Reference Model) is two things bundled together: a packaging format and a runtime contract. The packaging side is a zip with an `imsmanifest.xml` at its root that tells the LMS what's inside and how to launch it. The runtime side is a small JavaScript conversation between your content and the LMS — the package finds the LMS's JavaScript API, initialises a session, then reads and writes named data-model elements (score, status, bookmark, answers) until the learner leaves. Both SCORM 1.2 (released 2001) and SCORM 2004 (now on its 4th Edition) share that shape. What changed between them is the vocabulary of that runtime conversation and the rules around navigation.

Where the versions genuinely differ

Four differences matter in practice. Everything else is detail.

• Completion vs success status. SCORM 1.2 has a single field, `cmi.core.lesson_status`, that has to mean both 'did they finish?' and 'did they pass?' at once. SCORM 2004 splits this into `cmi.completion_status` (completed / incomplete) and `cmi.success_status` (passed / failed), so a learner can finish a module without passing its assessment — and your reports can tell the two apart.
• Resume data (suspend_data). This is the scratch space your content uses to remember where a learner was and what they answered. SCORM 1.2 caps `cmi.suspend_data` at 4,096 characters — about 4KB. SCORM 2004 4th Edition raises that ceiling to 64,000 characters, roughly sixteen times the room, which matters for long modules or anything that stores per-question state.
• Interactions and objectives. Both versions can record question-level `cmi.interactions`, but 2004 formalises them and links each interaction to per-objective results with a normalised measure (`cmi.score.scaled`, −1 to 1). In 1.2, interaction data is write-only in the spec — the LMS can store it but content can't read it back — and reporting support is patchy across LMSs.
• Sequencing and navigation. This is 2004's headline addition: IMS Simple Sequencing, a rule-based model for moving a learner across multiple SCOs. SCORM 1.2 has no formal sequencing model at all.

The LMS support reality

Newer does not automatically mean better supported. LMS vendor documentation differs by version and configuration: Moodle documents stronger native SCORM 1.2 support than SCORM 2004, while Rustici Engine-based products document broader SCORM 2004 capabilities and Canvas documents its own SCORM tool behaviour. Those vendor statements are not proof that an Intle package has passed on a named platform. Check the dated evidence matrix and run an import, launch, score, completion and resume test in the exact LMS configuration.

Choose SCORM 1.2 when…

• Your LMS is Moodle. Uploading the 1.2 package keeps you on Moodle's fully-conformant path and sidesteps the unfinished 2004 sequencing entirely. (Moodle's 'Number of attempts' setting also works only with SCORM 1.2 and AICC.)
• You're dispatching content onward to an older downstream LMS that only accepts 1.2.
• Your institution enforces a strict ~4KB resume-data limit regardless of version.
• You want the most conservative, most widely-tested option and your content is short enough that 4KB of resume data is ample.

Choose SCORM 2004 4th Edition when…

• Your LMS runs the Rustici SCORM Engine — SCORM Cloud, Blackboard Learn, many commercial LMSs — or otherwise fully supports 2004.
• You need completion and pass/fail reported as separate statuses — common for compliance evidence, where 'finished but failed' must be distinguishable.
• Your module is long or richly tracked and you want the 64,000-character resume headroom so bookmarking and answer-restore survive.
• You want per-objective results and a normalised scaled score feeding mastery rules.

How Intle builds both from one generation

Both packages come from separate runtime adapters built from the same canonical content, so the learning design, questions and scoring stay aligned while the CMI data-model dialect and manifest change. The main content is delivered as a self-contained SCO and resume state is compressed in stages for the 1.2 and 2004 limits. Dated package-level evidence is published on the LMS compatibility page, separately from the honestly pending named-platform matrix. Test the exact release in your own LMS before rollout.

The one-minute rule of thumb

If you're uploading to Moodle, use the SCORM 1.2 file. If you're uploading to SCORM Cloud, Blackboard Learn, or any Rustici-powered corporate LMS, use the SCORM 2004 4th Edition file. If you're on Canvas or genuinely unsure, start with 2004 and keep 1.2 as your fallback. Whether you're building an AI-generated quiz for a Moodle rollout, a compliance course for a Rustici-powered LMS, or a full multi-section course, you generate once and pick the version your platform prefers. For exact upload steps, see the Moodle guide and the Canvas guide.