How to Build a Decade-Based Movie Retrospective

A decade evidence grid defines date membership, selection dimensions, current version availability, and acknowledged gaps before the viewing order is set.

In short: Define decade membership by one date field—usually the work's original release year—and state inclusive boundaries. Generate candidates from your authorized collection, then select through a declared evidence grid covering time within the decade, form, region, creator perspective, runtime, and current version availability. Document gaps and present the result as your retrospective, not a definitive history.

A retrospective is a curated argument about a period. Its value comes from explicit selection rules and useful contrasts, not from claiming complete representation.

Define the decade contract

Write:

• start and end years, including both boundaries;
• the date field used for membership;
• whether restorations retain the work's original year;
• project size and time span;
• current source and profile scope;
• the dimensions you want to compare.

Do not mix original release year with reissue, restoration, source-added, or local acquisition dates. Dublin Core defines date as a point or period associated with an event in a resource's lifecycle, which is why the event must be named.

Generate a broad candidate set

Filter by the declared year range from a clean baseline. Check boundary-year controls and records with ambiguous or missing dates. The category-and-year filter guide provides a method for validating the date field.

Add plausible candidates before deciding what the decade “means.” Early selection based on a familiar narrative can hide useful counterexamples.

Build the evidence grid

Choose dimensions relevant to the collection. They might include early/middle/late decade, fiction/documentary, mainstream/independent, short/feature, or contrasting production regions. Do not force equal representation when the available collection cannot support it; record the gap.

The FIAF manual's work–variant–manifestation–item structure reinforces the distinction between a movie's identity and the particular version available for viewing. A restored manifestation should not silently change the work's decade membership.

Select with declared rules

Use three questions for every candidate:

1. What does it add that no current selection provides?
2. Is the work's date and identity sufficiently verified?
3. Is a suitable current version available for the project?

Remove near-duplicates unless comparison is intentional. If two movies occupy the same role, keep the one with stronger evidence or make the contrast itself part of the retrospective.

Arrange the viewing sequence

Chronological order reveals change over time, but it is not the only valid route. You can alternate contrasting forms, group by theme, or create pairs around a shared question.

Use a sequence card:

The link sentence matters more than a decorative collection title. It tells the viewer why the order exists.

Keep the project temporary

Create a named project manifest with start date, membership rule, selected titles, and completion state. Do not turn every retrospective into permanent category structure. The temporary-theme workflow shows how to archive the project cleanly.

For a creator-focused subset, use the director retrospective plan rather than overloading the decade grid with a second dominant organizing rule.

Review coverage honestly

At the end, record:

• dimensions well represented;
• known absences in the authorized collection;
• candidates excluded because no usable version was available;
• uncertain dates or relationships;
• questions for a future project.

This coverage note prevents a personal selection from being mistaken for a survey of the entire decade.

Norva may help search, filter, group versions, and organize compatible authorized sources, but the collection reflects what the user connects. Current source availability and metadata quality limit the retrospective and should be disclosed.

Common mistakes and limitations

• Using source-added date as original release year.
• Selecting only familiar examples before generating candidates.
• Treating a restoration as a new work in the later decade.
• Hiding collection gaps behind a universal claim.
• Building permanent categories for a temporary project.
• Ignoring the exact version and required tracks.

Frequently asked questions

Must a decade run from years ending in zero through nine?

Use that convention when it serves the project, or define another ten-year window explicitly. Clarity matters more than an assumed rule.

How many movies belong in the retrospective?

Choose a size the available time and evidence can support. Each selection should have a distinct curatorial role.

Should I order strictly by year?

Not necessarily. Chronology, contrast, theme, or paired comparison can work if the sequence is explained.