Post-Interpretive Curatorial/Museum Research — Observatory

← Research Tools and Experiments · Research

Research materials

Explore the research programme →

Read the research questions, methodology, staged validation plan and current software outputs. The worked study below explains a documentary and curatorial inquiry with inspectable source wording.

Read the worked study → · Read the validation plan →

This public page explains the working prototype and how to use it. Screenshots show the actual local interface with original schematic images and entirely constructed documentation records. Research takes place in the separately run local application.

New: guided workflowsFor museums & curatorsResearch programmeWhat it doesWorked studyHow to useFAQScreenshotsRequest accessCitation

Observatory study overview with a working question and constructed teaching-vessel documentation. — image at source

Experiment with this research tool

Museums, curators, individual researchers and institutions interested in experimenting with Observatory are welcome to contact Museum of One. On request, I can send the private GitHub repository or a copy of the tool itself. Describe your research question and intended setting so that a suitable starting point can be identified.

Request the repository or tool →

New in Observatory 2.1 · Guided workflows

A route through your inquiry.

Begin with a question. The guided workflow brings the next useful action into view, from gathering works and sources to following evidence and preparing an output.

Choose a starting route, then adapt it to the research. Keep the steps you need for object histories, curatorial decisions, exhibition conditions, encounters, documentary work or another inquiry. Open the relevant research tool from a step and return to the guide when you are ready.

Observatory guided workflow showing a sequence of research steps, actions to add a work or source, and two constructed demonstration records — image at source

Observatory Adapt this route controls showing a comparative encounter starting route and optional research modules — image at source

Start here: choose New study → Guide me step by step, or open an existing study and select Guided workflow. Use Adapt this route to choose what the inquiry needs.

Workflow notes remain private and are excluded from research outputs. Reviewing a step records your progress; it does not establish consent, permissions or the validity of a finding. These screenshots contain fictional demonstration records and no participant findings.

For museums, galleries and curators

Give curatorial judgement an evidence trail.

For teams and researchers who need to explain how an account, object record or exhibition decision was reached and retain the evidence and uncertainty behind it.

Ask a question. Bring things together. Follow the evidence. Start from the work and the situation, with room for several accounts to remain in view.

Museums and galleries

Connect collections research, documentary comparison and exhibition development. Keep object identities, source roles, rights and unresolved questions inspectable.

Curators and exhibition designers

Retain an installation’s conditions, the alternatives considered and the reasons for a decision. Revisit the account when a source, situation or attribution changes.

Learning, interpretation and research teams

Bring qualitative readings, properly consented encounter records and documentary evidence into the same inquiry without treating their different roles as interchangeable.

Bring a real curatorial question.

Start with one uncertainty or decision your team is already discussing. These are proposed uses, illustrated by the working interface and constructed records.

Reconcile an object record: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Reconcile an object record

Which description belongs to which source?

Retain an old title, a later title and their provenance without silently overwriting either.

How a museum team can use it

Retain an old title, a later title and their provenance without silently overwriting either. Compare the original identifiers, dates and wording before deciding whether a catalogue change is warranted.

Make a curatorial decision inspectable: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Make a curatorial decision inspectable

Why was this arrangement or wording chosen?

Record alternatives and responsible voices alongside the evidence.

How a museum team can use it

Record alternatives and responsible voices alongside the evidence. State what the decision can establish and what a later review would need to reconsider it.

Investigate exhibition conditions: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Investigate exhibition conditions

What was available to be encountered?

Retain plans or installation photographs with versioned spatial records.

How a museum team can use it

Retain plans or installation photographs with versioned spatial records. Record omission, sequence, sightlines and access; investigate any claim about a visitor effect through a separate study.

Follow a documentary correction: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Follow a documentary correction

What changes when the source is re-read?

Capture the exact passage and its locator.

How a museum team can use it

Capture the exact passage and its locator. Keep source wording, translation, paraphrase and researcher inference distinguishable; connect the correction to supporting and contrary evidence.

Compare accounts and readings: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Compare accounts and readings

Where do readers agree, disagree or hesitate?

Define a coding scheme appropriate to the inquiry, preserve independent reading records, and examine negative cases.

How a museum team can use it

Define a coding scheme appropriate to the inquiry, preserve independent reading records, and examine negative cases. The software does not automatically adjudicate disagreement or calculate a validated measure of meaning.

Study the afterlife of a claim: proposed museum workflow illustration — image at source

Proposed museum workflow · interface illustration

Study the afterlife of a claim

How does an account move between settings?

Record source-to-source and passage-to-passage relationships with their evidence.

How a museum team can use it

Record source-to-source and passage-to-passage relationships with their evidence. Preserve uncertainty about translation, influence and independence rather than treating every repeated phrase as a new witness.

Begin with a bounded inquiry.

1. Agree the question. Name the decision, evidence needed and limits of the study.
2. Document the materials. Retain identifiers, exact passages, source roles and permissions.
3. Compare and reconsider. Examine contrary accounts and record why a decision changes or stays open.
4. Review what travels. Prepare a permitted account and retain a separate complete private backup.

A record to bring into the meeting.

Bring the object, the competing accounts, the precise evidence and the reasons for a proposed decision. Review what remains uncertain before acting on the account.

Inspect the worked study → Explore the visual tour → Discuss research access →

Fit the tool to the inquiry.

This prototype supports research documentation and analysis. It is not a collections-management replacement, provenance certification, visitor-tracking system or automatic finding generator. Its usefulness in particular museum settings must be investigated, not assumed.

Inside the working prototype

Follow the research from question to record.

Fifteen actual application views across four stages. All records are constructed for documentation; the images depict no historical collection or real exhibition.

Bring a study together ↘Stay close to the evidence ↘Compare and reconsider ↘Take an account forward ↘

Bring a study together

Ask a working question: constructed documentation example — image at source

Bring the work into view: constructed documentation example — image at source

Record the situation: constructed documentation example — image at source

Stay close to the evidence

Stay with the source: constructed documentation example — image at source

Keep an assertion answerable: constructed documentation example — image at source

Preserve the reasons: constructed documentation example — image at source

Compare and reconsider

Trace an account carefully: constructed documentation example — image at source

Define the analytical terms: constructed documentation example — image at source

Compare readings: constructed documentation example — image at source

Retain what resists the account: constructed documentation example — image at source

Preserve chronology and uncertainty: constructed documentation example — image at source

Review an imported study: constructed documentation example — image at source

Take an account forward

Choose an account to circulate: constructed documentation example — image at source

Review the current selection: constructed documentation example — image at source

Inspect the working dossier: constructed documentation example — image at source

Actual screens explain the instrument; they are not evidence of a curatorial effect. Read image and record provenance →

What it does

• Inspect evidence. Render local PDFs and images, inspect stored text and preserve exact fragments with page or image-region anchors. No OCR is supplied.
• Object histories. Keep separate sourced assertions about titles, attribution, material, dimensions, location, ownership and exhibition history.
• Record a decision. Document alternatives, reasons, voices, disagreement, outcomes and the conditions for reconsideration.
• Trace circulation. Link accounts through typed quotation, translation, copying, challenge or possible-dependency relationships with their evidential limits.
• Space & material. Annotate plans and installation photographs; document what was selected, absent, sequenced or made accessible.
• Read & code. Use versioned definitions and named readings; retain negative cases, alternatives and missing evidence.
• Bring in a study. Review one supported encounter-study/v1 JSON record from Encounter Differential with its original source bytes, identifiers and measurement units.
• Prepare an output. Review selected records for an illustrated HTML dossier, CSV evidence table or JSON research package. Keep complete private backups separately.

In post-interpretive research, an account remains answerable to the work, its material situation, the source and the conditions of encounter. The workspace helps distinguish these records while preserving their connections. It does not prescribe one diagnostic vocabulary for every inquiry.

Worked study: two titles, one teaching vessel

Research question: How should an inventory title and a later display heading be recorded without treating either account as the object’s final meaning?

Constructed demonstration. The vessel is an original schematic; the two texts, dates and decision were written for software documentation. They are not archival evidence or participant testimony.

Procedure: retain both sources and exact fragments; record two sourced title assertions; compare their roles; write a provisional claim and decision with contrary possibilities. No participant data are needed for this documentary example.

Review and stopping rule: stop when each wording can be traced to its source and the decision’s limits are explicit. In a real study, further archival material, stakeholder accounts or an installation comparison could require revision.

How to use it

1. Ask a question
2. Bring things together
3. Follow the evidence
4. Compare accounts
5. Record the encounter conditions
6. Read, code and reconsider
7. Make the decision traceable
8. Prepare an output and keep a backup

1. Ask a question

Choose New study. Write the question you can presently ask; select a research direction and state the scope, limits and next step. Leave Guide me step by step selected to begin a guided workflow, then adapt the route as needed.

Procedure and research checks

Research design can change. Record the change and its reasons. Decide what evidence would challenge your proposed account before searching only for support.

2. Bring things together

Add a work or object, a source, an image, a document or a first observation. Preserve existing identifiers exactly and attach only material you are entitled to use.

Procedure and research checks

Add a citation, locator, dates, source independence and limits. Keep separate records where accounts conflict. The tool is not a live museum catalogue or a rights authority.

3. Follow the evidence

Open Research tools → Inspect evidence. Check the source, capture exact wording or an image region and save a fragment. Put your reading in the note field.

Procedure and research checks

PDF page anchors use file order, which may differ from printed page labels. Check extracted text against the rendered page. For scanned pages, use a checked transcription or image-region fragment.

4. Compare accounts

Use Object histories for sourced assertions, Trace circulation for relationships, and Notes & analysis for a claim with supporting and contrary evidence.

Procedure and research checks

A source relationship requires evidence. Separate a reported description from your endorsement of it, and preserve incompatible dates, titles and attributions.

5. Record the encounter conditions

Where relevant, add versioned presentations, actual encounters and a spatial record. Record participation and observation consent separately.

Procedure and research checks

A study record is not consent. Use Space & material to annotate an actual plan or installation view; do not interpret its rectangles as recorded visitor behaviour.

6. Read, code and reconsider

Define codes, attach readings to precise fragments, retain disagreement and examine contrary or missing evidence.

Procedure and research checks

Identify the reader and analytical pass. Do not overwrite an applied code definition to improve apparent agreement. Add a new version and explain the difference.

7. Make the decision traceable

Record the alternatives, choice, evidence, responsible voices, disagreement, uncertainty and any later outcome.

Procedure and research checks

State a review condition: a missing source, revised attribution, further account or practical installation check. A completed form does not validate a curatorial conclusion.

8. Prepare an output and keep a backup

Choose Prepare an output. Select the intended use, format and records, review the selection, then download. Use Workspace & backups for a complete private recovery copy.

Procedure and research checks

Reviewed circulation requires explicit recorded permissions and credits. It does not publish. A research JSON export is not a full backup; private backups include attachments and history and require protected storage.

Frequently asked questions

Is this a website I can conduct research in?

This is the public guide. The working application runs locally on your computer from its source code. Request the repository or a copy of the tool from Museum of One.

Must my study follow the demonstration?

No. Begin with your own question and select the research direction that fits it. Documentary, object-centred, exhibition, translation, encounter and mixed-method inquiries can use different parts of the workspace.

What is the quickest way to start?

Ask a question. Bring things together. Follow the evidence. A working question and one source are enough to begin; use the additional tools only when your inquiry needs them.

What files can I attach?

PDF, PNG, JPEG, WebP, TXT, CSV and JSON, up to 25 MB per file. PDFs render locally. Extracted text needs checking, and image-only pages have no automatic OCR.

Can I retain conflicting titles or attributions?

Yes. Add separate sourced assertions, retain each original identifier, and record the uncertainty or conflict. Do not merge accounts merely to make the record look complete.

How are quotations and translations handled?

A fragment retains exact source wording and an anchor. Notes, language, translation and translator are recorded separately. Check permission and credit before circulating either wording.

Does a marked plan record visitor movement?

No. A plan region is a researcher annotation. Sightlines, routes and access notes describe a documented or proposed situation; a visitor-effect claim needs its own evidence.

Does the app determine what an artwork means?

No. It helps preserve the connection between an account, its evidence and its limits. Observation, testimony, inference and curatorial interpretation must remain distinguishable.

Can I bring in Encounter Differential research?

The importer previews one supported encounter-study/v1 JSON record. It retains original identifiers and millisecond measurements. Archive wrappers, legacy sessions and unknown schema versions are not accepted; no artwork URL is fetched and no consent is inferred.

Can several people edit together?

This version is a single-user local workspace. Named reading records can document different readers, but there is no multiuser account system, live collaboration or cloud sync.

Does Reviewed for circulation publish anything?

No. It prepares a locally downloaded selection after checks against recorded permissions, credits and links. Human review of the context and any separately copied or identifying text remains necessary.

What happens when an encounter is withdrawn?

Withdrawn encounters and explicitly linked evidence derivatives are excluded from research outputs. Private records and audit history remain in complete backups. Withdrawal and archiving are not permanent erasure; review separately copied material manually.

How do I keep a recoverable copy?

Use Workspace & backups to download a complete JSON backup with attachments and history. Keep a dated copy on protected storage. Restore previews links, checksums and identifier conflicts before merging compatible records.

Has this been validated in a museum?

The public record documents a working prototype and its constructed example. It reports no completed museum pilot, participant findings or empirical validation. Facilitated usability and museum-transfer studies remain a research task.

Screenshots and demonstration provenance

The screenshots were captured from the working local Research Desk v2 interface and the Observatory 2.1 guided workflow on 21 October 2026, using separate documentation workspaces. The teaching vessel and gallery plan are original schematic drawings created for this guide. All object descriptions, source texts, coded readings, chronology and decisions shown are explicitly constructed.

No private study PDFs, scholarly correspondence, actual participant records or institutional collection images are reproduced. The display illustrates software procedures; it reports no museum finding. Select any image to inspect its full-size capture.

Scope and limits

This is a single-user local application with no cloud sync, public publishing, AI analysis or unattended collection. Local storage is not encryption or a multiuser authorization system. Backups can contain restricted material and must be stored accordingly.

Recorded rights, consent and clearance document decisions; they do not establish legal permission or ethics approval. Explicitly linked withdrawn material is excluded from research exports, but separately copied text needs manual review. Archiving and withdrawal preserve records; permanent erasure is not implemented.

File checksums and revision history help inspect integrity. They do not prove source accuracy, historical causation, interpretive correctness, empirical validity or institutional wrongdoing.

Citation and revision

Vale, Dorian. Post-Interpretive Curatorial/Museum Research — Observatory. Version 2.1, working prototype. Museum of One. Public documentation revised 21 October 2026. Canonical tool guide.

Software version and documentation revision are separate records. Cite the version and retrieval date of the copy actually used. The research programme remains in development; no persistent identifier or peer-reviewed validation is asserted here.

Research programme → · Encounter Differential → · Museum of One research programmes →

← All research tools and experiments · Request research access ↑