CoRE-Math picture archive workflow

This is the website-facing operational guide to selecting and preparing photographs. The permanent archive, original files, permissions evidence and activity records remain in Dropbox; this repository contains only the web derivatives and documentation needed to publish them consistently. It does not replace the wider CoRE-Math communication or image-policy material held in Dropbox.

The archive is stored in Dropbox at:

/CoRE-Math/06. Communication, Dissemination, and Visibility/Pictures/

It is deliberately simple: browse pictures in Dropbox, preserve originals, embed standard metadata in working masters, and keep web copies separate. The archive is not a website asset folder and does not replace the website repository.

Folder structure

Pictures/
  00-Inbox/
  _unidentified/
  _external-reference-only/
  YYYY/
    YYYY-MM-DD-activity-name/
      activity.json
      originals/
      masters/       # only when metadata/editing is required
      web/           # only when a derivative is required
      rights/
  archive-index.csv

Use a date only when it is known. If only month and year are known, use YYYY-MM-activity-name; do not invent a day. IPMS pictures remains outside this workflow until separately reviewed.

Adding new photographs

1. Ask the contributor to send the original file, not a screenshot, social-media download or web export. They should supply: activity, date and location; photographer; copyright holder; requested credit; source; and explicit permission/restrictions for web publication.
2. Place received files in 00-Inbox/ while checking the information. Do not alter the received file.
3. Create the year/activity folder and move the unchanged received file to originals/. Assign the next permanent identifier, for example CM-PH-000028.
4. Create activity.json from the template below. Store common activity information there once rather than repeating it in each image.
5. If metadata or a web image is required, copy the original to masters/. Add IPTC/XMP fields to the copy only. For an immutable received file that must remain untouched, store its identifiers and provenance in a same-name .xmp sidecar. Leave unknown fields blank or use unknown in the activity record; never infer authorship, copyright or permission.
6. Create a file in web/ only after permission for publication has been recorded. Copy that derivative, not the original, to the GitHub repository. Add an accurate caption, credit and alt text to the page/post.

Metadata and rights

The master must use standard IPTC/XMP fields where known:

• XMP-xmp:Identifier: permanent ID, such as CM-PH-000028.
• XMP-dc:Title and XMP-dc:Description: title and factual caption.
• EXIF:DateTimeOriginal / XMP-xmp:CreateDate: actual capture time only.
• IPTC Core location/city/country fields.
• XMP-dc:Creator, XMP-dc:Rights, XMP-photoshop:Credit and XMP-photoshop:Source.
• XMP-xmpRights:UsageTerms and XMP-xmpRights:WebStatement for documented terms.
• XMP-dc:Subject for controlled keywords, for example Institution|Makerere University and Programme|Industrial Mathematics.

Use the following rights-status values in activity.json: unknown, permission-requested, approved-web, restricted, CC/licensed, or public-domain. A webpage that happens to display a picture is not permission evidence.

Activity record template

Save this as activity.json in each activity folder. Keep unknown values as null or "unknown".

{
  "activity_id": "CM-ACT-2026-0001",
  "title": "Activity title",
  "date": "2026-10-02",
  "source_post_date": null,
  "location": "Institution, city, country",
  "institutions": ["Institution name"],
  "networks": ["CoRE-Math"],
  "programmes": [],
  "projects": [],
  "source_urls": [],
  "rights_status": "unknown",
  "rights_evidence": null,
  "notes": ""
}

Use date only for a verified activity date or date range. When a historical image is associated only with a dated website post, leave date as null and record that post date separately in source_post_date.

Browsing and catalogue

Browse visually in Dropbox by year and activity. For searching by title, person, institution, programme, rights status or source, open archive-index.csv in Numbers, Excel, LibreOffice or any text editor.

The catalogue is generated, never manually edited. Once ExifTool is installed and the Dropbox folder is synced locally, run:

python3 scripts/generate_photo_archive_catalog.py \
  "$HOME/Library/CloudStorage/Dropbox/CoRE-Math/06. Communication, Dissemination, and Visibility/Pictures"

The generator reads embedded metadata (and same-name .xmp sidecars when an original must remain untouched) plus activity.json records, then rewrites archive-index.csv. It does not modify photographs. Its rights_status column comes from the activity record; XMP usage terms are retained separately.

Preparing web copies

1. Check rights_status and the evidence before creating a derivative.
2. Work from masters/, preserving the original aspect ratio and retaining the original in originals/.
3. Create a reasonably sized JPEG or AVIF in web/; name it from the permanent identifier, for example CM-PH-000028-card-960.avif.
4. Record the GitHub destination in the activity record notes, then use the derivative in the website. Do not publish originals or rights correspondence.

One-time requirement

ExifTool is required for embedded IPTC/XMP writing and catalogue generation. Install it locally on the archive-maintenance Mac with Homebrew:

brew install exiftool

Then test it on a copied master, never on an original. The repository script is intentionally dependency-free apart from ExifTool and Python 3, which macOS provides.

If ExifTool is installed outside the shell PATH, set EXIFTOOL to its executable path when generating the catalogue.