eGGSA Gravestone Photograph Library, the Karoo band: 464 burial grounds across the Western, Northern and Eastern Cape, caption index
gravestone photographs (index) · consulted 2026-08-12
- Type
- gravestone photographs (index)
- Repository
- eGGSA / Genealogical Society of South Africa, graveswcape.eggsa.org, gravesnortherncape.eggsa.org, graveseastcape.eggsa.org
- Reference
- 38 in-band collections of 402; 464 cemetery albums, 425 distinct Google Earth Cemetery Initiative GPSIDs, 1,869 albums including surname sortations and per-grave sub-albums, 20,949 captioned photographs. Harvested 12 Aug 2026.
- Consulted
- 2026-08-12
The citation as it was given
THE SOURCE OF grave_photos.csv, one row per photograph. WHAT WAS TAKEN: the caption exactly as eGGSA wrote it, the filename, the photograph's URL and the album it sits in. NO PHOTOGRAPH WAS DOWNLOADED and none is republished here - the position taken on SRC-090 and unchanged: eGGSA state that no one may replicate their material to add to another website database. Names and dates read off a caption are facts and are recorded; the images are linked. HOW THE BAND WAS DRAWN: Willowmore - Beaufort West - Prince Albert - Laingsburg - Merweville - Fraserburg - Calvinia - Williston - Brandvlei, roughly lat -30.3 to -33.6, lon 19.1 to 23.9, as an EXPLICIT LIST of collection slugs rather than a keyword match. A keyword match was tried first and was wrong in both directions at once: it admitted FORT-BEAUFORT in the Eastern Cape frontier on 'beaufort', and it missed Rietbron, which sits between Willowmore and Beaufort West and whose name matches nothing. MOST OF THIS BAND IS OUTSIDE THIS PROJECT'S DISTRICT and is not claimed to be in it - 266 of the 464 grounds are Northern Cape. The value is the reverse lookup: is this surname on a stone anywhere in the Karoo. WHAT IS NOT COMPLETE, said plainly: on a cemetery sorted into surname albums the letter pages were taken, and below a letter a per-grave album was taken only where its title carries a surname this project holds . 340 albums were passed over and every one is listed with its link in data/_harvest/eggsa-skipped.jsonl. So an absence in this index is not evidence a person is not buried there - and eGGSA say the same of their own albums. GPS is DEGREES AND DECIMAL MINUTES on every album page: -32 6.062, 23 0.630 is -(32 + 6.062/60), not -32.6062. Read the obvious way it throws the Karoo about a hundred kilometres out. 384 grounds carry a reading, 54 have a GPSID and no GPS line, 38 have neither; the last two classes are recorded null and never guessed. CAPTIONS ARE NOT PARSED. `MERWE VD` is Van der Merwe and `nee`/`voorheen` mark successive married names, and a caption is stored as written so that the parse stays a later, reviewable step. The one change made is that `&` is decoded to `&`, which is the page's encoding of the caption rather than anything on a stone. The staged working files, the walk statistics and the skipped list are in data/_harvest/; that directory is a finding aid and is not part of the model.
Kept verbatim. This project contradicts published genealogies and corrects a family's own oral tradition, so every claim has to be checkable by someone who was not in the room.
Nothing in the model cites this source yet.