DacteloBy the Dactelo team at Taktek.
Counting the ruling archive before scaling the work
On 27 October 2026, Dactelo added scripts to count the ruling cache and draw a reproducible page sample. The cache index listed 44,804 documents. Of those, 44,766 had scanned PDFs, containing 274,936 page records. These are counts from the project’s cached archive, rather than a fresh census of the source website.
The inventory separates missing files, page totals and OCR-confidence bands. Low OCR confidence does not establish that a page is handwritten. Poor scans, unusual print and layout can also lower the score. A handwriting estimate requires inspecting a sample with an explicit labelling rule.
This changed the planning task: processing time and storage could be estimated against a defined page count. The first handwriting sample still needs human checking before supporting an external prevalence claim. The inventory therefore establishes the scale of the cached material, while leaving handwriting prevalence as a separate measurement.