Transcribe

Workflow

1. Determine the range of pages to transcribe from the arguments
2. Check which files already exist in pages_txt/ to avoid re-doing work
3. Read the page images from pages_760w/ (read up to 5 at a time in parallel)
4. Transcribe each page faithfully into the format above
5. Write each transcription to pages_txt/b30535827_NNNN.txt
6. After writing all files, spot-check 1–2 pages by re-reading the image alongside the transcription and fix any errors
7. Report how many pages were transcribed and any issues encountered

Tips

• The images are 760px wide JPGs — readable but sometimes marginal notes or small text can be tricky. When uncertain about a word, do your best and flag it with [?].
• The catch-word at the bottom-right of each page (a single word previewing the next page) is a printing artifact. Include it on its own line at the end of the file, NOT joined to the preceding sentence. The next page's transcription should still start with that same word as part of the running text. Example:

  ...since she sends us
  
  into
  

  Then the next page starts: into the World Cloathed... Note the blank line before the catch-word, consistent with the spacing between all other elements in the file.
• Running headers alternate between the two speakers' names on verso/recto pages.
• The printed page number appears in the header area, not the filename number.