UW–Madison Libraries Services during COVID–19

COVID–19 Library Dashboard
Library by Appointment: Access to Memorial Library now available

Transcription Guidelines

The primary purpose of EFacs transcription is to help users search for content that is relevant to their interests. A secondary benefit of such text is that users can cut and paste (and perhaps correct) content for re-use more easily than completely retyping. The electronic text may have accessibility or other benefits as well. Given that purpose, attempt to make your transcription as faithful as possible to the original, but note the following:

  • Transcribe the text in reading order. If the natural reading order is unclear, do your best.
  • Where possible, include line breaks that mirror the line breaks of the original (This will make it easier for a user to match one to the other). However, recombine words that were broken across two lines. (Likewise: if the original includes a blank line between paragraphs, mirror that in the transcript with a blank line.)
  • Because the primary purpose of the transcription is for searching, the best way to handle non-standard or abbreviated forms is to normalize them to the modern, complete form. This may (somewhat) reduce the utility for cutting and pasting, so if desired, both forms could be included with the provided form appearing in square brackets [ ]
    • E.g. A problem is a chanse [chance] for U [you] to do your best. (Attr. [Attributed] to Duke Ellingtion)
  • Searching will not be affected by capitalization, but one could choose to modernize the capitalization for the fully-normalized transcript, or retain the original capitalization for the original plus [normalized] version

Saving files

The transcriptions should be created and saved as:

  • Plain text files. No .doc or .rtf, nothing fancy. You’ll want to work in a plain text editor such as Notepad (Windows) or TextEdit (Mac)
  • One text file per image, named to matched the filename of the page image it represents. (Image 0001.jpg has the corresponding transcription file of 0001.txt)
  • Saved with UTF-8 encoding

File delivery

  • UW-Madison has Box and this is an easy way to share a folder of files.
  • Creating a folder in R:/temp works right out of the gate, but the transfer needs to be carefully coordinated as that directory is cleaned out on a regular basis.
  • Email programs can do some awkward things with text-file attachments, so the safest way to email your finished files is going to be as a .zip file.