Guidelines for submitting documents to Epsilon

If you have data, we want it! You might not realise you have data - transcriptions of letters and lists of letters both count as data, and we’d love to incorporate them if at least some of it has scientific content and touches on the nineteenth century. We’re keen to add to our collection, whether you’re a project or archive with a large collection, or a scholar sitting on a set of transcriptions produced as a research side product. If you have anything that seems appropriate for Epsilon, please get in touch with us at

Epsilon’s core data is all in TEI XML and we encourage anyone starting out to use that form (we’re happy to talk to anyone wanting to start a project about working in TEI). However, we can take data in multiple formats. The most important thing is the regularity of the data. Please see below for our requirements.

Non-XML Source Materials

Two separate collections of data are needed when you want to submit materials that aren’t in TEI XML:

  1. An excel or CSV (comma separation values) file containing the core fields for the correspondence data.
  2. Transcriptions, preferably in DOCX, ODT, HTML or DOC format.

Structured Data in Excel or CSV

The field names in the file must match the following exactly and those marked as ‘Required’ must be present.

Field Name Status Description
id Optional A unique identifying record number for the record
sender_surname Required Surname of the sender
sender_forename Required Forename(s) of the sender
recipient_surname Required Surname of the recipient
recipient_forename Required Forenames(s) of the recipient
sorting_date Required

An iso formatted version of the date of the letter. This value is used for sorting. It’s also used when the desired textual version exactly corresponds to it. For example, YYYY-MM-DD, YYYY-MM, YYYY will be automatically rendered as Day MonthName Year, MonthName Year and Year. If the desired date string is more complex than that, a value can be provided in the ‘date’ field.

NB: The -MM-DD format is not supported.

date Optional

A textual version of the date of the letter when it cannot be rendered from the sorting_date value. For example:

15-16 Oct 1870 15? Oct 1870 15 Oct? 1870 [1, 8, 15, 22, 29] Oct 1870 Late Oct 1870
sender_address Optional A textual version of the sender’s address
recipient_address Optional A textual version of the recipient’s address
source_custodian Required The name of the custodian of the item, possibly an institution or a private collection or the full bibliographical reference if the source is printed.
source_classmark Required Classmark for the item (if manuscript), otherwise, blank.
languages Optional The language(s) of the letter
extent Optional The length of the letter, expressed in either pages or folios.
filename Required Filename of the transcription of the letter (if any).

Try to be consistent when entering the data.