Ein Beitrag von Ilya Khait
The paper overviews the ongoing development of JTF (JSON Transliteration Format) and related infrastructure: jtf-lib, the format’s API, jtf-signlist, a cuneiform sign list utility, and uqnu, a cuneiform transliteration editor. The motivation behind this initiative, an offshoot of the MTAAC project, is to make digital cuneiform data more accessible for both automatic processing and editing by humans.
Currently, the only machine-readable specialized transliteration standard is ATF (ASCII Transliteration Format), a textual data format of two dialects designed to use with the two main Assyriological digital platforms, CDLI and ORACC respectively.¹ ATF provides rich semantic markup for cuneiform data and metadata and serves well for archival purposes, but as a complex text format it is laborious to write, prone to errors, almost impossible to safely extend with new features and difficult to parse. No less importantly, the field still lacks a reusable code library for basic operations on ATF, e.g. to modify the data.
This bottleneck of Digital Assyriology can be significantly broadened by providing the means for simpler work on data on both the human and the machine side. JTF, created to assist solving the latter issue, is a JSON-based structured document format with a detailed mapping that fully mirrors both ATF dialects and is compatible with them. It is not intended to be edited by users directly, but offers serialized data that is programmatically simple to access, manipulate and extract.
Below is an example of a short transliteration fragment in ATF from CDLI P000001, a proto-cuneiform lexical list:²
@tablet
@obverse
@column 1
…
3′. 1(N01) , KINGAL#
Its JTF counterpart would be the following:
{“_class”:”object”, “type”:”tablet”, “children”:[
{“_class”:”surface”, “type”:”obverse”, “children”:[
{“_class”:”column”, “name”:”1″, “children”:[
…
{“_class”:”line”, “name”:”3′”, “children”:[
{“_class”:”field”, “children”:[
{“_class”:”sequence”, “type”:”short”, “children”:[
{“_class”:”chr”, “type”:”num”, “value”:”1″, “unit”:
{“_class”:”chr”, “type”:”sListRef.protoCuneiform”, “value”:”n”, “index”:”01″}}
]}]},
{“_class”:”field”, “children”:[
{“_class”:”sequence”, “type”:”short”, “children”:[
{“_class”:”chr”, “value”:”kingal”, “damage”:”#”}
]}]}]}
JTF represents transliterations as tree structures with detailed mapping of different data. It uses JSON dictionary elements (named by their _class attribute) such as object, surface, column, field and line to describe the text’s relation to its medium, appearance and structure, while others, e.g. chr and sequence stand for inline data elements, respectively glyphs and their grouping. All JTF classes are provided with a set of attributes to store the necessary information on the element at hand. Besides the text tree, JTF includes a metadata dictionary with fields for CDLI ID, language, genre, bibliography, museum numbers, and other information.
JTF comes together with jtf-lib,³ a NodeJS API and CLI for JTF-related operations. At the moment, it facilitates essential functionality. Thus, the module has a NearlyJS parser for ATF in order to validate and convert ATF data (from a string, a local file or CDLI) to JTF. To automatically process changes to JTF content and to ensure its integrity, a special formatter is used. In development are the JTF CRUD (create, read, update and delete) API, as well as a converter from JTF to ATF.
A related standalone NodeJS module, jtf-signlist,⁴ is a cuneiform sign list API with features such as Unicode normalization, sign value checkers and bibliographical references.
A React web application that relies on these tools, uqnu,⁵ provides a GUI for users to work on cuneiform data. By now, it embeds components for importing ATF (loader, checker and editor), converting it to JTF and rendering the transliterations in a convenient way. The ATF importer facilitates mapping and asynchronous reading of multi-text ATF files (more than 123 000 texts). The user can switch between loaded files and see the list of texts included in each of them, then open each text to see its ATF content, check if the formatting is valid, modify, and save it back to the file. The importer automatically converts ATF data to JTF. The prospective core of the application, at the time a tab where the user can see the JTF rendered in a human-friendly form, is soon to be expanded with widgets for editing in order to serve as a JTF-based document processor. Additional features and integration with existing tools, e.g. for linguistic and paleographic annotation, is also envisaged for the future.
The infrastructure is to lay the foundation for a web-based open access digital apparatus for Assyriologists to use both privately and as a crowdsourcing platform that would accelerate the accumulation and reuse of cuneiform digital data.
Kurzvita
Ilya Khait attended the Russian State University for the Humanities and finished his PhD. at the Leipzig University with his thesis Typology of Old Babylonian Divination Apodoses in 2018. He is currently working at the Leiden University. His interests include Cuneiform Studies, (Computational) Linguistics, Digital Humanities, Linked Data & Ontologies, Lexicography, Writing & Paleography and Archeology.

