Weigh in on ISO639-3 language code implementation

[Martha Tenney]

There has been a request to implement ISO639-3 (ISO639, set 3) language codes in ArchivesSpace, with the goal of being able to accurately represent more languages (ISO639-3 includes over 8000 codes whereas ISO639-2b, the set currently used by ArchivesSpace, includes under 500). We want to hear from you: read more below and share your thoughts via this form (https://forms.gle/khf92JYHFYTYBuEd7) by Monday, June 1.

Most ISO639-2b language codes map to ISO639-3 language codes. However, there are a couple significant differences:

• ISO639-2b includes collective language codes which have no analogous entries in ISO639-3.
• In ISO639-2b around 20 of the most common languages have codes derived from English (fre for French) as opposed to the ISO639-3 codes which are derived from the name of the language in the language itself (fra for French). These would need to be converted.

The Metadata Standards subteam has come up with a few possible paths, and we are seeking your input:

• Option 1: Fully implement ISO639-3 and deprecate ISO639-2b. Make choices about the collective language codes not in ISO639-3. In a future release (allowing for ample lead time to allow community members to decide what to do about collective language codes), run a database migration to convert all ISO639-2b values to ISO639-3. Write mappings into import and export.
• Option 2: Develop a plugin for ISO639-3 or other language sets implementation and migration, with caveats that they may break MARC or other imports/exports and/or create issues in the PUI and that additional mapping would be needed in order to create valid imports/exports. Allow institutions to perform their own adoption and mapping at their own discretion.
• Option 3: No change/do not adopt ISO639-3 now (leaving open the possibility of future adoption).
• (Note: we also considered an option to keep ISO639-2b values as a required field and add a second column in the database for ISO639-3 values. We determined that this would be overly complicated for both staff and public interface users and would either require duplicate language fields or converting the languages to a linked record rather than a controlled value and creating a new table in the database.)

Any path to adopting ISO639-3 presents questions for MARC imports/exports, EAD imports/exports, and other external standard imports/exports. More context for these questions and a discussion of the three options can be found here. A note, this is a preliminary inquiry and any changes would likely take some time to develop and implement, particularly because they would likely require data mapping and cleanup.

Please share your thoughts via this form (https://forms.gle/khf92JYHFYTYBuEd7) by Monday, June 1. We would also appreciate it if you could circulate this form to others at your repository who have experience with MARC cataloging, EAD encoding, and/or ISO 639/language codes more generally.

Thank you!

Martha

Martha Tenney

ArchivesSpace Standards and Testing Archivist

martha...@lyrasis.org