# Methods note

## Purpose
This dataset was prepared as a **bibliographic publication companion** to the article *“The Adriatic Sound World: Listening to the Sea (1300–1650)”*. Its purpose is to make the article’s source base more transparent, not to publish a full passage-level or annotation-level reproduction of the interpretive process.

## Source material used for dataset preparation
The final dataset was compiled from two source databases:
1. `Sourcebook_Primary sources_list with citations_complete.xlsx`
2. `Sourcebook_The Sea_from itineraries_Adriatic sea-with focus on Dubrovnik.xlsx`

The wider inspected corpus was taken from the primary sourcebook and cleaned into a one-row-per-work bibliographic dataset. The article corpus was identified from the Adriatic sea sourcebook and then aligned against the wider inspected corpus. Contextual/local works used in the article were added explicitly as separate rows.

## Release design
The released CSV combines:
- the wider inspected travelogue corpus;
- the focused article-corpus travelogue subset;
- the contextual/local non-travelogue works used in the article.

This design keeps the dataset simple while allowing readers to distinguish between:
- the full inspected travelogue corpus;
- the article’s focused corpus;
- the article’s contextual/local supports.

## Normalisation principles
- One row per work.
- Unicode and punctuation were normalised for readability.
- Author forms were regularised conservatively.
- Language records the **original source language**, not the language of the edition used.
- Genre labels were simplified into a small controlled vocabulary.
- `edition_used` and `publication_details` were kept concise and pragmatic, not expanded into full catalogue records.

## On supplemental rows
A small number of `SUP-*` rows were introduced where works present in the article sourcebook were not preserved as distinct cleaned rows in the wider inspected corpus dataset, but were still needed for a transparent article companion. These rows are marked `wider_inspected_corpus = no` and `in_article_corpus = yes`.

## Limitations
This is not a critical bibliography in the strict bibliographical sense, nor a full evidentiary dataset of quoted passages. It is a cleaned bibliographic corpus release designed to accompany the article and document the source base behind it.