Milarepa: ten-page transcription and translation draft

阅读说明

这是十页藏文木刻本的图文对照试稿，不是完整译本。藏文栏保留暂定转录，其他十二种语言提供初步译文。无法确定的片段保留为 [?]，该符号可能代表一个字或更长的缺文；不根据上下文补写。页首、页尾的断句没有补全。某些行只能提供片段，某些行尚不能翻译。

请先核对图片和藏文转录，再审阅英文及其他译文。英文为中转文本，其他语言可能继承其错误。可点击整页图片查看原始分辨率。每一文本片段均有稳定编号；具体疑点附在对应行后。此稿使用人工智能辅助生成，尚未经专家审核；它不表示您认可转录、翻译或本项目。

Как читать пакет

Тибетская версия — предварительная транскрипция. Остальные двенадцать — рабочие переводы. Сначала нужно проверить чтение по сканам, затем смысл и терминологию. [?] обозначает неразрешённый фрагмент неизвестной длины, а не обязательно одну букву. Пропуски и оборванные на краю страницы фразы не восстановлены по догадке. Наличие файла на языке не означает полного перевода страницы.

Каждый фрагмент связан с конкретной страницей и местом на изображении. Английская версия служит промежуточной основой большинства переводов; ошибки этой основы могут повторяться в других языках. Точность OCR и переводов численно не измерялась. Проверка структуры файлов не заменяет филологическую и языковую вычитку.

Reading and review method

The source scans remain unchanged. Tibetan, English and Russian readings were drafted with image inspection and a historical OCR hint. Other translations use the provisional English pivot while preserving segment identifiers and unresolved spans. They are not independent confirmations of the source. Read uncertainty notes alongside every segment, including apparently fluent passages.

The package distinguishes the main text, captions, marginal titles and later page labels. A represented region is not necessarily a fully transcribed region. No missing context has been supplied from other editions. Religious and biographical statements are translated as statements in the source, not verified empirical claims.

To propose a correction, cite the page identifier and segment identifier, then supply the proposed Tibetan reading and, if useful, its meaning. This is a small OCR and textual-preservation pilot, not a released training dataset. Publication, contribution reuse and model training are separate decisions.

Source and provenance

• Work: མི་ལ་རས་པའི་རྣམ་མགུར། — life and songs of Milarepa; BDRC edition W1CZ1988, volume I1CZ2041. Authorship metadata: Tsangnyön Heruka. The edition date has not been established here.
• BDRC access policies explicitly use W1CZ1988 as an example of a public-domain work without additional restrictions. This is the basis for distributing these particular images; it is not inferred from the licence on catalogue metadata.
• BDRC edition metadata links the electronic reproduction IE0OPI5E6DA852 to OpenPecha I5E6DA852, commit 08030c148470ed5a987bdf5f36903eeec2799732. Its metadata identifies Google OCR and CC0/public-domain status. The OCR is historical and noisy; it is a reading aid, not a verified transcript.
• The ten images are the same sample listed in the Matreya archive. They include clear, degraded and difficult pages, plus illustrated layouts. Original file identifiers and folio labels are retained.
• Fonts bundled for Tibetan and other complex scripts are derived from Noto Serif under SIL OFL 1.1. Licence notices accompany them. Chinese, Japanese and Korean use the device's installed fonts. The package makes no external font requests.

Machine-readable source and translations: source.json. File checksums: sha256.json. Open a language above for aligned text and a plain-text download. Unzip the whole folder before opening this index.