This is an in-progress open source translation of Anatoly Kuznetsov's "Babi Yar" using automated translation with human assistance. The source files can be compiled to both Russian and English epub and pdf digital files and you can even build them yourself.
Everything from "To the question of Heaven and Earth" and onwards is machine-translated
and not yet edited. It's open game for anyone interested. For the sections
preceding, I used a combination of Google, Amazon, Deepl, and Yandex translation. For
everything after 05.txt in the /src/ directory, I used Google translate (gcloud).
I didn't intend to release this until it was in a better shape, but I figure it's now or never: putting this out in the open might be helpful to someone and anyone can contribute, and the epub and pdf can be read and shared as-is, as long as no one minds that the last 3/4 of the book still needs human attention.
This was the most important and affecting book that I've ever read, and Kuznetsov's warnings to us are vital and timeless. That such an important work is not in print in English, and wasn't available digitally seemed criminal. (The English translation is now available, thanks to archive.org!, but there's still no epub).
Please be aware that the content included here is very disturbing and can be overwhelming. But it is what happened and it's important that it's accessible.
If you do nothing else, read all of the "From the Author" chapters.
You can build your own PDF or epub on Linux (or using the docker image @ jenh/eoat:v2).
To build the epub and PDF
-
Install eoat-tools dependencies described in the
EOAT READMEAlternatively, you can use a pre-build docker image with most required tools installed:
docker pull jenh/eoat:v2 -
Run
scripts/eoat-build $langwhere the language isenorru.
- Only 10% of the English content has been reviewed and cleaned up by a human so far.
- None of the Russian version has been formatted or checked.
- The Russian version PDF title is in English.
- The LaTeX and CSS files could be made much nicer.
- Because I'm using a hokey way to select matching language strings from source files, we may be missing text or have extra text. Started looking at lingua-language-detect as a better alternative, but it will take a little time.
- Many more identified and unidentified!
Because I did a lot of this work before I finished EOAT, most of the source text was obtained using the OCR output created by Nina & Leon Dotan http://ldn-knigi.lib.ru We are all very indebted to them for making this and other resources available!
Before I went down the rabbit hole with EOAT for automated OCR and translation, I relied on Mort Yao's wonderful translate-shell for translation and Govarthani Rajesh's Text Scanner (OCR) for easy and selective OCR.