Project Gutenberg's epubmaker & PG RST

[Tom Morris]

I haven't seen this mentioned anywhere.  It's apparently what PG uses to generate their epub formats from either HTML or RST (mostly HTML).

- epubmaker - http://pydoc.net/Python/epubmaker/0.3.20/

    http://www.gutenberg.org/tools/

- epubmaker online conversion service - http://epubmaker.pglaf.org/

- PG RST spec / test doc - http://pgrst.pglaf.org/publish/181/181-h.html

Related PGDP docs

- HTML best practices doc - http://www.pgdp.org/~jana/best-practices/pages/best-practices/

- http://www.pgdp.net/wiki/The_Proofreader%27s_Guide_to_EPUB

Tom

[Eric Hellman]

Tom,

 this is very useful, thanks. 

We're experimenting with running epubmaker with Travis-CI. Not quite there yet.

Somehow I'd missed http://upload.pglaf.org/   That fills in things I'd been wondering about.

Eric

[Tom Morris]

On Thu, Mar 26, 2015 at 9:07 AM, Eric Hellman <er...@hellman.net> wrote:

Somehow I'd missed http://upload.pglaf.org/   That fills in things I'd been wondering about.

The sequence, as I understand it, is:

1. Get a copyright clearance line - http://copy.pglaf.org/

2. Generate a text - pgdp.net for >50% - then validate by:

  - run gutcheck

  - run HTML validator - http://validator.pglaf.org/ (clone of W3C service)

  - run CSS validator - http://jigsaw.w3.org/css-validator/

  - run link validator - http://validator.pglaf.org/checklink

  - run dry run of epub converter - http://epubmaker.pglaf.org/index.php

3. Upload the text for processing & publication - http://upload.pglaf.org/

At DistributedProofreaders, the sequence is:

- scan book pages, either privately or by acquiring images from Internet Archive, Canadiana Online or other archives

- OCR (mostly using ABBYY FineReader

- three rounds of proofreading

- two rounds of formatting

- post processing of resulting UTF-8 text to create an HTML version (usually)

All proofing and formatting is done to the project manager's specs, largely based on global standards used by DP.

Post processors have a fair amount of latitude to generate the HTML as they see fit, as long as it falls within what can be handled by PG's epubmaker.

Tom

p.s. Looks like PG toyed with the a similar thing to Gitenberg - http://www.gutenberg.org/wiki/Mercurial_Repository_How-To -- perhaps a result of previous discussions about crowdsourcing http://search.gmane.org/?query=crowdsourcing&group=gmane.culture.literature.e-books.gutenberg.volunteers