- From: Hadrien Gardeur <hadrien@demarque.com>
- Date: Mon, 20 Sep 2021 19:03:10 +0200
- To: Laurent Le Meur <laurent.lemeur@edrlab.org>
- Cc: "George Kerscher (montana)" <kerscher@montana.com>, Dan Lazin <dlazin@google.com>, W3C EPUB 3 Working Group <public-epub-wg@w3.org>
- Message-ID: <CAHqp8zguDdqz44Z1YYwq+FqsknbjjWTVXjmEjrh5dNp2pCTgbw@mail.gmail.com>
Hello, I think that it's quite misleading to think that because publishers provide a page list, reading systems will use them instead of generating their own position list. As I've pointed out in a separate section of our working document <https://docs.google.com/document/d/11GypOjE9xOTaINATl5bxVIA3Mc9jzNBGCr6GT_KNaQ4/edit#bookmark=id.2gd7n6l9y7gi>, they're not the same thing: - string vs integer - complete vs incomplete (a page list can skip multiple references) - authored vs generated Both will co-exist, even in a world where most EPUB contain a page list. Before we even discuss "how" we generate these references, we should acknowledge that we can't change every reading system out there and that they have good reasons for keeping both references side by side. This feels very similar to the whole discussion about markup in table of contents, where we request something from RS that's pure wishful thinking right now. Best, Hadrien Le dim. 19 sept. 2021 à 19:38, Laurent Le Meur <laurent.lemeur@edrlab.org> a écrit : > Hi Georges, > > I certainly agree with you: publishers should provide page lists. This > will be the case from 2025 (max) for every ebook distributed in Europe, > since page lists have been selected as a mandatory feature of accessible > ebooks. > > From what I've seen so far, companies that ingest titles (i.e. > e-distributors) never modify the content provided by publishers. In the > trade sector, they don't have such a right. > > In consequence, "calculated pages" may soon be something RS rarely have to > compute. This is IMO another good reason for creating an algorithm that is > not a software nightmare, and move to more interesting use cases (sharable > annotations maybe?). > > Best regards > Laurent > > Le 18 sept. 2021 à 16:15, kerscher@montana.com a écrit : > > Hi, > > I think I have said this many times, but we should first encourage authors > and publishers to insert the page numbers, and then RS must not then > calculate their own. > > Next, companies that ingest titles should check if page numbers have been > authored, and they should modify the titles using a common algorithm, > assuming they have the rights to make modification. > Only then should RS should use calculated virtual pages. > > Best > George > > > >
Received on Monday, 20 September 2021 17:03:36 UTC