402.1 Adventures with ePub3 -- Runner Up: Best Short Paper Award

The role of standards in digital preservation is widely acknowledged. The current version of the ePub standard, used for publishing and disseminating eBooks, is ePub3, specifically 3.1 (January 2017). A marked difference from ePub2 is support for fixed layout files and, whilst several different ePub readers are available, not all have upgraded to provide full support for ePub3. In late 2017 and early 2018 the British Library’s digital preservation team undertook research into the impact of using an ePub viewer without explicit support for ePub3 on a mixed sample of ePub3 files. The sample comprised 54 files: 20 of these were of fixed layout, the remainder utilized reflowable layouts. For the analysis, content was accessed using two different open source ePub readers: Calibre, which has a wide user base but does not currently explicitly support ePub3, and Readium, which does explicitly support ePub3.
ePub3 files with a reflowable layout broadly rendered to an acceptable standard using both readers. There was one notable exception for both readers, and investigations indicated this was likely due to a problem with the file itself rather than the rendering software. Problems manifested in a more serious way when using Calibre to access just under half of the fixed layout ePub3 files. The rendering of these items inhibited access to intellectual content for example by overlaying it on other content, misrepresenting it, or not including it at all. Other issues that were initially apparent with the other half of the fixed layout sample were mostly resolved by the simple quick fix of switching from ‘page’ view to ‘flow’ view. By contrast, Readium was able to display all of fixed layout files correctly.
The research serves as a reminder that whilst standards remain an essential tool in the digital preservation toolbox, updates to a standard may necessitate changes to the rendering software in use. It further underlines the importance of accurate characterization so that repositories can identify formats at a version level or be able to identify items with explicit rendering needs beyond those served by their default rendering viewers.
This paper was the runner up for the Best Short Paper award.

数字保存领域中对标准作用的认可已广为人知。当前用于出版和传播电子书的 ePub 标准版本为 ePub3，具体为 3.1 版（2017 年 1 月）。与 ePub2 相比，ePub3 的显著差异在于对固定布局文件的兼容性。尽管市面上有多种 ePub 阅读器，但并非所有都已升级以提供对 ePub3 的全面支持。2017 年晚些时候至 2018 年初，英国图书馆的数字保存团队对在不具备对 ePub3 明确支持的情况下使用 ePub 阅读器对一组混合的 ePub3 文件的影响进行了研究。该样本包含 54 个文件：其中 20 个为固定布局，其余则采用了可重排布局。在分析过程中，使用了两款不同的开源 ePub 阅读器访问内容：Calibre（拥有广泛的用户基础，但尚未明确支持 ePub3）和 Readium（明确支持 ePub3）。两款阅读器均能将具有可重排布局的 ePub3 文件广泛地渲染至可接受的标准。然而，两款阅读器均存在一个显著的例外，调查表明这可能是由于文件本身的问题而非渲染软件。当使用 Calibre 访问接近一半的固定布局 ePub3 文件时，问题变得更加严重，这些文件的渲染阻碍了对知识内容的访问，例如通过叠加在其他内容上、扭曲其表现或完全不包括它。其他问题在固定布局样本的另一半中最初显现，但通过简单地将视图从“页面”模式切换至“流”模式，大多数问题得以解决。相比之下，Readium 能够正确显示所有固定布局文件。这项研究提醒我们，虽然标准仍然是数字保存工具箱中的基本工具，但标准的更新可能需要更新所使用的渲染软件。它进一步强调了准确表征的重要性，以便存储库能够识别版本级别的格式，或能够识别需要超出其默认渲染查看器所提供功能的明确渲染需求的项。本文荣获最佳短篇论文奖的亚军。