The first Area of Action in the 2019 UNESCO Recommendation on OER
(https://en.unesco.org/themes/building-knowledge-societies/oer/recommendation) calls for “building the capacity of stakeholders to create, access, re-use, adapt and redistribute OER.” From a technical perspective, that capacity was already addressed more than 10 years ago by the ALMS framework (Hilton et al. 2010), which aims to simplify creation, reuse, and remix of OER with desirable technical criteria for OER tools and OER source files. However, building upon Hilton et al. (2010), Ovadia (2019) still observes “technical challenges [that] will prevent more widespread adaptation of material.” Addressing those challenges, emacs-reveal serves as OER infrastructure and as case study for the transfer of best practices in software engineering to the domain of OER.

Emacs-reveal is a free/libre and open source software (FLOSS) bundle with CI/CD (continuous integration, continuous deployment) infrastructure on GitLab (a) to generate OER from input files in the lightweight markup language Org Mode and (b) to publish them as GitLab Pages on the web. The primary target format of emacs-reveal are HTML presentations (using the JavaScript presentation framework reveal.js) with voice-over, which are accessible on any platform with a recent browser, online as well as offline (after download of a selfcontained archive). “Ordinary” web pages and PDF variants serve as secondary target formats. As presented previously (Lechtenbörger 2019), emacs-reveal automates the time-consuming task of license attribution when reusing OER figures based on machine-readable RDFa metadata.

Beyond that earlier work, three OER courses at GitLab (https://gitlab.com/oer/oer-courses) demonstrate that lightweight markup mechanisms not only allow to reuse figures, but they also enable reuse along other dimensions: First, with include mechanisms, smaller learning objects can be remixed according to varying scopes and purposes into larger units without copy and paste; thus, updates and improvements do not exist in isolated places but in a single, shared source (which is preferably managed in a version control system such as Git). Second, with lightweight markup languages, layout information is kept separate from learning contents. Thus, (a) content experts do not need to be layout experts and (b) remixing can focus on contents regardless of specific layout conventions of, say, different collaborating institutions. Third, emacs-reveal adopts licensing recommendations of the REUSE project (https://reuse.software/) in the form of so-called SPDX headers in every source file, from which license information of remixed OER is generated automatically.