Upcoming Webinar of direct interest!
1 Like
At OERSI we also think about GitLab.com or any other instance of GitLab as a potential repository for OER. With some simple rules for tagging repositories and adding metadata we are able to harvest GitLab instances. A proof of concept is OERSI GitLab example.
Really good to know!
Are users finding it easy enough to use for both producing the content and adopting/adapting it?
In my experience, git concepts tend to be offputting to learning pros. The most efficient introduction I’ve found so far is this very short video from Steel Wagstaff to use “issues” in Github as a way to share ideas about Pressbooks (via @amysong, I think).
Filing Pressbooks Feature Ideas and Bug Reports in GitHub - YouTube
During this webinar (in French; at 2903s), UdeM librarian Pascal Martinolli shared his use of Github as an OER repository to solve two main issues: stable URL and version archiving.
(He also mentions the issue with Github’s ownership. Surely, he knows about Gitlab. There’s actually a provincial instance from the Quebec government and, as far as I can tell, we should be allowed to use it.)
“git concepts tend to be offputting to learning pros.” You can say that again and again. I’ve never met anyone who actually teaches in a K12 (elementary and secondary) classroom who has ever heard of Github. There are a few tech integration people who know about it and even some who might use it. People who are in the classroom for most of the day likely won’t have any interest in learning about Github, and I won’t blame them.
.
Good to know. A bit more common in HigherEd… still not that wellknown.
Maybe the solution is a user-friendly frontend. Gitlab could serve as a backend with nice features as described @axel.klinger. We sure need excellent frontends in the OER world. (#Discoverability and all that.)
The OERSI shared by @axel.klinger does not require (I believe) any end users to even touch GitLab/Github or even say the word “git” – If you look at the two examples he shared, they are both web sites at external links (technically one is published at GitHub but it is a stand alone web site).
If I understand right, an organization that has an OER Collection would import records to GitLab instance that could then be harvested/searched from OERSI. But all of the OERs exist elsewhere, it is very much a referatory. What he is showing is a functioning federated OER search.
This is covered in Axel’s 2022 presentation
Gotcha! In that sense, it sounds a bit like eCampus Ontario’s use of Drupal as a backend.
Still thinking about the overall workflow from different sides. If I understand @BarbaraClass’s need, the institutional repository needs to fulfill different requirements in terms of technical infrastructure (including interop) and usage (including search and possible leapfrog from AI use). In fact, it sounds to me like part of the question revolves around local instances as opposed to federated ones, in relation to what organisational leaders perceive as a requirement.
(I realize I’m probably extrapolating. It’s an “educated guess” based on prior discussions we’ve had. And I guess we should get in touch at some point.)
All this to say… Is there an OEG workgroup on platforms, by any chance? If not… would that be helpful? Maybe with some tie-ins with OpenEd Tech, Open Data, etc.?
The approach is indeed to consider GitLab or even GitHub as a repository. Even if in the examples given above the actual contents are in external Git projects, these could also be in the same project as the metadata. GitLab/GitHub can thus be used both as a repository or as a referatory.
Certainly, git itself is quite sophisticated if one uses the full collaboration possibilities, but for interactive courses the platforms GitLab/GitHub (whereby the former is open source and can also be set up locally/regionally) offers far more possibilities for collaboration, versioning, project management, extension of functionalities through templates etc even with simpler use cases than any other kind of repository known to us, such as DSpace, document management systems or learning management systems, and last but not least due to its success in the last 10 years in open open source development, has proven its success on collaborating in open projects. GitLab may not be the one and only solution, but it is already existing and open for the public with a lot of great desired features.
As already mentioned, the above examples are still a proof of concept, but they also already show how close we are to a possible solution for a global open source based infrastructure.
Our focus is higher education, and the hurdles there are certainly lower than in the school sector. We see ourselves how difficult it is to inspire teachers at universities for GitLab and therefore we also try to promote these technologies and platforms among students by showing them how they can already produce better, more dynamic and future-proof results with study projects or master theses based on RAW texts, separating content from layout (Markdown to LaTeX) and thus support the next generation of scientists and teachers in higher education in the direction of Open Science and Open Education.
Theoretically, you can also share PowerPoint and Word documents on GitLab/GitHub, but the big advantage is of course text-based content such as Markdown or LaTex, from which other output formats such as HTML, PDF or EPUB can be generated by templates.
An example that might also be interesting for students is an (incomplete) reconstruction of my student research project, comparable to a bachelor thesis, which I converted from a Word document to Markdown in less than three hours. The output formats HTML, PDF and EPUb all look better than the 20-year-old original one and I didn’t have to worry about formatting. If students learn to work with those technologies today, the next generation of teachers will be well prepared for Open Education 
Last but not least, by tagging content with metadata according to CreativeWork - Schema.org Type, it will also be easier to find on Google, e.g. “studienarbeit fassade” or “gitlab für texte”.
One goal for me would be to create an introduction for the next OE Global Conference, on how to create a basic course/lesson setup, publish it an make it findable on OERSI in less than 5min. Here is a short preview of the first two steps in about 2min GitLab-OER-Teil1.mp4 · main · Axel Klinger / GitLab für Open Educational Resources · GitLab. Sorry, that it is also still available in german, but you may get an idea of how this could work.
1 Like
I am also very interested in an international working group on platforms and infrastructure for OER and would like to get involved.
The Saylor Academy uses GitHub as a repository for OER textbooks. (They’re more or less a mirror of the old Flat World Knowledge ones that were released as BY-NC-SA ten or so years ago and preserved by Andy Schmitz.) Click on any of the book titles and you’ll go straight to the book in HTML on GitHub:
1 Like
Thanks a whole lot for the explanations, Axel!
In case it’s not obvious: I’m really excited about such git-based workflows. I’m not criticizing any of it. Just “thinking out loud” about what might be involved upstream/downstream for a variety of people (including the institutions creating their own repositories).
(And, yes, if some of us in Quebec end up publishing OERs through Gitlab, I’ll do my best to make sure to connect with OERSI.)
Which all makes a whole lot of sense for well-structured documents, especially academic articles, particularly in STEM.
Reminds me of Quarto, which I just heard about recently (through the Fediverse, obviously). It uses Pandoc. So, Markdown to these other formats.
Some people produce OpenTextbooks (and, presumably, other OERs) via PreTeXt.
https://alir-jsturcotte.profweb.ca/frontmatter.html
At this point, I’m even thinking about where Pressbooks and Libretexts fit.
I know Bernard Pochet (librarian and university press at Liège) also created a whole workflow integrating Markdown/pandoc and, as I recall, ownCloud.
Quel outil pour rédiger (en science) ? | infolit.be
GitHub Pages was also an option.
At this point, he’s using Pressbooks.
Bienvenue sur e-publish – L’édition numérique ouverte (uliege.be)
Also related: a discussion with the Scenari Community about MikTeX.
Somehow, it feels like all these paths are converging to a cohesive “solution space”, with room for all sorts of workflows.
If so, that could be great news for the OE movement.
And, at the risk of frightening some people, it even sounds like something ISO’s SC36 committee might take on. Standardization can have tremendous benefits when done appropriately.
1 Like
I can’t say it’s not been done before but currently we do not have work groups… but why can’t we start? What would they do? It seems like there are many independent projects like the OERSI and more.
I think this is a great idea, support it fully. Best Ebba
1 Like
Hi Axel.
I followed the link in your post, and have a question.
Re. “…it is important to consider that even if the source materials are deposited in a repository, they usually do not provide an approach to contribute to the original document and improve it. Instead, they tend to encourage the creation of further derivatives,…”
I’m struggling to understand the difference. Wouldn’t improvement of the original document be a derivative, an updated version?
d
1 Like
Hi David, if I make a copy of the original, correct a spelling mistake and republish the copy, then there are two similar versions. If I offer the correction to the author of the original and he accepts it, then there is only one new version of the original. Best, Axel
1 Like
Just as an example, if we have a version control system, OERSI would always reference to the one latest version, while older versions would be available behind the scenes: current version Open Educational Resources Search Index and older versions Releases · progit/progit2-de · GitHub. There will be no different version in OERSI to confuse the users. A copy and republication would lead to two entries in OERSI, which could confuse users.
1 Like
Thanks for the clarification.
This makes much sense, though we might speculate how much change makes it more than a modified version, not to mention the complexities if David makes a new work by adding chapters off of say version 2.122 and I make a changes version from 2.198… Or that many OER textbooks are built from a start of significant reuse of multiple works.
And versioning might get to the question of tracking the reuse of OER which now at many places seems to lean on a volunteer system of notifying by a web form (vague memories long ago David, of Scott Leslie pondering how this could be done with documents).
Axel, I am intrigued to see this ability of OER to be tracked by OERSI by adding a tag in Github or Gitlab, and having appropriate metadata (looking at the few titles that show up, at least they are not all computer science). It’s an uphill effort to get many authors to take on the Git system, it might be interesting to set up some way for people to “git” an experience with something smaller in scale than a whole textbook.
Versioning systems on a larger scale would help in so many ways,.
I just came across Liascript used for the OER GitHub tutorial which looks like an elegant way to render course/OER content from a Github source, a bit like @paulhibbitts Docsify This.
All of this takes me back to a circa 2016 project where I was advocating sourcing content in Markdown from a git repo, and rendering it in WordPress, Hugo, etc. Later the project ended up in an LMS… so I dream of more Git based solutions.
1 Like
I suggest that the number of people on the planet who might use OER with actual students in a classroom who also understand the meaning of the phrase, “sourcing content in Markdown from a git repo, and rendering it in WordPress, Hugo, etc,”
It’s not only a matter of authors taking on the Git system; most people have never heard of Git. Go to an elementary school near where you live and ask how many teachers know what Git is. Then, try to convince them why they might want to know what it is.
And, I don’t see the benefit of keeping track of ‘versioning.’