Next generation journal publishing and containers

Some challenges of working on the next generation of research infrastructures can be solved most effectively by talking to other people. That is why o2r team members Tom and Daniel were happy to learn about the announcement of an Open Journal Systems (OJS) workshop organised by Heidelberg University Publising (heiUP).

The o2r team was a little bit the odd one out. Other workshop participants either had extensive OJS development experience, or were not developers at all but running production systems of many OJS journals across the German university landscape. But that could not keep us from telling everyone about Executable Research Compendia, of course. We briefly summarised our plans to extend OJS with ERC capabilities, but we also had new stuff to share! Tom is considering to put his geo-informatics skills to use and extend the metadata of OJS articles with geospatial features in his Bachelor thesis. This would allow to display the spatial area of articles on a map, and even browse articles by their location(s). Learn more about these ideas in our slides.

Today team members @nordholmen @herrniers meet the German #OJS developer and user community at a workshop organised by @heiUP_HD @ojs_pkp

Of course we want to talk #spatial data and #ERC in OJS!

🌍❤️📚#SpatialIsSpecial #ResearchCompendium #ScholCommhttps://t.co/5FenW7WUae pic.twitter.com/BrxkAiSE72

— o2r (@o2r_project) February 20, 2020

But Tom and Daniel also came with a mission: to jumpstart the struggling OJS developments with the help of some experiences OJS developers. None of our team has extensive experience with PHP, so getting control over the huge OJS codebase and setting up a proper OJS development environment with debugging was an important task they’ve been pushing aside since autumn last year. And we got it! [Note to self: don’t forget to enable remote_enable and remote_autostart for Xdebug in the file /etc/php/7.3/cli/conf.d/20-xdebug.ini for debugging to work - then the default VSCode configuration with port 9000 will just work (-:]. On top of that, Tom got a very helpful introduction to writing OJS plug-ins, and Daniel now has a good graps on the currently developed Docker images for OJS. The Docker images are not a simple project, since the PKP team plans to support multiple webserver implementations, multiple PHP versions, and all OJS versions still in production somewhere… phew! Daniel even opened a pull request and suggests a different way to support both remotely and locally built images. This prepares us well for the moment when we want to run OJS on our own servers - in containers of course. So the expectations were high, but eventually they were not disappointed. Daniel was glad to see some familiar faces from the OJS-de.net community he met at a previous workshop in Heidelberg. The new contacts made were more just as important as the helpful practical tipps towards becoming real “OJS devs” 🐱‍💻.

Other groups at the workshop reported very interesting results, for example on the connection of OJS with proper digitial archives (with promising mentions of also archiving data… and code?), more flexible publishing workflows with own tools (mentioning Pandoc, which might make these flexible pipelines a first step towards (R) Markdown-based OJS publications ⚙️), and using search indexes such as Solr and Elasticsearch within OJS (which also have geospatial capabilities 💯). As you can see, we’re very hopeful future collaborations will spark from these educational and entertaining encounters.