![\"\"](\"https://blog.front-matter.de/content/images/2026/04/image.png\")
One blog was added in March. This increases the number of participating blogs (after adjusting for retired blogs) to 186 , the number of archived posts has grown to 49,606 \u2013 Rogue Scholar is getting closer to the big milestones of 200 participating blogs with 50,000 posts!
Library and Information Sciences, English.
https://orion-dbs.community/blog/
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in March was on infrastructure improvements. The monitoring of the Rogue Scholar infrastructure was improved by deploying a self-hosted observability platform for logs, metrics and errors with dashboards and alerting using the Grafana open source platform:
Th dashboard for key metadata metrics initially released in March 2025 was improved visually and launched for communities, including blog communities:
![\"\"](\"https://blog.front-matter.de/content/images/2026/04/image-1.png\")
This makes it much easier for readers to get an overview for each blog participating in Rogue Scholar, and for blog authors to see gaps in metadata coverage that they can improve.
This week the blog self-management in Rogue Scholar was improved, enabling blog owners to update all relevant blog metadata.
The technical updates mentioned above are part of an effort to align Rogue Scholar better with the InvenioRDM repository platform. This will make it easier in the long run to sustain and update Rogue Scholar, as an increasing proportion of the required functionality is built into InvenioRDM and developed and used by other repositories.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
Rogue Scholar uses the community functionality of the InvenioRDM repository platform to manage science blogs and their associated blog posts. In the community settings one can configure basic metadata such as blog name, description, logo and URL.
![\"\"](\"https://blog.front-matter.de/content/images/2026/04/Bildschirmfoto-2026-04-01-um-17.36.07.png\")
By default this information is automatically extracted from the blog feed (RSS, Atom or JSONFeed), and since July 2025 blog authors can change these settings directly in Rogue Scholar, e.g. because their blogging platform makes it difficult to add a logo or description.
Starting this week, more information can be updated: besides the required blog feed URL the optional ISSN and the main OpenAlex subfield of the blog. This is provided by community custom fields, and the subfield uses the OpenAlex subfield vocabulary built into Rogue Scholar and also used for blog post classification.
![\"\"](\"https://blog.front-matter.de/content/images/2026/04/Bildschirmfoto-2026-04-01-um-17.35.20.png\")
Some metadata still can't be changed by blog authors: the license (CC-BY) as it is fixed for all Rogue Scholar content, the Rogue Scholar join date, blog language, feed format (Atom, RSS or JSONFeed) and blogging platform (WordPress, Blogger, etc.) as they are determined automatically, and the DOI prefix, as this can't be as easily changed as other metadata.
InvenioRDM community members can have the roles reader, curator, manager and owner, and only owners have access to these settings. For a blog author to become an owner I have to send out an invitation to an existing Rogue Scholar account, which since the beginning of the year means an account managed by a self-hosted Keycloak instance.
This work integrates with another change to blog communities introduced last week, community dashboards:
![\"\"](\"https://blog.front-matter.de/content/images/2026/04/Bildschirmfoto-2026-04-01-um-18.07.55.png\")
The dashboard shows the metadata from the community form, in the case of language, subfield and blogging platform now as a clickable link to all posts in Rogue Scholar with that blogging platform, etc.
What community owners still can't do is adding or updating blog posts as this is done automatically via their blog feed. And adding a new blog to Rogue Scholar still requires filling out a submission form (at the bottom of every Rogue Scholar page), as blog authors have to agree to provide the full-text in the blog feed with a CC-BY license, submissions are not automatically accepted, and minimal one-time setup work is required.
Please use Slack, email, Mastodon, or Bluesky if you are a blog author with Rogue Scholar and want to become the owner of your blog community, or have any questions or comments.
About 12 months ago Rogue Scholar introduced dashboards with summary statistics of the science blogging archive. At the end of 2025 these dashboards migrated to the Rogue Scholar landing page. Today Rogue Scholar is launching community dashboards \u2013 showing the same summary statistics by community. This feature works with all types of communities: blog, subject and topic communities.
Blog community dashboards help blog authors see major gaps in their metadata coverage, e.g. the percentage of posts that include an author ORCID, ROR for author affiliation, or references. The Upstream blog (where I am one of the editors) for example does a good job including ORCID, ROR and references.
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-27-um-14.47.30.png\")
For readers, community dashboards provide an overview of its activities. The Computational Theory and Mathematics Subject Community published two posts in the last month, and the first post in 2005.
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-27-um-14.48.30.png\")
As before, these dashboards don't show metadata coverage that is always 100% at Rogue Scholar. All nearly 50,000 blog posts have an open license (CC-BY, a few CC0), all posts have an abstract, all posts have a language, and all posts have a subject classification using the OpenAlex categories.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
Running scholarly infrastructure such as Rogue Scholar beyond the initial launch phase \u2013 what is typically called Day 2 operations \u2013 requires the following activities:
As Rogue Scholar is self-hosted infrastructure the natural fit for these activities is self-hosted observability frameworks. One central element is the Grafana dashboard monitoring service health:
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-16-um-17.23.23.png\")
The dashboard collects metrics (ViktoriaMetrics) and logs (ViktoriaLogs) from the InvenioRDM respository platform that powers Rogue Scholar, and monitors the service availability from an external uptime monitor. Errors are sent to the self-hosted GlitchTip service using the Sentry integration built into InvenioRDM.
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-16-um-17.29.54.png\")
Standard prometheus metrics were added to InvenioRDM using the prometheus-flask-exporter library, and log files for the InvenioRDM web and worker services are generated in structlog JSON format and reformatted into ViktoriaLogs format using Vector.
This new infrastructure not only simplifies monitoring service health, but also makes easier specific investigations, for example exploring the bot traffic Rogue Scholar receives each day:
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-16-um-17.41.36.png\")
Future work could include more detailed monitoring of celery workers or a dedicated dashboard for authentication and access monitoring.
The Rogue Scholar observability service is generic enough to be useful for any InvenioRDM repository instance, uses self-hosted open source software, and can be deployed via the included Docker Compose file. The source code is available at https://github.com/front-matter/invenio-observability, for question send an email or reach out via the inveniosoftware Discord Forum.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
This is the February 2026 issue of the monthly newsletter from the Rogue Scholar science blog archive. The newsletter reports on new blogs that have joined the platform, important technical updates in Rogue Scholar infrastructure, community updates, and other news relevant to Rogue Scholar users.
No blog was added in February. This leaves the number of participating blogs to 190, the number of archived posts has grown to 49,342 \u2013 Rogue Scholar is getting closer to the big milestones of 200 participating blogs with 50,000 posts!
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in February was on translations \u2013 the Rogue Scholar website is now translates into 11 languages with blog posts in nine languages. The translations still need lots of refinements but include many areas that were previously not translated, e.g. the sidebar on the homepage, footer, and subject areas. The homepage now shows the most recent posts in the selected language:
![\"\"](\"https://blog.front-matter.de/content/images/2026/03/Bildschirmfoto-2026-03-05-um-09.38.13.png\")
As you can see in the screenshot, the landing page now also shows summary statistics that were previously available on a separate dashboard page. All numbers use live data, and clicking on them returns a search result for that data.
Rogue Scholar is powered by the InvenioRDM open source repository software, and last week the InvenioRDM community met for its annual partner meeting in Graz, Austria. Watch the Invenio blog \u2013 also archived in Rogue Scholar \u2013 for a report in the coming weeks. One important topic was the next major release (v14) of the software in June. Rogue Scholar is running the latest pre-release version of the software \u2013 see the Rogue Scholar footer for the current version.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
This week the InvenioRDM community meets in Graz for five days to discuss the open source repository platform. Front Matter has been part of the InvenioRDM community since August 2021, and Rogue Scholar relaunched on the InvenioRDM platform in October 2024.
On the first day of the workshop we will have a session where InvenioRDM instances can present the highlights of what they are doing, e.g. customizations of the InvenioRDM platform in their repository instance. Here are some of the special Rogue Scholar features launched since the last workshop March 2025 in Hamburg that I want to highlight:
Rogue Scholar registered content with Crossref since it launched. In the last quarter of 2025 I worked on integrating the functionality directly into the InvenioRDM platform. This is a drop-in replacement for DataCite DOI registration, or Crossref and DataCite DOIs can be registered in the same instance. There is more need for discussion at the workshop around advanced features \u2013 e.g. multiple DOI prefixes by InvenioRDM community - but I hope we can finally merge the GitHub pull-request so that Crossref DOI registration is available for all InvenioRDM instances in the upcoming v14 release this summer.
Major releases of the InvenioRDM software happen once a year in the summer, and v14 is on track to be released in a few months. As in previous years, there will be a workshop session dedicated to release planing. Since a few weeks ago, Rogue Scholar is running the latest pre-release version of the software, and is showing the version number in its footer.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-23-um-09.22.02.png\")
And as of last week, Rogue Scholar is powered by Python 3.14, which was released in October and since been added to all tests of Invenio packages.
Also in the footer is a menu to select the repository language. Rogue Scholar has content in currently 9 languages, plus two more translations (Czech and Swedish) for a total of currently 11 languages. The translations not only cover the user interface, but also dedicated text areas (such as the footer, FAQ and search guide), vocabularies and formatted citations.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-23-um-09.33.08.png\")
There are still gaps in the translations, which then fall back to English, but I hope to have them sorted out for the v14 release.
A new feature I launched in January is an update to the \"Recently added content\" section on the homepage \u2013 this section now only shows recently added content in the selected language:
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-23-um-09.37.26.png\")
In the last quarter of 2025 I did extensive work on adding the OpenAlex subject classification to Rogue Scholar. The functionality is documented in detail elsewhere, so I focus on the main features here:
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/subject_areas.png\")
The dashboard showing key summary statistics launched last year is now integrated into the Rogue Scholar homepage. The dashboard uses existing OpenSearch queries to show live data, clicking on a number shows the underlying data. The dashboard doesn't use any external dependencies.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-23-um-10.06.25.png\")
The only special number on the dashboard is cited posts, as this depends on the Crossref Cited-By service and a custom integration. And the dashboard is of course also translated into other languages.
The InvenioRDM platform provides a lot of functionality for managing local user accounts, and for integrating with external services. Still, there are two challenges:
My conclusion for Rogue Scholar was that I wanted to a migrate to a single Identity and Access Management (IAM) external service that should be open-source and self-hosted. After exploring several options I ended up deploying Keycloak as an IAM service. Commercial IAM services quickly become costly once your number of user accounts grows, and many open source IAM solutions either were too limited for the features I needed for Rogue Scholar, or felt too complicated to configure and deploy.
My self-hosted Keycloak instance uses a keycloak-invenio Docker image with several Keycloak extensions that add functionality I needed:
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-23-um-10.35.13.png\")
I can now enforce two policies with user accounts in Rogue Scholar:
Over the coming months I will fine-tune Keycloak both in terms of functionality and user interface. One feature already developed but not needed for Rogue Scholar is enhancing the magic-link extension to allow account creation based on a whitelist of email domains.
In 2025 Rogue Scholar changed how it hosts the InvenioRDM repository platform. All required services are hosted by a dedicated server from Hetzner and hosted in Germany, and S3 object storage also by Hetzner. The Docker images needed to run InvenioRDM are now orchestrated by Coolify, which provides an easy way deploy services based on Docker Compose configuration files to the cloud. In the next few months I will work on a one-click install option for InvenioRDM in Coolify, extending the work on InvenioRDM Starter I began in 2024. Before Coolify I used the Kamal tool to orchestrate InvenioRDM. While Kamal was working perfectly fine for my needs, I wanted a solution that is easier to deploy, and this is the case with Coolify.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments, in particular with regards to InvenioRDM.
About 85% of the almost 50,000 Rogue Scholar blog posts are written in English, while German, Spanish and French are the most popular languages other than English \u2013 for example more than 5,000 German-language blog posts have been archived.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-16-um-14.31.12-1.png\")
The language is automatically detected by the browser, but can also be manually changed in the Rogue Scholar footer:
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-16-um-14.35.33-1.png\")
The ten languages used in Rogue Scholar blog posts plus five languages where most of the InvenioRDM user interface (the repository platform that powers Rogue Scholar) has been translated can be selected.
This week the following homepage areas have been translated into German, Spanish and French:
In addition, the frontpage now shows the most recent publications into the selected language \u2013 you can of course always search for content in any language.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-16-um-14.34.49.png\")
Also this week, Rogue Scholar started a weekly newsletter of newly published blog posts in German, Spanish or French, you can sign up here.
With these changes much of Rogue Scholar is now available in multiple languages, especially useful for users who are not comfortable searching and reading content in English. And the architecture changes make it easy to add additional languages going forward.
This is a reminder that work on Rogue Scholar becoming a German non-profit organization in 2026 is slowly progressing. While we work on the legal documents needed to register as a German non-profit, you can express your interest in joining Rogue Scholar as an individual or organization via a form link (Membership Interest Form) from the footer of every Rogue Scholar page.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments, in particular with regards to translations in Rogue Scholar.
This is the January 2026 issue of the monthly newsletter from the Rogue Scholar science blog archive. The newsletter reports on new blogs that have joined the platform, important technical updates in Rogue Scholar infrastructure, community updates, and other news relevant to Rogue Scholar users.
One blog was added in January. Welcome! This brings the number of participating blogs to 190, and the number of archived posts to 49,135 \u2013 Rogue Scholar is close to the big milestones of 200 participating blogs with 50,000 posts!
Perspectives on science, data, and technology that don't fit anywhere else.
Sociology and Political Science, English.
http://chrismarcum.com/marcum-blog/
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in January was on refining Rogue Scholar Identity and Access Management (IAM) via Keycloak, launched in December. Instead of one-time passwords (a six digit code), Keycloak now sends emails with magic links that automatically sign in the user into Rogue Scholar when clicked in the email (the link expires after 15 min).
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-11-um-14.54.13.png\")
Rogue Scholar completed the migration to Coolify Cloud, so that (almost) all services are self-hosted now, making operating Rogue Scholar both simpler and more affordable.
The subject classification using OpenAlex fields, subfields and topic introduced in November was further improved, automatically mapping subfields to OECD Fields of Science (FOS), and providing clickable links to query by FOS, Subfield or Topic, e.g. FOS Computer and information sciences.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-01-29-um-09.45.31.png\")
Work on Rogue Scholar becoming a German non-profit organization in 2026 is slowly progressing. While we work on the legal documents needed to register as a German non-profit, you can express your interest in joining Rogue Scholar as an individual or organization via a form link (Membership Interest Form) from the footer of every Rogue Scholar page.
![\"\"](\"https://blog.front-matter.de/content/images/2026/02/Bildschirmfoto-2026-02-11-um-15.40.36.png\")
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
The Rogue Scholar science blog archive has improved the subject classification of blog posts, using an open approach to subject classification developed by CWTS and OpenAlex, and combining this with the widely used OECD Fields of Science.
Classification of scholarly works into defined subject areas is important for two groups of use cases:
A good example of the first use case is a study recently published by Catharina Ochser and colleagues about scholarly blogs in Germany, and one of the main outcomes was the disciplinary distribution of these blogs:
![\"\"](\"https://blog.front-matter.de/content/images/2025/12/jd-02-2025-0053001.png\")
The majority of the blogs they found were in the humanities and social sciences. One important reason for this is de.hypotheses, the German-language version of the large blog portal Hypotheses, offering hosting and other services to humanities and social sciences blogs.
Five years ago in the Make Data Count project, we were interested in the subject classification of research data, as usage of datasets is highly dependent on community size and community standards. One big problem was that the standardized subject classification of datasets was mostly lacking, so we started making this easier for DataCite members: push for the OECD Fields of Science as the standard subject area vocabulary that can be mapped from other vocabularies, and show the OECD Fields of Science as a facet in the search results of DataCite Search.
The second use case is primarily driven by readers. After reading an interesting science blog post, they might want to read more posts with similar content. Or they have an idea about the content they are looking for, but it is poorly described using keywords. Blogging platforms have long provided this functionality, but have been limited to a single blog and have used old technology, rather than semantic search.
Rogue Scholar launched OECD Fields of Science subject area support for every blog post in February of this year. While this was a good start, there were three significant shortcomings:
When submitting a blog to be archived by Rogue Scholar, one of only a few questions I ask blog authors is for the subject area of their blog using the OECD Fields of Science list. Similar to the subject classification of journals rather than journal articles (which, unfortunately is more often done), this approach has limitations, in particular for multidisciplinary blogs/journals. One prominent example in Rogue Scholar is the SVPOW blog, which has been writing about sauropods since October 2007, but also frequently publishes about Open Access.
The six top-level and 42 second-level subject areas in the OECD Fields of Science are good enough for many high-level questions, such as R&D spending or doctoral students per country and/or over time. But for more fine-grained questions, more granularity in the subject classification is needed. There are for example almost 20,000 social science and almost 5,000 computer and information science blog posts in Rogue Scholar, which give a very broad picture.
Using a more granular subject area classification applied to each publication is a major challenge, as can for example be seen in the major curation effort needed to classify the biomedical literature into Medical Subject Headings (MeSH) at PubMed.
The third major limitation is that publishers can't easily distribute subject classification information, as Crossref metadata can't hold that information. DataCite is better in this regard, but unfortunately subject area metadata is still only included in a small fraction of DataCite DOIs.
For Rogue Scholar, I always envisioned a semi-automated approach based on machine learning and human curation of edge cases. Having the full-text of all blog posts available in machine-readable form makes this possible, but picking the right subject area classification and building a classification workflow from scratch are still major challenges.
The ideal subject area classification is more granular than OECD, covers all scholarly disciplines, is widely used (and documented), and openly available. My main candidate five years ago, when I did the subject area work for DataCite, was the Australian and New Zealand Standard Research Classification (ANZSRC) Fields of Research with more than 1000 subject areas, and used for example by the repository Figshare. The German Research Foundation (DFG) Subject Areas used by the Registry of Research Data Repositories (re3data), on the other hand, have some limitations, primarily that they were developed for a different use case - panels for grant evaluation.
The CWTS/OpenAlex classification announced in January 2024 with about 4500 topics categorized into subfields and fields that map to the All Science Journal Classification (ASJC) is a very attractive alternative, especially since OpenAlex uses it, makes an open API available, and more importantly makes the pre-trained model available for download with an Apache 2.0 open source license. This makes it straightforward to run the classifier on Rogue Scholar infrastructure and not depend on external APIs with associated costs and rate-limits.
In January 2025, OpenAlex published a blog post describing the mapping of OpenAlex subfields to the OECD Fields of Science, providing a bridge to the widely used OECD classification.
![\"\"](\"https://blog.front-matter.de/content/images/2025/11/3849275338_dec5fa17a1_k.jpg\")
Like many interesting collaborations, a conversation with Nees Jan van Eck from CWTS at a Dagstuhl seminar on Open Scholarly Information Systems in September was the trigger for me to finally start the work on using machine learning to classify Rogue Scholar blog posts into subject areas.
Two months later I now have a classifier running with the CWTS/OpenAlex model provided via Hugging Face, and integrated into the Rogue Scholar infrastructure.
The InvenioRDM repository platform that powers Rogue Scholar has built-in subject classification with the OECD Fields of Science since v6.o released in August 2021, but this can easily be extended with custom vocabularies. Rogue Scholar now has these four hierarchical OpenAlex vocabularies built in:
![\"\"](\"https://blog.front-matter.de/content/images/2025/12/subject_areas.png\")
Any InvenioRDM instance can install these vocabularies in yaml format and make them available in the record editor and API, and I will also include them in the InvenioRDM Starter pre-built Docker image.
I have used the subfields (that map to ASJC) to classify all participating blogs, and updated the Rogue Scholar submission form to use the OpenAlex subfields instead of the OECD Fields of Science. Please fill out the form (also linked from the footer of every Rogue Scholar page) if you don't agree with the suggested subfield (multiple subfields are not yet supported).
Rogue Scholar has created communities for each of the 252 subfields, and has reused existing OECD subject communities where it makes sense. It will take a few weeks before all blog posts are properly sorted into these new OpenAlex subfield communities.
![\"\"](\"https://media.tenor.com/xFRvGx7feb4AAAAC/sorting-hat.gif\")
The OpenAlex classifier is now running in the Rogue Scholar infrastructure and has started classifying using the blog post title and abstract (as the model was trained with this information). After classifying a few hundred blog posts, I found a score of at least 0.25 to be a good cutoff. With lower scores the number of false-positives was too high, whereas with higher scores (e.g. 0.40) the number of false-negatives was too high. More systematic testing of cutoff scores will follow in a few months, but it is clear that more than half of the classified posts currently have a score that is too low, and false-positives have to be removed. An example of the latter is a post about the Barcelona Declaration on Open Research Information that was classified as topic Optics and Image Analysis with a score of 0.40.
What I found so far is that non-English language posts classify well (14% are in other languages), and that posts that can't be classified are often not sciency enough, e.g. they talk about an upcoming webinar or general tools or techniques.
The OpenAlex subfield(s) are then mapped to OECD Fields of Science using the mapping from OpenAlex to link Rogue Scholar blog posts to systems using that subject classification.
The subfield of the blog and the topic of the post, together with the corresponding topic subfield(s) and OECD Fields of Science (FOS) are shown in the sidebar of the Rogue Scholar post \u2013 again it will take a few weeks to generate this information for all posts.
![\"\"](\"https://blog.front-matter.de/content/images/2026/01/Bildschirmfoto-2026-01-29-um-09.45.31.png\")
The post is also added to the corresponding subfield communities, in the example above Information Systems and Conservation. The community ID corresponds to the OpenAlex subfield ID (here 1206), and OpenAlex IDs can also be used in Rogue Scholar Search:
![\"\"](\"https://blog.front-matter.de/content/images/2026/01/Bildschirmfoto-2026-01-29-um-09.54.25.png\")
Similarly you can click on the OECD FOS or Topic in the sidebar to trigger a Rogue Scholar Search by that OECD FOS or OpenAlex Topic, respectively. In the latest version, search results are automatically sorted by publication date rather than best match.
![\"\"](\"https://blog.front-matter.de/content/images/2026/01/Bildschirmfoto-2026-01-29-um-09.51.30.png\")
As I mentioned before, the next few months will be spent updating the subject classification data, improving the automatic subject classification, and generally respond to user feedback. I am also happy to share this work \u2013 both the OpenAlex subject vocabularies and self-hosted machine learning model \u2013 with other InvenioRDM instances.
For community validation I wrote an integration into the Rogue Scholar Slack (join via this link) using the workflow automation service n8n, also self-hosted on Rogue Scholar infrastructure.
![\"\"](\"https://blog.front-matter.de/content/images/2025/11/Bildschirmfoto-2025-11-26-um-13.19.44.png\")
![\"\"](\"https://blog.front-matter.de/content/images/2025/11/Bildschirmfoto-2025-11-26-um-13.24.22.png\")
As this might also be of interest to other repositories running the InvenioRDM software, I also built a similar integration into the Inveniosoftware Discord server, testing the classification with a few Zenodo records:
![\"\"](\"https://blog.front-matter.de/content/images/2025/11/Bildschirmfoto-2025-11-26-um-13.37.03.png\")
As I am using the same metadata and classifier model as OpenAlex, I can compare their classifications \u2013 all Rogue Scholar posts should be archived in OpenAlex. The example post Reinvestigating the reported transition state structure of a concerted triple H-tunneling mechanism was classified into topic Chemical Reactions and Isotopes by both OpenAlex and Rogue Scholar. Also interesting is a systematic comparison of abstracts and full-text for classifier input, preliminary testing didn't show much difference, and the model was trained with abstracts.
Finally, this approach (OpenAlex subfield and topic classification and pre-trained model made available with an open license) can also work for other Crossref members interested in subject classification, and might convince Crossref to add subject area classification to the metadata schema, something the recently launched Crossref Metadata Advisory Group is thinking about.
In 2026 Rogue Scholar will use this new infrastructure to address important use cases around Understanding and Discovery.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments regarding this work, and join the Rogue Scholar Slack if you want to provide feedback on concrete classifications.
Like most digital infrastructure nowadays, Rogue Scholar and the underlying digital repository platform InvenioRDM are built and deployed using Docker containers. There are many ways to orchestrate (manage) Docker containers; two common options are Docker Compose for local development and Kubernetes for production infrastructure.
The problem is that for many small to medium-sized services \u2013 including Rogue Scholar \u2013 Docker Compose is too simple and Kubernetes is too complicated. Many alternatives cover that middle ground, in particular from the huge cloud providers (hyperscalers) such as AWS (Elastic Container Service or ECS), or smaller cloud providers such as Fly.io, where Rogue Scholar initially launched in 2023.
There are also multiple options to self-host Docker containers on your own infrastructure, rather than with a public cloud provider; you lose convenience but gain control over your data and save money. Last June, Rogue Scholar took that path, relaunching on our own servers (rented from Hetzner) and using the Kamal container orchestration tool.
Kamal has become popular since it was released in September 2023 and has worked really well for Rogue Scholar. But it falls short in helping with one important goal I have for Rogue Scholar and the underlying InvenioRDM repository platform: provide the best deployment alternative to Docker Compose and Kubernetes that makes it easy to get started with InvenioRDM in production.
One recurring theme in my work on scholarly infrastructure over the last 15 years has been KISS (keep it simple, stupid). Scholarly infrastructure is more often than not painful to use. Not only for scholars who want to create and consume scholarly content, but also for the people and organizations who want to help scholars work with scholarly content.
This is obviously a huge topic, and today I want to focus on one small aspect of it: it should be straightforward to configure and deploy a digital repository. For this I started the InvenioRDM Starter project in 2024.
![\"\"](\"https://blog.front-matter.de/content/images/2026/01/Bildschirmfoto-2026-01-22-um-10.32.53.png\")
The initial focus was on the difficulties of building and configuring Docker images. InvenioRDM Starter provides pre-built Docker images and a Docker Compose configuration file that allows everyone to have InvenioRDM running on their local computer in less than 30 min.
Today, I am announcing the next big step for InvenioRDM Starter: make it straightforward to deploy InvenioRDM Starter to production within 60 min with the help of Coolify. Coolify is an open-source and self-hostable alternative to public cloud providers. It works seamlessly with Docker Compose, including the InvenioRDM Starter compose file, and you can use your own server(s) and/or dedicated servers you rented. Coolify is easier to use than Kamal as it provides a web interface instead of relying on YAML configuration files. And Coolify provides a growing list of services that can be installed with one click.
![\"\"](\"https://blog.front-matter.de/content/images/2026/01/Bildschirmfoto-2026-01-22-um-11.07.33.png\")
Currently this means additional services you need for your InvenioRDM instance, in the case of Rogue Scholar for example Keycloak identity and authentication management, Listmonk mailing list, Plausible web analytics, and n8n workflow automation \u2013 all open source software that can be self-hosted.
Going forward I want to write an InvenioRDM one-click service for Coolify so that installing InvenioRDM as production service becomes as straightforward and affordable as installing other popular open source software, e.g. the Ghost blogging platform that powers this blog.
Please reach out via Slack, email, Mastodon, or Bluesky if you have any questions or comments regarding Coolify or InvenioRDM Starter.
One blog was added in March. This increases the number of participating blogs (after adjusting for retired blogs) to 186 , the number of archived posts has grown to 49,606 \u2013 Rogue Scholar is getting closer to the big milestones of 200 participating blogs with 50,000 posts!
Library and Information Sciences, English.
https://orion-dbs.community/blog/
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in March was on infrastructure improvements. The monitoring of the Rogue Scholar infrastructure was improved by deploying a self-hosted observability platform for logs, metrics and errors with dashboards and alerting using the Grafana open source platform:
Th dashboard for key metadata metrics initially released in March 2025 was improved visually and launched for communities, including blog communities:
This makes it much easier for readers to get an overview for each blog participating in Rogue Scholar, and for blog authors to see gaps in metadata coverage that they can improve.
This week the blog self-management in Rogue Scholar was improved, enabling blog owners to update all relevant blog metadata.
The technical updates mentioned above are part of an effort to align Rogue Scholar better with the InvenioRDM repository platform. This will make it easier in the long run to sustain and update Rogue Scholar, as an increasing proportion of the required functionality is built into InvenioRDM and developed and used by other repositories.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
Rogue Scholar uses the community functionality of the InvenioRDM repository platform to manage science blogs and their associated blog posts. In the community settings one can configure basic metadata such as blog name, description, logo and URL.
By default this information is automatically extracted from the blog feed (RSS, Atom or JSONFeed), and since July 2025 blog authors can change these settings directly in Rogue Scholar, e.g. because their blogging platform makes it difficult to add a logo or description.
Starting this week, more information can be updated: besides the required blog feed URL the optional ISSN and the main OpenAlex subfield of the blog. This is provided by community custom fields, and the subfield uses the OpenAlex subfield vocabulary built into Rogue Scholar and also used for blog post classification.
Some metadata still can't be changed by blog authors: the license (CC-BY) as it is fixed for all Rogue Scholar content, the Rogue Scholar join date, blog language, feed format (Atom, RSS or JSONFeed) and blogging platform (WordPress, Blogger, etc.) as they are determined automatically, and the DOI prefix, as this can't be as easily changed as other metadata.
InvenioRDM community members can have the roles reader, curator, manager and owner, and only owners have access to these settings. For a blog author to become an owner I have to send out an invitation to an existing Rogue Scholar account, which since the beginning of the year means an account managed by a self-hosted Keycloak instance.
This work integrates with another change to blog communities introduced last week, community dashboards:
The dashboard shows the metadata from the community form, in the case of language, subfield and blogging platform now as a clickable link to all posts in Rogue Scholar with that blogging platform, etc.
What community owners still can't do is adding or updating blog posts as this is done automatically via their blog feed. And adding a new blog to Rogue Scholar still requires filling out a submission form (at the bottom of every Rogue Scholar page), as blog authors have to agree to provide the full-text in the blog feed with a CC-BY license, submissions are not automatically accepted, and minimal one-time setup work is required.
Please use Slack, email, Mastodon, or Bluesky if you are a blog author with Rogue Scholar and want to become the owner of your blog community, or have any questions or comments.
About 12 months ago Rogue Scholar introduced dashboards with summary statistics of the science blogging archive. At the end of 2025 these dashboards migrated to the Rogue Scholar landing page. Today Rogue Scholar is launching community dashboards \u2013 showing the same summary statistics by community. This feature works with all types of communities: blog, subject and topic communities.
Blog community dashboards help blog authors see major gaps in their metadata coverage, e.g. the percentage of posts that include an author ORCID, ROR for author affiliation, or references. The Upstream blog (where I am one of the editors) for example does a good job including ORCID, ROR and references.
For readers, community dashboards provide an overview of its activities. The Computational Theory and Mathematics Subject Community published two posts in the last month, and the first post in 2005.
As before, these dashboards don't show metadata coverage that is always 100% at Rogue Scholar. All nearly 50,000 blog posts have an open license (CC-BY, a few CC0), all posts have an abstract, all posts have a language, and all posts have a subject classification using the OpenAlex categories.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
Running scholarly infrastructure such as Rogue Scholar beyond the initial launch phase \u2013 what is typically called Day 2 operations \u2013 requires the following activities:
As Rogue Scholar is self-hosted infrastructure the natural fit for these activities is self-hosted observability frameworks. One central element is the Grafana dashboard monitoring service health:
The dashboard collects metrics (ViktoriaMetrics) and logs (ViktoriaLogs) from the InvenioRDM respository platform that powers Rogue Scholar, and monitors the service availability from an external uptime monitor. Errors are sent to the self-hosted GlitchTip service using the Sentry integration built into InvenioRDM.
Standard prometheus metrics were added to InvenioRDM using the prometheus-flask-exporter library, and log files for the InvenioRDM web and worker services are generated in structlog JSON format and reformatted into ViktoriaLogs format using Vector.
This new infrastructure not only simplifies monitoring service health, but also makes easier specific investigations, for example exploring the bot traffic Rogue Scholar receives each day:
Future work could include more detailed monitoring of celery workers or a dedicated dashboard for authentication and access monitoring.
The Rogue Scholar observability service is generic enough to be useful for any InvenioRDM repository instance, uses self-hosted open source software, and can be deployed via the included Docker Compose file. The source code is available at https://github.com/front-matter/invenio-observability, for question send an email or reach out via the inveniosoftware Discord Forum.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
This is the February 2026 issue of the monthly newsletter from the Rogue Scholar science blog archive. The newsletter reports on new blogs that have joined the platform, important technical updates in Rogue Scholar infrastructure, community updates, and other news relevant to Rogue Scholar users.
No blog was added in February. This leaves the number of participating blogs to 190, the number of archived posts has grown to 49,342 \u2013 Rogue Scholar is getting closer to the big milestones of 200 participating blogs with 50,000 posts!
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in February was on translations \u2013 the Rogue Scholar website is now translates into 11 languages with blog posts in nine languages. The translations still need lots of refinements but include many areas that were previously not translated, e.g. the sidebar on the homepage, footer, and subject areas. The homepage now shows the most recent posts in the selected language:
As you can see in the screenshot, the landing page now also shows summary statistics that were previously available on a separate dashboard page. All numbers use live data, and clicking on them returns a search result for that data.
Rogue Scholar is powered by the InvenioRDM open source repository software, and last week the InvenioRDM community met for its annual partner meeting in Graz, Austria. Watch the Invenio blog \u2013 also archived in Rogue Scholar \u2013 for a report in the coming weeks. One important topic was the next major release (v14) of the software in June. Rogue Scholar is running the latest pre-release version of the software \u2013 see the Rogue Scholar footer for the current version.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
This week the InvenioRDM community meets in Graz for five days to discuss the open source repository platform. Front Matter has been part of the InvenioRDM community since August 2021, and Rogue Scholar relaunched on the InvenioRDM platform in October 2024.
On the first day of the workshop we will have a session where InvenioRDM instances can present the highlights of what they are doing, e.g. customizations of the InvenioRDM platform in their repository instance. Here are some of the special Rogue Scholar features launched since the last workshop March 2025 in Hamburg that I want to highlight:
Rogue Scholar registered content with Crossref since it launched. In the last quarter of 2025 I worked on integrating the functionality directly into the InvenioRDM platform. This is a drop-in replacement for DataCite DOI registration, or Crossref and DataCite DOIs can be registered in the same instance. There is more need for discussion at the workshop around advanced features \u2013 e.g. multiple DOI prefixes by InvenioRDM community - but I hope we can finally merge the GitHub pull-request so that Crossref DOI registration is available for all InvenioRDM instances in the upcoming v14 release this summer.
Major releases of the InvenioRDM software happen once a year in the summer, and v14 is on track to be released in a few months. As in previous years, there will be a workshop session dedicated to release planing. Since a few weeks ago, Rogue Scholar is running the latest pre-release version of the software, and is showing the version number in its footer.
And as of last week, Rogue Scholar is powered by Python 3.14, which was released in October and since been added to all tests of Invenio packages.
Also in the footer is a menu to select the repository language. Rogue Scholar has content in currently 9 languages, plus two more translations (Czech and Swedish) for a total of currently 11 languages. The translations not only cover the user interface, but also dedicated text areas (such as the footer, FAQ and search guide), vocabularies and formatted citations.
There are still gaps in the translations, which then fall back to English, but I hope to have them sorted out for the v14 release.
A new feature I launched in January is an update to the \"Recently added content\" section on the homepage \u2013 this section now only shows recently added content in the selected language:
In the last quarter of 2025 I did extensive work on adding the OpenAlex subject classification to Rogue Scholar. The functionality is documented in detail elsewhere, so I focus on the main features here:
The dashboard showing key summary statistics launched last year is now integrated into the Rogue Scholar homepage. The dashboard uses existing OpenSearch queries to show live data, clicking on a number shows the underlying data. The dashboard doesn't use any external dependencies.
The only special number on the dashboard is cited posts, as this depends on the Crossref Cited-By service and a custom integration. And the dashboard is of course also translated into other languages.
The InvenioRDM platform provides a lot of functionality for managing local user accounts, and for integrating with external services. Still, there are two challenges:
My conclusion for Rogue Scholar was that I wanted to a migrate to a single Identity and Access Management (IAM) external service that should be open-source and self-hosted. After exploring several options I ended up deploying Keycloak as an IAM service. Commercial IAM services quickly become costly once your number of user accounts grows, and many open source IAM solutions either were too limited for the features I needed for Rogue Scholar, or felt too complicated to configure and deploy.
My self-hosted Keycloak instance uses a keycloak-invenio Docker image with several Keycloak extensions that add functionality I needed:
I can now enforce two policies with user accounts in Rogue Scholar:
Over the coming months I will fine-tune Keycloak both in terms of functionality and user interface. One feature already developed but not needed for Rogue Scholar is enhancing the magic-link extension to allow account creation based on a whitelist of email domains.
In 2025 Rogue Scholar changed how it hosts the InvenioRDM repository platform. All required services are hosted by a dedicated server from Hetzner and hosted in Germany, and S3 object storage also by Hetzner. The Docker images needed to run InvenioRDM are now orchestrated by Coolify, which provides an easy way deploy services based on Docker Compose configuration files to the cloud. In the next few months I will work on a one-click install option for InvenioRDM in Coolify, extending the work on InvenioRDM Starter I began in 2024. Before Coolify I used the Kamal tool to orchestrate InvenioRDM. While Kamal was working perfectly fine for my needs, I wanted a solution that is easier to deploy, and this is the case with Coolify.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments, in particular with regards to InvenioRDM.
About 85% of the almost 50,000 Rogue Scholar blog posts are written in English, while German, Spanish and French are the most popular languages other than English \u2013 for example more than 5,000 German-language blog posts have been archived.
The language is automatically detected by the browser, but can also be manually changed in the Rogue Scholar footer:
The ten languages used in Rogue Scholar blog posts plus five languages where most of the InvenioRDM user interface (the repository platform that powers Rogue Scholar) has been translated can be selected.
This week the following homepage areas have been translated into German, Spanish and French:
In addition, the frontpage now shows the most recent publications into the selected language \u2013 you can of course always search for content in any language.
Also this week, Rogue Scholar started a weekly newsletter of newly published blog posts in German, Spanish or French, you can sign up here.
With these changes much of Rogue Scholar is now available in multiple languages, especially useful for users who are not comfortable searching and reading content in English. And the architecture changes make it easy to add additional languages going forward.
This is a reminder that work on Rogue Scholar becoming a German non-profit organization in 2026 is slowly progressing. While we work on the legal documents needed to register as a German non-profit, you can express your interest in joining Rogue Scholar as an individual or organization via a form link (Membership Interest Form) from the footer of every Rogue Scholar page.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments, in particular with regards to translations in Rogue Scholar.
This is the January 2026 issue of the monthly newsletter from the Rogue Scholar science blog archive. The newsletter reports on new blogs that have joined the platform, important technical updates in Rogue Scholar infrastructure, community updates, and other news relevant to Rogue Scholar users.
One blog was added in January. Welcome! This brings the number of participating blogs to 190, and the number of archived posts to 49,135 \u2013 Rogue Scholar is close to the big milestones of 200 participating blogs with 50,000 posts!
Perspectives on science, data, and technology that don't fit anywhere else.
Sociology and Political Science, English.
http://chrismarcum.com/marcum-blog/
The a backlog of new blog submissions is still not resolved, so please be patient. You can always reach out via Slack, email, Mastodon, or Bluesky to ask about the status of your submission.
One focus of the technical work in January was on refining Rogue Scholar Identity and Access Management (IAM) via Keycloak, launched in December. Instead of one-time passwords (a six digit code), Keycloak now sends emails with magic links that automatically sign in the user into Rogue Scholar when clicked in the email (the link expires after 15 min).
Rogue Scholar completed the migration to Coolify Cloud, so that (almost) all services are self-hosted now, making operating Rogue Scholar both simpler and more affordable.
The subject classification using OpenAlex fields, subfields and topic introduced in November was further improved, automatically mapping subfields to OECD Fields of Science (FOS), and providing clickable links to query by FOS, Subfield or Topic, e.g. FOS Computer and information sciences.
Work on Rogue Scholar becoming a German non-profit organization in 2026 is slowly progressing. While we work on the legal documents needed to register as a German non-profit, you can express your interest in joining Rogue Scholar as an individual or organization via a form link (Membership Interest Form) from the footer of every Rogue Scholar page.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments.
The Rogue Scholar science blog archive has improved the subject classification of blog posts, using an open approach to subject classification developed by CWTS and OpenAlex, and combining this with the widely used OECD Fields of Science.
Classification of scholarly works into defined subject areas is important for two groups of use cases:
A good example of the first use case is a study recently published by Catharina Ochser and colleagues about scholarly blogs in Germany, and one of the main outcomes was the disciplinary distribution of these blogs:
The majority of the blogs they found were in the humanities and social sciences. One important reason for this is de.hypotheses, the German-language version of the large blog portal Hypotheses, offering hosting and other services to humanities and social sciences blogs.
Five years ago in the Make Data Count project, we were interested in the subject classification of research data, as usage of datasets is highly dependent on community size and community standards. One big problem was that the standardized subject classification of datasets was mostly lacking, so we started making this easier for DataCite members: push for the OECD Fields of Science as the standard subject area vocabulary that can be mapped from other vocabularies, and show the OECD Fields of Science as a facet in the search results of DataCite Search.
The second use case is primarily driven by readers. After reading an interesting science blog post, they might want to read more posts with similar content. Or they have an idea about the content they are looking for, but it is poorly described using keywords. Blogging platforms have long provided this functionality, but have been limited to a single blog and have used old technology, rather than semantic search.
Rogue Scholar launched OECD Fields of Science subject area support for every blog post in February of this year. While this was a good start, there were three significant shortcomings:
When submitting a blog to be archived by Rogue Scholar, one of only a few questions I ask blog authors is for the subject area of their blog using the OECD Fields of Science list. Similar to the subject classification of journals rather than journal articles (which, unfortunately is more often done), this approach has limitations, in particular for multidisciplinary blogs/journals. One prominent example in Rogue Scholar is the SVPOW blog, which has been writing about sauropods since October 2007, but also frequently publishes about Open Access.
The six top-level and 42 second-level subject areas in the OECD Fields of Science are good enough for many high-level questions, such as R&D spending or doctoral students per country and/or over time. But for more fine-grained questions, more granularity in the subject classification is needed. There are for example almost 20,000 social science and almost 5,000 computer and information science blog posts in Rogue Scholar, which give a very broad picture.
Using a more granular subject area classification applied to each publication is a major challenge, as can for example be seen in the major curation effort needed to classify the biomedical literature into Medical Subject Headings (MeSH) at PubMed.
The third major limitation is that publishers can't easily distribute subject classification information, as Crossref metadata can't hold that information. DataCite is better in this regard, but unfortunately subject area metadata is still only included in a small fraction of DataCite DOIs.
For Rogue Scholar, I always envisioned a semi-automated approach based on machine learning and human curation of edge cases. Having the full-text of all blog posts available in machine-readable form makes this possible, but picking the right subject area classification and building a classification workflow from scratch are still major challenges.
The ideal subject area classification is more granular than OECD, covers all scholarly disciplines, is widely used (and documented), and openly available. My main candidate five years ago, when I did the subject area work for DataCite, was the Australian and New Zealand Standard Research Classification (ANZSRC) Fields of Research with more than 1000 subject areas, and used for example by the repository Figshare. The German Research Foundation (DFG) Subject Areas used by the Registry of Research Data Repositories (re3data), on the other hand, have some limitations, primarily that they were developed for a different use case - panels for grant evaluation.
The CWTS/OpenAlex classification announced in January 2024 with about 4500 topics categorized into subfields and fields that map to the All Science Journal Classification (ASJC) is a very attractive alternative, especially since OpenAlex uses it, makes an open API available, and more importantly makes the pre-trained model available for download with an Apache 2.0 open source license. This makes it straightforward to run the classifier on Rogue Scholar infrastructure and not depend on external APIs with associated costs and rate-limits.
In January 2025, OpenAlex published a blog post describing the mapping of OpenAlex subfields to the OECD Fields of Science, providing a bridge to the widely used OECD classification.
Like many interesting collaborations, a conversation with Nees Jan van Eck from CWTS at a Dagstuhl seminar on Open Scholarly Information Systems in September was the trigger for me to finally start the work on using machine learning to classify Rogue Scholar blog posts into subject areas.
Two months later I now have a classifier running with the CWTS/OpenAlex model provided via Hugging Face, and integrated into the Rogue Scholar infrastructure.
The InvenioRDM repository platform that powers Rogue Scholar has built-in subject classification with the OECD Fields of Science since v6.o released in August 2021, but this can easily be extended with custom vocabularies. Rogue Scholar now has these four hierarchical OpenAlex vocabularies built in:
Any InvenioRDM instance can install these vocabularies in yaml format and make them available in the record editor and API, and I will also include them in the InvenioRDM Starter pre-built Docker image.
I have used the subfields (that map to ASJC) to classify all participating blogs, and updated the Rogue Scholar submission form to use the OpenAlex subfields instead of the OECD Fields of Science. Please fill out the form (also linked from the footer of every Rogue Scholar page) if you don't agree with the suggested subfield (multiple subfields are not yet supported).
Rogue Scholar has created communities for each of the 252 subfields, and has reused existing OECD subject communities where it makes sense. It will take a few weeks before all blog posts are properly sorted into these new OpenAlex subfield communities.
The OpenAlex classifier is now running in the Rogue Scholar infrastructure and has started classifying using the blog post title and abstract (as the model was trained with this information). After classifying a few hundred blog posts, I found a score of at least 0.25 to be a good cutoff. With lower scores the number of false-positives was too high, whereas with higher scores (e.g. 0.40) the number of false-negatives was too high. More systematic testing of cutoff scores will follow in a few months, but it is clear that more than half of the classified posts currently have a score that is too low, and false-positives have to be removed. An example of the latter is a post about the Barcelona Declaration on Open Research Information that was classified as topic Optics and Image Analysis with a score of 0.40.
What I found so far is that non-English language posts classify well (14% are in other languages), and that posts that can't be classified are often not sciency enough, e.g. they talk about an upcoming webinar or general tools or techniques.
The OpenAlex subfield(s) are then mapped to OECD Fields of Science using the mapping from OpenAlex to link Rogue Scholar blog posts to systems using that subject classification.
The subfield of the blog and the topic of the post, together with the corresponding topic subfield(s) and OECD Fields of Science (FOS) are shown in the sidebar of the Rogue Scholar post \u2013 again it will take a few weeks to generate this information for all posts.
The post is also added to the corresponding subfield communities, in the example above Information Systems and Conservation. The community ID corresponds to the OpenAlex subfield ID (here 1206), and OpenAlex IDs can also be used in Rogue Scholar Search:
Similarly you can click on the OECD FOS or Topic in the sidebar to trigger a Rogue Scholar Search by that OECD FOS or OpenAlex Topic, respectively. In the latest version, search results are automatically sorted by publication date rather than best match.
As I mentioned before, the next few months will be spent updating the subject classification data, improving the automatic subject classification, and generally respond to user feedback. I am also happy to share this work \u2013 both the OpenAlex subject vocabularies and self-hosted machine learning model \u2013 with other InvenioRDM instances.
For community validation I wrote an integration into the Rogue Scholar Slack (join via this link) using the workflow automation service n8n, also self-hosted on Rogue Scholar infrastructure.
As this might also be of interest to other repositories running the InvenioRDM software, I also built a similar integration into the Inveniosoftware Discord server, testing the classification with a few Zenodo records:
As I am using the same metadata and classifier model as OpenAlex, I can compare their classifications \u2013 all Rogue Scholar posts should be archived in OpenAlex. The example post Reinvestigating the reported transition state structure of a concerted triple H-tunneling mechanism was classified into topic Chemical Reactions and Isotopes by both OpenAlex and Rogue Scholar. Also interesting is a systematic comparison of abstracts and full-text for classifier input, preliminary testing didn't show much difference, and the model was trained with abstracts.
Finally, this approach (OpenAlex subfield and topic classification and pre-trained model made available with an open license) can also work for other Crossref members interested in subject classification, and might convince Crossref to add subject area classification to the metadata schema, something the recently launched Crossref Metadata Advisory Group is thinking about.
In 2026 Rogue Scholar will use this new infrastructure to address important use cases around Understanding and Discovery.
Please use Slack, email, Mastodon, or Bluesky if you have any questions or comments regarding this work, and join the Rogue Scholar Slack if you want to provide feedback on concrete classifications.
Like most digital infrastructure nowadays, Rogue Scholar and the underlying digital repository platform InvenioRDM are built and deployed using Docker containers. There are many ways to orchestrate (manage) Docker containers; two common options are Docker Compose for local development and Kubernetes for production infrastructure.
The problem is that for many small to medium-sized services \u2013 including Rogue Scholar \u2013 Docker Compose is too simple and Kubernetes is too complicated. Many alternatives cover that middle ground, in particular from the huge cloud providers (hyperscalers) such as AWS (Elastic Container Service or ECS), or smaller cloud providers such as Fly.io, where Rogue Scholar initially launched in 2023.
There are also multiple options to self-host Docker containers on your own infrastructure, rather than with a public cloud provider; you lose convenience but gain control over your data and save money. Last June, Rogue Scholar took that path, relaunching on our own servers (rented from Hetzner) and using the Kamal container orchestration tool.
Kamal has become popular since it was released in September 2023 and has worked really well for Rogue Scholar. But it falls short in helping with one important goal I have for Rogue Scholar and the underlying InvenioRDM repository platform: provide the best deployment alternative to Docker Compose and Kubernetes that makes it easy to get started with InvenioRDM in production.
One recurring theme in my work on scholarly infrastructure over the last 15 years has been KISS (keep it simple, stupid). Scholarly infrastructure is more often than not painful to use. Not only for scholars who want to create and consume scholarly content, but also for the people and organizations who want to help scholars work with scholarly content.
This is obviously a huge topic, and today I want to focus on one small aspect of it: it should be straightforward to configure and deploy a digital repository. For this I started the InvenioRDM Starter project in 2024.
The initial focus was on the difficulties of building and configuring Docker images. InvenioRDM Starter provides pre-built Docker images and a Docker Compose configuration file that allows everyone to have InvenioRDM running on their local computer in less than 30 min.
Today, I am announcing the next big step for InvenioRDM Starter: make it straightforward to deploy InvenioRDM Starter to production within 60 min with the help of Coolify. Coolify is an open-source and self-hostable alternative to public cloud providers. It works seamlessly with Docker Compose, including the InvenioRDM Starter compose file, and you can use your own server(s) and/or dedicated servers you rented. Coolify is easier to use than Kamal as it provides a web interface instead of relying on YAML configuration files. And Coolify provides a growing list of services that can be installed with one click.
Currently this means additional services you need for your InvenioRDM instance, in the case of Rogue Scholar for example Keycloak identity and authentication management, Listmonk mailing list, Plausible web analytics, and n8n workflow automation \u2013 all open source software that can be self-hosted.
Going forward I want to write an InvenioRDM one-click service for Coolify so that installing InvenioRDM as production service becomes as straightforward and affordable as installing other popular open source software, e.g. the Ghost blogging platform that powers this blog.
Please reach out via Slack, email, Mastodon, or Bluesky if you have any questions or comments regarding Coolify or InvenioRDM Starter.