10 reasons why the National Library of the Netherlands moved its Wikimedia-related publications from SlideShare to Zenodo, and keeps them on Wikimedia Commons

Although Wikimedia and/or SlideShare can meet some of these needs, over the years I gradually found out that they were not really potent enough to meet my ever increasing publication needs for my work to be discoverable and reusable by as many people, institutions and machines as possible, now and in the future, in a durable way, with low risk of their URLs becoming 404s or not findable at all.

Furthermore, the KB is a scientific institution and thus part of the international science and research communities. So as an employee of such an institution, I felt a need to also make my publications more strongly integrated into these communities. Both Wikimedia Commons and SlideShare fail to meet this need, neither of them has a strong user base in academic publishing.

Finally, as a European user of SlideShare, without going into all (legal) details, I felt some stress about having my content hosted on servers located in the US and being under that jurisdiction. This stimulated me to look for strictly EU-based hosting alternatives.

The new needs and insights I developed over these years slowly but surely urged me to change my approach to storing and sharing my content. As the KB has been a part of the Wikimedia community for many years, an important requirement was to keep all Wikimedia-related KB output available through Wikimedia Commons.

However, for the KB content hosted on SlideShare, I needed to find a better place, closer to home, on EU-based servers, with a service meeting as many of my demands as possible.

After doing some (non-exhaustive) comparison, using the Generalist Repository Comparison Chart among others, I picked Zenodo as the best match for my needs.

In summary, Zenodo is a general-purpose EU-hosted data repository built on open-source software that accepts all forms of research output, including presentations, research papers, data sets, research software, reports, and any other research related digital artefact. For each submission, a persistent digital object identifier (DOI) is minted, which makes the stored items easily citable.

My choice for Zenodo was also influenced by my wish to comply with KB’s earlier choice of Zenodo as its preferred repository; employee-generated publications have been stored in its KB community since 2017.

Gradually I started moving my Wikimedia-related content from SlideShare to Zenodo. Along the way I also improved and harmonized the descriptions and other metadata associated with the files. Of course I made sure to submit all my content to the KB community as well. I completed this task in the summer of 2024, making all these presentations, articles, reports, tutorials, videos and other publications now accessible from Zenodo.

1) Sharing a wide variety of file formats, beyond the typical “presentation-oriented” PDF and JPG, such as more “data-oriented” formats like JSON, TSV, Markdown, Jupyter notebooks, or zipped files. Zenodo is very flexible, as it supports uploading files in any format. Wikimedia Commons has a smaller set of supported formats and SlideShare only allows PowerPoint, PDF and Word files.

Example: The Workshop OpenRefine en Wikimedia Commons (19-11-2022) hosts files in tar.gz, pdf, xlsx, txt and json formats.

2) Publishing constituent workshop materials under one unique and persistent identifier. When running a typical data-oriented workshop, it is very convenient – both for the participants and for post-workshop archiving – to publish all materials (slides, notes, guide for participants, csv with data, Jupyter notebooks, Openrefine project files, etc.) as a combined package in a single location under a single durable URI or DOI. This is a real added value of Zenodo, as both SlideShare and Wikimedia Commons allow only one file per upload.

Example: This OpenRefine Introduction Workshop (2023) packages six files (pdf, xlsx, txt, and md) into a single upload.

3) Sharing large files, with personal use cases including video recording of live presentations of hundreds of MBs, or datasets containing large numbers of hi-res images and their descriptive metadata files (1+ GB). SlideShare has a 300MB limit and Wikimedia Commons allows 100MB by default, or up to 5GB when using chunked uploading. Zenodo has a default 50GB file size limit  for each publication, which is more than enough for my needs.

Example: Hosting the 397 MB video Wikidata in de praktijk bij de Koninklijke Bibliotheek, masterclass Wikidata (28-05-2021) is well within Zenodo’s file size limits.

4) Restricting access to files and saving drafts are features in Zenodo that allow editors to work on (drafts of) records in a closed-off environment before releasing them to the public. Wikimedia Commons does not offer this, as every file there is publicly accessible, and under version control right from the start, making it unsuitable for drafting temporary or permanent non-open publications.

Example: I’m currently preparing a Wikidata introduction workshop for Maastricht University and already uploaded some draft materials for this event to Zenodo. When the preparations are finished, I will openly share the files with the participants via https://doi.org/10.5281/zenodo.11224632  

5) Wider licensing options compared to Wikimedia Commons and SlideShare. By default, I publish and share my content under open licenses, either CC-BY, CC-BY-SA or even CC0. This makes my publications suitable for Wikimedia Commons as well as for SlideShare.

However, in some cases, I would like to have a bit more control over derivative or commercial reuses of my works, for which CC-BY-ND and/or CC-BY-NC licensing are appropriate instruments. This makes these kinds of works unsuitable for sharing on Wikimedia Commons, but SlideShare would still be an option, as these more restrictive licenses are allowed there.

Zenodo accommodates all of the above scenarios, as it allows for both very open (CC0) and less open (CC-BY-NC/ND) licenses, as well as over 600 alternative options, both for presentation-oriented and data-oriented publication. So in the vast majority of cases, there is always a suitable license to choose from.

The only exception is the ‘All rights reserved’ case. It is allowed by SlideShare, but not on either open-oriented platform. However, I can use this limitation of Zenodo and Commons as a stimulant to always create publications that are suitable for sharing (be it under a CC-BY-ND/NC clause) and reuse when possible (without the ND clause).

6) Adhering to FAIR principles. Zenodo’s Principles page explains in detail how the repository adheres to the FAIR principles of being Findable, Accessible, Interoperable and Reusable.

This makes Zenodo more FAIR-compliant than Wikimedia Commons. Although all uploads get their own unique persistent M-identifiers (so-called Concept URIs, such as https://commons.wikimedia.org/entity/M144836996), these play a less manifest and significant role than their DOI equivalents in Zenodo. Multilingual findability is limited by the English-oriented category names used to group files.

On the positive side, metadata of each upload is indexed and searchable via the Wikimedia Commons search engine immediately after publishing. Additionally, all files and their metadata can be requested for free via a set of REST-APIs without the need for a user account.

However, interoperability is non-ideal as Commons uses metadata standards that are not necessarily compatible with ontologies in the cultural heritage and humanities domains. Also Commons does not offer support for standards like IIIF, although some attempts have been made in the past. All Commons files can be reused because open (Creative Commons) licenses are mandatory.

Finally, thanks to the Structured Data on Commons effort, Wikidata has been integrated into Commons. This means all FAIR strongpoints of Wikidata (“Wikidata fulfills the most complete degree of FAIRness”) are now exposed to Commons as well.

As for SlideShare, it is a purely presentation-oriented and visual-first environment, with no explicit attention to being FAIR. It does for instance not use unique and persistent identifiers for its uploads, nor does it offer rich and explicit metadata for them. Publications are not linked to open, external vocabularies, and you will need to have an account to be able to download content and their (very basic) metadata.

8) Contributing to international research: The KB is a scientific organization and, as such, part of the international science and research communities. As an employee of such an institution, Zenodo allows me to automatically integrate my publications into scholarly aggregators, such as ORCID, DataCite, OpenAIRE, BASE or FID Benelux, making them more visible for study and research purposes, without any extra effort on my part.

Both Wikimedia Commons and SlideShare are not well integrated into academic workflows and don’t have strong user bases in the scholarly community. By exposing my Wikimedia-related publications to this target group, I hope to raise more awareness of the benefits that Wikimedia can bring to their fields of work.

9) Fully hosted on EU-based servers: Zenodo is powered by the CERN Data Centre, physically located in France. This means it is subject to CERN’s legal status as an intergovernmental organization with immunity from the jurisdictions of the nation courts of the EU Member States (source). This effectively safeguards Zenodo from being taken down due to legal procedures initiated by hostile agents.

Although Wikimedia Commons runs in data centers partly located in the EU (Amsterdam and Marseille), the platform legally falls under U.S. jurisdiction, as its operator, the Wikimedia Foundation, is a nonprofit public charity under U.S. law.

Of course, these differences are hardly surprising because Zenodo was built as a scientific data-focused service, giving metadata the same (or even higher) priority as the object. But for users coming from the visual-first background of SlideShare, where most textual metadata is rather hidden, the differences are striking.

Fortunately, using its REST API and some Python code, it would be very feasible to generate a thumbnail gallery view for all my publications in Zenodo, mimicking the KB Wikimedia presentations & publications archive discussed at the start of this article.

In this article, I describe my journey in transitioning from SlideShare to Zenodo for hosting and sharing my professional Wikimedia-related content, while continuing to maintain KB’s presence on Wikimedia Commons.

As my publishing requirements evolved over time, the need for a more robust, versatile, and research-oriented platform became clear. While SlideShare initially met my basic needs for visibility and discoverability, its limitations—such as a narrow selection of file formats, lack of advanced metadata support, and concerns about long-term accessibility—prompted me to look for a better alternative.