Dear all,
To add just a small technological note to what Dominik rightfully mentioned above, I would like to point that, even if it is not so famous, GitLab offers slightly more features than GitLab. Disclaimer: I am not associated in any way with GitLab, but I am just sharing the conclusions I got after publishing 16 digital resources (dictionaries, dictionary corpus, GRETIL search interface, and a search interface for the Indology mailing list).
Besides the more convenient technical features GitLab has, it features a community edition which can be installed anywhere, by anyone. And this is pure gold for the research projects, especially after the funding is over.
At the Academy of Mainz, Germany, where I currently work, ALL the digital resources we produce are stored in Git-type repository, in a GitLab server that is made available at the land's level, for academic purposes. When I worked with the Heidelberg University we had a GitLab instance, too, for the digital resources we developed. This is also currently the case with the universities or university libraries, respectively, of Bielefeld, Dresden, Paderborn, and I am sure for others in Germany. I think this is a model that can be successfully used.
So, my technical advice would be:
1. Use Git-type repositories for research data of any kind: raw research data (scans, audio files, etc.), processed research data (transcriptions in TEI format, etc.), and published research data (transcriptions in HTML format, static indexes for searching in the browser, audio files in web-friendly formats, images in web-friendly formats, etc.).
2. Separate the raw and processed data to the published data, by storing the published data only in Pages storage and regenerating it from the processed data at any change of the processed data. Do not store data in the publishing format, which is usually HTML.
3. Keep the research data in a repository that is separated to the repository with the static website that presents the digital resource In this way, one can have more views of the data, namely: more publishing formats (HTML, PDF, ePub), linking of the respective digital resource with other resources, in a corpora, etc. Also, one will not have to process again the data for every tiny change of the presentation website.
Key concepts to take away: Git Storage of data, static publishing of data.
Examples of static digital resources that I published (the search is done with an in-browser search engine I made and static indexes):
Best regards,
Claudius Teodorescu