Review of "Decentralized creation of academic documents using a Network Attached Storage (NAS) server"
This article addresses quite a specific problem: collaborative document editing without third-party servers. The reasons for not wanting to rely on Web servers, even if they're under the control of the document author, is fairly well explained and motivated.
Addressing authoring solutions to MS Word users is diving in the deep end, but a worthy goal as this reaches a lot of researchers. However, I question that the idea of installing software is too complicated for most people.. do you not think it's possible to get the installation process for FW to be as simple as for MSWord?
I realise that this article is not about FidusWriter, but rather collaborative document editing via a NAS, but it would be helpful to explain more about what FidusWriter does and how it is normally run, ie. not on a NAS, and some comparison.
The authors seem to feel strongly about data ownership, mentioning privacy potentially being violated by corporations and governments. It's not clear however that FidusWriter provides a publishing solution, only an authoring solution. That is, once an article is finished it seems that a PDF is exported and turned over to third parties anyway? Relatedly, I would love to know more about how the authors envision this fitting in with the rest of the scholarly communication process, as mentioned in the abstract, particularly peer-review. Could a similar setup be purposed to permit reviewers to control their review contents as well? I assume FidusWriter takes care of access control, citations, formulas and figures, though the article does not state that.
The article mentions that there are other web services for academic authoring, but doesn't name any. I'd be interested to know which ones the authors analysed. Is it only the ones in the footnote? In which case only Authorea and Overleaf are problematic in the ways mentioned for running on third-party servers. I don't really see evidence that the authors have performed a comprehensive search for alternatives in this space. There are certainly clientside document authoring applications which run on personal data stores, such as Laverna which can talk to RemoteStorage servers, and dokieli (which I'm pretty sure the authors have heard of) which can talk to LDP servers, not to mention 'decentralised Google Docs' type things like NextCloud and CosyCloud. It would be worth finding out how easily any of these could be extended to add collaborative editing if they don't have it already, since they have the decentralisation part covered.
Thus, perhaps it's worth comparing running FidusWriter on a NAS to running various generic personal datastores. It seems to me that FidusWriter rather ties the editing application with the storage, which prevents the user from easily switching applications. It's not clear if FW follows a standard protocol for data exchange with the server either. These aren't really the problems of the authors in this context, but it is a downside of a non-standard system even if it's open source and installed on the user's machine. Useful future work might be taking the idea of a NAS forward with other storage/server and client/application options.
That said, I would love to see a docker image prepared to easily set up FidusWriter on any server, if that doesn't already exist!
Limitations are appropriately mentioned as cost and power of NAS devices, though no specifics are given.
Finally, the authors assume a status-quo vision of academic authoring (eg. the "need" for blind peer review) and proceed from there with a decentralisation angle. More interesting would be a paragraph or two about how an approach like this is setting the stage for future advancement in the space. I like the idea of meeting in the middle; addressing immediate problems authors have with their current tooling, whilst laying some foundations for progressing in the direction opened up to us by Web technologies.
Minor comment: what is meant by "Linux boxes" (in quotes)? Are you getting at things like Raspberry Pis?
And I can't help but ask: did you collaboratively author this article using FidusWriter on a NAS? Some screenshots of doing so would add a whole new layer of credibility to your analysis :)