I am thinking a lot these days about my own mortality. That’s not because I am obsessed with death or anything of the sort. It’s more that my health problems are accumulating and I do not know how long my body is going to hold out against them. It might be that I can clear all my current infections. But it turns out that norovirus, which I have had for four months, can be life threatening in the case of immunocompromised patients. Hence I keep thinking: what if it happens? What will the world look like after I am gone? The answer is probably not very different, but with some localized effects on those who care for me. But what about all that I have written, not just in formal academic publications, but on this blog, where I believe that there is matter worth persisting?
Blogging in this context, though, has become a painful and laborious exercise for every post. First, I must write good metadata, including generating a DOI using Martin Fenner’s commonmeta tool. For aesthetic purposes, I must also find a freely licensed image that somehow ties in with the post. This image must then be scaled to the correct size for open graph sharing on social media. I then write the post itself in markdown, which is the most enjoyable part of the whole process. I never use AI to write my blog posts. The act of writing and thinking in public is important to me and is not something that I wish to delegate to a machine or in any way to demean by removing this human connection. That said, I do sometimes dictate the posts because my hands are not in great shape. And this uses machine learning technologies to transcribe my voice. When ready I publish the post. I still get great satisfaction from the instantaneous nature of this publishing form. I just press go and it’s out there. Nobody can tell me not to publish it.
However, this is hardly the end. I want to ensure that my work is digitally preserved in as many ways as possible. The first line of defense on this front is that I am a member of the Rogue Scholar network. Again, an initiative set up by Martin Fenner, Rogue Scholar ingests the content of my posts, so the full text is stored offsite. Rogue Scholar also registers my DOIs so that, in the event of my demise, these can be updated to provide persistent linking to my content. I believe that Martin is planning on setting up a German foundation that will have charitable custody over Rogue Scholar for the foreseeable future, and be responsible for implementing a business model that will allow its persistence. Because, of course, Rogue Scholar is only as persistent as is the will to run it and the capacity of people to spend their time working on it. It is good to know also that the Internet Archive is primed to preserve my work. All of my most recent posts have a Wayback Machine link that I believe is accurate, although it’s based on a speculative slugging process for the URL.
I then upload my blog post in PDF and markdown formats to two repositories: KC Works and Birkbeck’s BIROn. Each of these has a detailed submission flow that involves complex metadata input in order to ensure that the work is discoverable and filed correctly. In short, each of these causes me a lot of pain every time I write a blog post, and substantially adds to the effort of posting. Yet each of them gives me a renewed sense of confidence in my persistent digital afterlife. KC Works is based at Michigan State University. BIROn is based at Birkbeck, University of London. The latter, where I am a Professor in the English department, has just celebrated its two-hundredth birthday. That is to say, it is a persistent institution. And I feel more hope in its continued existence than I do in, for example, Amazon. Indeed, Amazon is just thirty or so years old and AWS 20 years old. Amazon may feel as embedded in our daily lives as the air we breathe, but the truth is, it is simply not as old as most universities and its persistence is also not guaranteed. Of course, the current abominable Labour government in the UK poses an enormous threat even to old and venerable institutions of higher education. They seem not to care in the slightest about the survival of these institutions and are willing to throw away centuries of tradition. I hope, though, that the institutions with whom I have lodged my content can survive. It is also fortunate, of course, that repositories like these are ingested into Core; another third party with a lodged copy of my work.
One of the weakest spots currently is that my blog, which is statically generated using Jekyll, is hosted on AWS. I know Amazon are evil, I should not be using them. However, their infrastructure is just so strong and their pricing so affordable. But the major technological weakness here is that I cannot preload my Amazon account in order to continue paying for their services in the event of my demise. That is, when I die, the most likely thing at the moment is that my blog will last until the end of the month, and then the original canonical source of truth, currently at https://eve.gd, will disappear. I need to fix this, as this is an unacceptably short period for the central block to remain visible. One of the other ways I will fix this is by uploading all of the source for the site to GitHub (owned by Microsoft, 18 years old) under version control so that if worst comes about, the original markdown is available.
This is all a lot of work, just to jot some thoughts on my blog. Indeed, it has a deterrent effect on my writing because I know that I cannot just log in, type what I want and press a button. As a result, one of the ways that I have to incentivize myself is by making my CV generation automatic and dependent on this process. That is, my academic CV is generated automatically from the metadata in BIROn. So, if I want my writing to be documented, I have to follow this process. Good psychological self-incentivizing.
I have also managed to automate some of the initial dissemination process. I have put together an application that automatically posts to BlueSky, Mastodon, and LinkedIn when I write a new blog post. The app also does great threading so that I can write long social media posts and have them automatically split, as needed. Furthermore, this app will perform the resizing and file size limit checking needed for open-graph posting on social media so that there is a nice preview image when I make my posts. This saves me at least a few minutes every time I write a blog post.
What’s missing is an easy way automatically to deposit my blog posts in multiple metadata repositories without having manually to go through the process every time. I used to have an automatic workflow for BIROn. It was based on Selenium and basically worked by knowing where to click in the deposit process each time. Since the repository moved to using Microsoft for its authentication process, though, with two-factor authentication, this has become much trickier. I suppose I might try again to automate this at some point, as it would be a vast life improvement to be able to have this done automatically. I know that my colleague Ian is working on automatic API deposit or ingest for KC Works. Again, if there was a proper API to do this, that would be incredible and would alleviate the speculative and hacky way of doing this through browser automation.
So, why bother? Why not just set up a Substack and whack the material on there, knowing that Substack will probably outlast me, sadly (they seem to make money from hosting Nazi newsletters)? The answer is where I started: because I want my material to persist and I do not trust third-party providers like Substack or Ghost to take any steps to preserve my work. But this highlights a productive tension between the ease of blogging on platforms like Substack or even WordPress and the sadly necessary pain involved in blogging in a way that will be more durable and that, I would hope, will last for longer. Using a static site generator means that my site will never be hacked, most likely, because there is no server-side code to exploit. Depositing multiple copies in many different places works on the principle of LOCKSS: Lots of Copies Keeps Stuff Safe. Ideally, we could automate much more of this process and make it so that your average non-techie academic was able to blog and to have some assurance that their work would be deposited in multiple safe stores and last beyond their lifetime. At the moment, there is a huge chasm between the ease of blogging in these spaces, and the hard graft of what you actually need to do to ensure (or at least to stand a chance of achieving) actively the persistence of your work.