The last 4 days have been crazy, overstimulating, mind-blowing. A day in Coventry for a MozFest fringe event, a day in London at the V&A, 2 days at MozFest. 1 more day of art immersion left to go… I’m not entirely sure I’m going to capture it all. It’s after 01:00 on Monday morning and I’m wired to the moon and unable to sleep until I empty a little out of my head. This blog post will hopefully be one of a few…
This week included time at the Open Web for Learning and Teaching Expertise Hub (OWLTEH) event in Coventry. As well as doing a little with my friend Samantha Ahern from UCL I was invited to sit on a panel session with Tony Hirst to talk about open web infrastructures – “Containing the Future“.
I won’t recount the talk. You can watch it for yourself if you are so compelled. Suffice to say that the stuff about containerisation in my view is only interesting in so far as it’s a means to an end, and the end that Tony and I were actually excited about was Jupyter Notebooks.
— openetc (@openetc) October 25, 2018
[Edit: Here’s the YouTube version of the talk now]
I had the pleasure of spending some time with Jim and Tony the night before *ahem* preparing for our talk. I also had some time with Tony on the afternoon of the event itself and he opened my eyes up truly to the breadth of what is possible with Notebooks. I thought I had a pretty good handle on the possibilities for data-handling and stats applications within various of our science and social science courses, but the point that he started showing me code-generated inline music* and notation I fell deep down the rabbit hole.** Genuinely, I learned a lot and very much appreciated it.
I think the poster display of 1990s web pages at the conference must also have acted like some unconscious stimulus as we came to the conclusion that a lot of our excitement around Notebooks was not dissimilar to the same excitement we had in the early days of the web. It was possible then to cut, paste and hack your way to a functioning webpage, based on remixing examples from the HTML source of other pages. Browser capabilities were being built out relatively quickly and it felt like the possibilities were expanding in all sorts of interesting, but still relatively accessible ways. This feels the same.
It’s possible to find an existing Notebook or two and hack your way to something approximating what you want. I started exploring some simple Notebooks for screen-scraping headlines in bulk from news sites and took a deeper dive into the Economist ‘Big Mac Index‘ notebook. I can see in both cases how these are useful in and of themselves, as well as editable examples of how you could approach other similar activities. This is exactly what my colleague Melissa Terras sees the potential for – building out tools that allow non-coders to carry out computational analysis (in her case she’s interested in whether text analysis could be wrapped in a Notebook, with the only requirement for use being some simple initial variable setting – search keywords etc).
At the end of the talk Mia Zamora asked about possibilities for generating writing based on Notebooks. I am lost in the possibilities here. Thankfully Tony has his sh*t together and has the references to hand.
— Tony Hirst (@psychemedia) October 26, 2018
There is also now a newsletter. Sign up.
— Tony Hirst (@psychemedia) October 28, 2018
* I think it’s worth acknowledging that this blog post was written in 2016. 2016? Has nobody been listening?
I mention this a bit in the talk. Noteable is a Jupyter Notebooks service at the University of Edinburgh. It’s a centrally-provided service linked into our VLE via an LTI connection. It’s currently being piloted with a selection of courses by my colleague James Slack. We are interested in talking to people who might be interested in using it as a shared service – please do get in touch…
- More about Noteable here: https://noteable.edina.ac.uk/
- More about our pilot project using it in learning and teaching here: https://thinking.is.ed.ac.uk/noteable/
A Vague Sense of Unease
[Edit: The talk is now available on YouTube]
Wendy Liu‘s talk earlier in the day was very timely. Whilst Tony and I didn’t have time to touch on this – we were getting too excited about Jupyter Notebooks – I retain a sense of unease about a lot of the open source that gives us the infrastructure of the web (including Noteable – which is reliant on Docker and Kubernetes amongst other things). I find difficult to put my finger on this unease, but it goes something like this:
- Platforms tend towards monopoly
- Platforms tend towards enclosure
- Many of the large platform companies release software infrastructure as open source
To what extent might the release of open source infrastructure tools from large platforms be shaping / influencing the infrastructure of the internet over time?
Development of open source technologies from companies like Google or Facebook were paid for with advertising revenue, which was likely generated through the exploitation of data. Can we really say open source web infrastructures are built on principles of fairness, openness, freedom?
Some open source from big platforms (a non-exhaustive list):
- Amazon: https://aws.github.io/
- Google: https://opensource.google.com/projects/explore/featured
- Microsoft: https://opensource.microsoft.com/
- Facebook: https://opensource.fb.com/
- Apple: https://developer.apple.com/opensource/
- Uber: https://uber.github.io/
- Lyft: https://lyft.github.io/
- Airbnb: https://airbnb.io/projects/
- LinkedIn: https://github.com/linkedin
- Twitter: https://twitter.github.io/
(Flying back to Vancouver. Photo taken by me, no rights reserved.)