Noting: the Quarto Rich Authoring Environment for Generative Texts

And so it seems that internal discussions are starting to take place about possible Jupyter flavoured production workflows (I wasn’t part of the working group; my 30 page comment response is available to anyone who wants a copy;-). This is led mainly by a unit that has to date had no real experience or apparent interest in the Jupyterverse, what it signifies or where it and related approaches might be heading at a pace faster than the internal one.

One of the comments I keep expecting to hear (and have heard in the past) is about the lack of rich editing tools in the official Jupyter enviornments.

This argument may have held some weight a couple of years ago, albeit ignoring the fact theat there are some extensions for supporting rich editing in official Jupyter UIs, although they do have some “issues”; the classic notebook rich editor, for example, is an HTML editor rather than a markdown editor, so it craps up your markdown as tagged HTML if you make any changes to the text.

But as more and more tools provide the ability to connect to a Jupyter server (for example, VS Code), or edit alternative formats that can be converted to Jupyter notebook .ipynb documents using converters such as Jupytext, the lack of rich authoring environments can no longer be argued. (That said, I now expect arguments to go the other way, the editing environments are too complicated.)

RStudio is an example of one such environment that can be used to edit one of the Jupytext convertable document formats —Rmd (R Markdown) — in a rich editor (see for example Authoring Notebooks from RStudio). VS Code is another. Python code can be executed against a specified Python environment, although I’m not sure whether you can connect to a particular IPython kernel over a Jupyter server connection?

However, I also note more recently the appearance of Quarto, “[a] scientific and technical publishing system built on Pandoc”. The Quarto visual editor which provides “a WYSIWYM editing interface [“What You See Is What You Mean”, apparently] for all of Pandoc markdown, including tables, citations, cross-references, footnotes, divs/spans, definition lists, attributes, raw HTML/TeX, and more”.

Furthermore, “[t]he visual editor also includes support for executing code cells and viewing their output inline”. Currently, the Quarto editor is accessed as an RStudio extension, although it seems as if the editor will be made available as a standalone editro at some point.

For a review of using VS Code for authoring rich content, see for example Authoring Notebooks in VS Code. See also Connect to a remote Jupyter server and VS Code as an Integrated, Extensible Authoring Environment for Rich Media Asset Creation. Note that VS Code can also be used to edit and run Jupyter notebook .ipynb documents.

I’ll try to do a full review over the next week or two…

PS In passing, I note that the same unit leading the “Jupyter production” project also seems to be making a landgrab for content production in the form of “learning developers who will …  take the lead on course content, informing overall course design, and creating effective learning journeys through digital media and learner communications and engagement.”

PPS I do, of course, selectively misquote; the advertised jobs are for “three Senior Learning Developers to join our cross-functional course design and production team working on microcredentials and short courses on FutureLearn” so in the short term the idea is that they get to try things out producing ‘Non-core” materials, rather than materials intended for full qualifications. I note in passing that my department is apparently behind a DataSkills by The Open University bootcamp, funded by the Department for Education, in which learners will “work towards the internationally recognised Microsoft Azure Data Fundamentals (DP-900) and Microsoft Data Associate (DA-100) certifications” using Microsoft Power BI. Good to know. I wonder if we should be changing our own data management course away from open source technologies and instead use these same technologies to bring our various offerings into alignment? And if not, why not? (There are good arguments why we might not want to sully our full qualification modules with vendor certified training content (as some would have it…), but the move to non-academic units creating our “commerical” training (sorry, bootcamp) content is… interesting…