Skip to main content

A New Kind of Scientific Publishing

October 30, 2023 · Dion Whitehead

Scientific papers are static artifacts. You read them, but that is about the limit of your interaction.

<!-- truncate -→

You can embed code via links into e.g. PDFs, but code, like knowledge, fades into an insurmountable complexity barrier over time:

  • OS versions change, become outdated
  • documentation, already sparse, becomes incomprehensible
  • configuration and input files become versions so old, no-one even remembers

Even papers dealing with purely computational research accumulate entropy over time.

Publishing journal articles as websites is a step in the right direction, but so far there is vast potential unrealized.

A new way of publishing

  • embed running code directly into the “article”
    • “articles” are really a combination of applications/workflows/websites
  • articles run in browsers, which are incentivized more than any other platform to run as much old stuff as possible (”Do not break the internet”), while remaining secure
  • rely on code modules that are open source, and published in sources that are meant to be “forever”

Metapages: publish entire reproducible applications as URLs

https://docs.metapage.io/docs

Metaframes: websites as visualization with editable code

The following is a visualization of a network. The code can be edited by you, and the updated code runs directly in your browser. To edit or copy the code below, click in the top right below:

This is an example of a metaframe: an editable URL, where configuration/code/data is stored in the URL. A metapage is a set of possibly connected metaframes:

Example 2: python and visualization in the browser, completely self-contained, and editable by you

This is discussed in more detail here

Example 3: visualization, user interaction

The following is an embedded metapage application that shows a visual of the sun, and allow you the user to rotate it with your hands (via your webcam):

It consists of three interconnected, communicating metaframes called a metapage. Below is the date flow:

Conclusion

There is a huge space for durable, high information scientific publishing, that can be shared, remixed, and would allow both casual and deep interactions.