Skip to content


The future of text is what we dream and what we make–and what infrastructures we can build to use to realise it. Crucially, the future of text can only grow from what the infrastructures of text allow. We cannot write upon dreams alone.

This is why metadata is crucial. After all it is the metadata which makes the data useful, it is the metadata which specifies what the data is–metadata is the infrastructures of how we can interact with textual knowledge.

our approach to metadata :

Our choice then is that we can continue with our current dead flat documents or we can enrich them by writing at the back of the documents what they are. This is low to no cost metadata which is equally low cost to utilise–it just takes institutional muscle to overcome the inertia of the current state of affairs to set this in motion.

If we are to expand our minds further by expanding text, we should look at the foundations of our textual systems and what they can make possible.


Text is mind expanding

When we write, we connect with our own thoughts as though our thoughts are spread out before us on the page, expanding our brains working memory, giving us instant access to all the text we can see.

When we read we connect with minds across time and space. Text is locked in place yet gives us freedom to explore by slowing down and pondering a sentence, or jumping around to different sections, all at will, in a way speech cannot, as speech recedes into the mist of time.

Socrates did not like the lack of interaction with text but of course we can interact with text, just not in the same way we can interact with another person through speech. As Yuval Noah Harari points out, writing was invented to solve different problems from speech. Where speech allows for questioning live, text allows for a different kind of questioning–through reading further–at any point in time. Where speech requires another person to be present, text does not have such a constraint.


Expanding mind expanding text

When we read, we understand the text in ways that digital systems cannot, even the advanced AI systems we see demonstrated in the first quarter of the 21st century. We understand semantics and layout contexts so that we can, for example, instantly understand that the name of a person on the first page, or at the very end, is likely the author of what we are reading. We can also understand what text is a heading, allowing us easy navigation.

One powerful way to expand text would be for such meanings to be made clear to the computer system.

This is what the Visual-Meta approach allows. It takes data describing the document from a manuscript and when the manuscript is exported this data is retained and added as an appendix to the PDF, or whatever other format is used and formats it in the academic standard BibTeX format.

Taking the data from the manuscript is important since there is much information in the manuscript (fx a Word document) which a reader can benefit from. Today though, this data is discarded on export. It is technically possible to embed some of this metadata in resource forks but it is hardly ever done, for reasons I am still learning about.