A Reply to Lorand

Published by Frode Hegland on October 30, 2018

In reply to the comment on http://wordpress.liquid.info/why-work-on-improving-our-tools-for-thought/#comment-4167954257

Lorand, I don’t have a problem with the importance of knowledge graphs as you put it.

I agree that what matters is to find a way to visually encode the author’s intention in such a way that the reader can most effectively critically access it. I completely agree that a visual approach to this is the way to go and I don’t think we disagree that textual components of this are likely going to be very powerful as well.

I have started some work on this for my Author word processor as Dynamic Views and I am working with Christopher Gutteridge and Gyuri Layos on their visual graph systems for jrnl.global

You say that “text is outdated because of the:”
– need of serialization,
– lack of exact definitions and structures,
– missing ability to follow morphing background.

I don’t agree, with the first and third, as my dynamic views system (and all other visual graphs show) but the second point I do agree with but I just wonder what nodes or symbols or statements or ‘things’ we have available to use which are rigidly defined and structured.

If you want to talk with Ted that’s quite a challenge these days. He is extremely busy with his work but if you like I can tell him to have a look at what you suggest if you send me something very specific.

Your angry tone betrays a frustration I can identify with but I can’t see much in terms of specifically what you suggest should be done. Please note, you trash talk linear text but you reply to me using exactly that. Please show me exactly, even with pencil drawings, I don’t mind, how you think you and me should be talking. No hyperbole, just description. I would very much like to engage new and different thoughts when presented clearly.

Watching Our Language

Published by Frode Hegland on June 5, 2018

We discussed naming and terminology on the Doug@50 call yesterday, or rather, I moaned about our lack of language for the work, and here are two pertinent realisations:

The Thing It Is

In Author, when we switched into the dynamic view, which is essentially a graph or a map where the headings become nodes which can be flexibly re-ordered and connected, I initially called the headings ‘nodes’ when in this view since they are indeed nodes when viewed this way. However, for the user, this turned out to make no sense, since they still ‘are the same thing’, it’s just the interaction which has changed. This is important because we need to use current language as a start for where we are going and we need to see this from a new-world perspective, not try hard to shoe-horn old concepts into a new conceptual space. ‘The thing is the thing’, no matter where it is or how it’s viewed.

Scale

The other issue is one of scale. In a conversation today the obvious issue of how some of us are dealing with and are interested in large scale knowledge graphs and some of us, myself included, are interested in extremely small graphs, such as the dynamic view scale where items on the screen can usefully be represented by text and not a cloud or other shape. I’m sure this has been obvious to many for a while, but it just became clear. The large scale knowledge graph work (LSKG?) is more concerned with AI analysis while the small scale work is more concerned with human-visual analysis.

Hyper-Connected

To keep these two quite different projects connected it would be great to find a way to make them interoperable in an active way if possible, or passive if that is all that’s useful and this is where the notion of a hyperGlossary comes in. The idea is to have a way for a user to manually add or define ‘nodes’ to a system which can be used to serve the nodes to other systems (Chris’s notion of a node server). A primary way to do this could be via a hyperGlossary where an author can define the personal meaning of a word or a term for the sake of elucidating the reader (as glosses originally were used) with a short definition, a long definition and then with specific connections to other terms. These connections would then build a basic graph which can be interactively viewed visually or even connected to large scale graphs.

This could be a very useful way of scaling the scale issue.

Flight