Tinderbox Forum

Another type of note = node

I know this point has been discussed a number of times. It might be a hard problem to implement in the current design. I just cannot stop thinking about it( I hope you don’t mind repeating the old story: this is off the wall after all).

The Plex in PersonalBrain is, however, more pliable than any of the views in Tinderbox. That is, you can scale it to see whatever level of complexity you need. You can even view your data in a quasi outline. The Map View in Tinderbox restricts you to one level of detail at a time. To see more of you information you need to switch to Chart or Outline views.

Steve Zeoli: https://welcometosherwood.wordpress.com/2010/06/08/tinderbox-vs-personalbrain-grudge-match-or-play-nice/

So, the point is to design a new note type. Call the nodes. They are different from the regular notes for they act as normal notes in the map view while they are containers in the outline view. They would have dedicated, system internal, links to make them look different.

Assume outline like this:

Now (regular Containers): we have a map like this:

When containers become nodes

I am not suggesting a new view. No, I want an improvement on the map view. The nodes being variants of Containers: and, one could even be converted to the other (in the ideal world):

The problems here are twofold:

  1. When we convert a container to a “node,” a great many new objects might be added to our view, since containers may contain hundreds or even thousands of items. Some of the newly-added objects might overlap or obscure objects that are already in our map; what happens to them? Some of the newly-added objects might be plopped down on top of adornments that have OnAdd actions; how do we handle that? If we move a daughter “node” in map view, what impact does that have when we convert the root “node” back into a container?

  2. The information hiding implicit in the handling of containers in map view is often desirable, both for clarity and for performance. In actual practice, flattening the hierarchy and representing it with links tends to make link networks that are too dense to comprehend. (There’s more of this in The Tinderbox Way, and will be still more the the forthcoming new edition.)

There’s a lot of research literature in the area. The hyperbolic view used in The Brain was the subject of two competing patents, which may have impeded adoption; that’s one reason we haven’t done more with non-Euclidean views. A competing formalism uses mechanics and simulated annealing for automatic layout, but it (like the hyperbolic view) works best for sparsely-linked, nearly-hierarchical documents.

Some of the newly-added objects might overlap or obscure objects that are already in our map; what happens to them?

The conversion will always expand the sub-notes downwards. The user should be aware of the possible overlaps. So, the conversion should be made (possible) only on the lowest Containers. The Convert command can even be deactivated (you seem to have the technology already to deactivate commands based on the position of the notes in the map) for higher containers. Drag it to the bottom of the map, the Convert command activates.

If we move a daughter “node” in map view, what impact does that have when we convert the root “node” back into a container?

I have no idea on this

representing it with links tends to make link networks that are too dense to comprehend.

I don’t think this is a real problem because the user should know what would be too dense to manage in a single map. TB will give the option: then, it is the responsibility of the user to make good use of it. But, I will be looking forward to the new version of The Tinderbox way: the first version was already a favorite. “Notes are a process for clarifying thinking and for refining inchoate ideas” still resonates with me every day–and, it has shaped me in good ways.

two competing patents
hmm, I am not aware of this one. I don’t know what to say.

TB is already great. We already have great tools to think and work with. Thanks for the reply.

Here are three screenshots taken from an atlas, a book of maps I’ve been making with Tinderbox for several days, just after I discovered your post and while I was wondering how to customize a big picture without TheBrain in order to have a big view.

The first one is an overview of my atlas. The second is a little part of one of its thematic areas. The last one is taken from another container in which I gather several topics, concepts or notions that I link to my main map. In this way, I can view, for each author or each book or article which are the associated concepts. I can also have an overview of internal links between authors or notions.

Your post challenged me! Another illustration of what Tinderbox is capable of: a great flexibility.


Hello Dominique

This is very interesting and I make some similar maps, but probably without what appears to be novel functionality. Can you just walk me through what is going on here?

  1. I notice you have double-ended arrows, but am guessing these are “faked” in some way. How precisely?
  2. I notice you have the little numbers in circles, which I am guessing is a unicode character. Is that right?
  3. Most importantly, what is the basic idea in the second picture? If I follow the central, downward arrow from Goodman’s Ways of Worldmaking, I get to a note symbolised by the number in a grey circle; and that note is linked to a note inside a container? In this way you have “node” indicator which is linked a container that has additional details?

Thanks for sharing your atlas.


Hello David, thank for your feedback.

  1. As you noticed, if there is a part of « fake » in this map, it’s in so far as I’m trying to build an atlas that could compete against TheBrain, even if those two tools do not do the same tasks at all nor in the same way, but this atlas is not yet finished. I’m processing it. I’m aware that Tinderbox excels in gathering and analyzing notes and I think too that it is very useful when it comes to learn by building patterns of knowledge. Therefore, I’d like to build, sample by sample, step by step, a map that would have sense to me considering my position with regard to Knowledge and culture. As if I have to know my position on an atlas: considering the big picture of the world, where am I?
    I can’t explain the presence of the double-ended arrows: maybe an effect of Dance function which I often use? As you can see on the picture below, the same effect appears, but it’s because PHILOSOPHIE ANALYTIQUE is also linked to Goodman, Nelson.

2./3. The little numbers are external links to a container in which I deposit concepts, notions, fields of research and so on. It’s another map, external to the first one.

I use it too in order to build, sample by sample, an atlas of notions. In this way, I can visualize and memorize links between authors or books and notions. For instance, I have to memorize the fact that Nelson Goodman uses the notion of Cadre de référence (reference framework?) in the first chapter of Manière de faire des mondes. No need to build a map with TheBrain: Tinderbox helps me mind mapping in my own way.

Hello Dominique

Thank you for the quick reply. Thank you also for this example. I think I am engaged in a similar project to you which I had called a conceptual map, though I think ‘atlas’ sounds better. The way you have done it will certainly help me to improve how I have been doing it.

One thing I do that may be an extension is that for some areas, where the concepts are dense and many, I have more than one atlas, as it were several possible conceptual cartographies. The exercise of making these is itself illuminating, i.e. part of the thought work. To do this, I have a master store of notes which is replicated as aliases by an agent to each possible map. Because these are aliases and thus cannot have their own children, the agent only replicates the top level. With your approach, I could have a second (or more) level in my master store which would still be accessible from any of my possible maps.

I am not very skilled with controlling the appearance of things in Tinderbox yet. I like the way your links touch the circled number and the way new links emerge from the circled number. Can I ask how you do that? I.e. how are you creating/formatting the circled number container note?

(Cadre de référence is a quite literal translation, viz. frame of reference.)

Thank you.

The numbers are a automatic embellishment for link stubs. Stubs are drawn if the other end of a link - source or destination - lies outside the current map. The general convention seems to be that the incoming stubs are drawn coming into the top of a note and outgoing stubs are drawn leaving the bottom of the note. The number in the circle is the total number of link stub(s). The outgoing count has been around a while but the same for incoming stubs is new to v7.5.0+. So, if you see pictures of maps with stubs without counts these are simply as in older versions to Tinderbox, as the stub counts can’t be suppressed).

In the image below, note ‘two’ has 2 outbound and 2 inbound links. It has 1 inbound and 1 outbound link from/to notes on the same map. It also has 1 inbound and 1 outbound link from/to notes on other map(s).


(see more on link stubs)

Double-ended arrows are are an accidental artefact. If a note has an inbound link arriving at the same point where an outbound link (to the linking note or some other note):


In the top case xx links to yy and vice versa. Due to their relative alignment both links are drawn in the same place and each’s single arrowhead creates the appearance of a two-way link. In the the second case aa links to bb, whilst cc links to aa. the arrowhead of the cc-> aa link again gives the appearance of a two-way aa-bb link.

There have been feature requests for links to be two-way, though this is not possible as at v7.5.2 (i.e. as I write). Although such a request may seem simple, especially to those who use maps as a drawing space, Tinderbox has powerful hypertext underpinnings engineered on single-direction links. Thus such a feature request is more complicated than imagined purely from a ‘drawing’ perspective.

1 Like

Hello Mark

Thank you for stepping in with the explanation, which is helpful. My links have been “flat” so far so I had not seen this functionality–or if I have I forgot. I guess I was thinking of the container as having its own content, as opposed to just containing more notes.

I’ve seen the discussion of double-ended links on the forum and understand the reason for their unidirectionality in Tinderbox. I can live without them, though I’d be glad to have them.

I think the contrast with drawing is misleading. The hypertext origins urge the idea that there are directed, narrative routes through text (I believe). For those of us using notes to describe concepts/ideas, we often think of links as representing relationships between concepts. These are often symmetric, reversible, or interdependent. For example the so-called “biconditional” which represents the relation of “if and only if” has these properties, i.e. it is true when read in both directions; its symbol in logical notation is a two-headed horizontal line, :left_right_arrow:. So there is a semantic content to the double-ended links, but not that which is relevant to hypertext (it seems to me, not being expert in hypertext).

Thanks again for your fast intervention and explanation.

1 Like

To do this, I have a master store of notes which is replicated as aliases by an agent to each possible map.

Great idea! As I divided my main atlas into separated fields of research, I could create an agent that would gather all my notes in one place. But the links would not appear in that place. Would it be possible to create an agent in such a way it gathers every container in order to have a duplicated picture of the whole map? In this way, I would have several pieces of maps and a big one. No matter: I have what I was looking for.

I don’t believe it is possible to do this. Aliases do share all their original’s text links plus its basic links (but only if the alias doesn’t have links of its own). Thus, when you make an agent map of an existing map the links are not shown as they are between originals and your new map only shows aliases.