Interdependency of notes / Links with two arrowheads

(Alexander Heusel) #1

Dear Forum,

currently im am using TB to make sense out of a paper I read. The paper depicts a schematic for a procedure which consists of three methods which I represent by Notes. The methods are interdependent which I’d like to model with links with two arrowheads. Is that possible, or am I stuck with two links with single arrowhead per interdependency?

Thanks for your advice!


(Mark Anderson) #2

The ‘arrows’ in maps are links and not simple graphical elements as in a drawing app.

Tinderbox links are always directional, with the arrowhead at the destination end of the link. Some customisation of links is possible - see here - though this doesn’t include having arrowheads at both ends of a link.

(Alexander Heusel) #3

Ok, so I didn’t miss something :smile: . I thought it would be so obvious that I just did not see it. Thank you very much for the quick answer!

(eastgate) #4

Of course, you can easily make a pair of links to represent the bidirectional relationship!

(Alexander Heusel) #5

I tried this already but the result was graphically not very pleasing, i.e. clean to me and it didn’t communicate what I wanted to express. The diagram I had in mind was rather prototypical: three ellipses situated in the corner of a virtual triangle connected by double-sided arrows. My solution was to create a prototype for an intermediate node which expresses an interdependency. I placed the node in between the other nodes an connected them accordingly.

Nevertheless: thanks again for your suggestions! :slight_smile:

(Nicholas Sloan) #6

I agree with Alexander about the unfortunate appearance of a symmetrical pair of directional links. The simplest solution seems to me to be an option None in the Connector menu which currently offers only arrow or circle. This must have been suggested before. Is there a reason why it has never been implemented?

(Mark Anderson) #7

I’m not sure I follow. Tinderbox links, by long-standing design, are uni-directional. As suggested above a bi-directional link can be shown by two links in opposing directions. One link between items with no head marker would make no sense in Tinderbox—other than as an element drawn onto the map UI. To date, Tinderbox has avoided the slippery slope of becoming a drawing package; data can easily be exported to vector drawing tools for extra visualisation.

That said, I don’t see any reason (other than confusion to the reader) of a ‘no head’ option for links.

(Nicholas Sloan) #8

Thanks for speedy reply Mark. To repond to your last sentence first, I read that as encouraging news: I would love to see a ‘no head’ option. There should not be confusion to the reader (or at least to the writer) because a) the option would have to be deliberately chosen and b) the nominal direction of the arrow could always be determined by clicking on the relevant note.

Now for your para one: yes, I was aware that it is an article of faith in Tinderbox that arrows are uni-directional, but in life connections are not. Tbx maps may represent many scenarios in which the conceptual connection between two notes is entirely symmetrical. This can be represented with twin arrows, but this is at best a workaround, and visually a dog’s breakfast.

I am too new to Tbx to challenge the orthodoxy any further, but it seems as though a ‘no arrow’ option might be very welcome to many while not offending the principles of others (not least because they don’t need to use it).

(eastgate) #10

The new options for adjusting link placement may make the use of directed links more palatable.

(Pat Maddox) #11

I admit that sometimes I don’t use links because I don’t know which way they should go. Yes, I know that I can use agents to find and create links if I ever need to change them… it’s just the decision itself is enough to make me hesitate and ultimately not do it.

I guess my brain is just broken :slight_smile: or not always powerful enough to grasp the TB paradigm.

That’s for linking directly from one note to another. If I link text to a note, that one is obvious, because it matches how I’ve always linked text using HTML.

(Nicholas Sloan) #12

(I was waiting to reply to this topic again until I had read the Links section of The Tinderbox Way. I’m not sure that has a direct bearing on the issue, but it helps to have a better grasp of the background to hypertext links generally.)

Ok, retreating to first principles, it is a fact of life (as previously stated) that a hypertext link is a connection from one entity to another entity, i.e. a directional link.

It is equally true that inTinderbox, connectors (the visual expression of links) are used to represent a whole range of relationships between notes, some of which are directional (eg parent to child) and some simply equivalent (e.g. sibling to sibling).

The problem seems to be that we only have a directional mechanism for representing all of these relationships, so that if we wish to express equivalence with, for instance, a headless arrow, there is an anomaly between the visual representation which may be true to the relationship but not to the underlying mechanism.

I am aware that there are many other ways of expressing equivalence, both visually (similar appearance, spatial proximity etc) and via other routes of data perception, but I don’t think it is unreasonable to suppose that there are circumstances in which an equivalence connector would be desirable.

As has been suggested, a pair of symmetrical connectors is one answer —and yes Mark, more flexible layout does help, a bit— but while this removes the anomaly between mechanism and relationship and satisfies the symmetrical aspect of equivalence, it may give a visually false message because it still suggests a dynamic reciprocal relationship between the two entities: either a feedback loop or a two-way road. Even a double-headed arrow sharing the same connector line falls foul of this objection.

I would plead strongly for a headless connector option. A headless connector is arguably no more a contradiction of the underlying link than an invisible connector. Another option for purists would be a pair of headless connectors, one of which is made invisible.

(Mark Anderson) #13

As a map is a rendering of that part of the hypertext forming the doc and not a drawing space, how does this single link work in terms of the underlying link base? Or will this single link be a visual affordance for the map UI and not a true link (or pair of links). I ask not because I’m against the notion of a two-way link but simply to understand how it works other than a bit of screen art (whose purpose in that context I do get).

Also, two notes can have multiple links between them (i.e. >2) especially if using differing link types to capture different semantics. I’m not sure where a 2-way link visual kludge fits with that.

(Paul Walters) #14

I’m not opposed to a “headless connector” (because I have no use case for it, and wouldn’t care if it’s an option). But I don’t know what the headless connector means.


What does that mean?

(eastgate) #15

Might composites sometimes be a useful replacement for bidirectional links?

(Nicholas Sloan) #16

My understanding (I may be wrong) is that when a user selects a default Connector Type —say Agree or Disagree— or sets up a new custom type, it is all the same to the underlying link base: a link is a link is a link. The type is of course recorded and available to agents etc, but the semantic significance is known only to the user and not to Tinderbox.

If this is correct, then the proposed “equivalence” connector would be just a link like any other, though the user would understand the significance to be “the connected notes share the property I am interested in”. And just as the user could choose to style “disagree” with a red or dotted line, so they could choose to style “equivalence” with a headless arrow, in other words a plain line. The lack of direction is only confusing if the direction matters, which in this case it does not.

You could call the plain line a visual affordance, but no more so than any other styled connector. I think using twin links to express equivalence is clumsy in this context, and best reserved for more dynamic reciprocal relationships. Simplest is usually best.

@ Paul Walters: sorry if I was unclear. Headless connector = connector line with no arrowhead (or circle). A—B means whatever the user intends it to mean, just as any directional arrow has a meaning that varies with context and user intention.
@ Eastgate: yes composites might be a viable replacement, one of many probably. I am aware that I weaken my case by not supplying you with a clear practical illustration. I’ll work on that.

(Mark Anderson) #17

This reads to me that you’re talking about a map view drawing feature rather than a Tinderbox link. The wider issue is that Tinderbox is a (hypertext) toolbox. Different groups of users employ it in different ways, some overlapping others in discrete areas but all within the same toolbox. A challenge is to ensure seemingly small changes for one group (e.g. visual affordances for map view) don’t have a wider effect (e.g. incompatibility with existing link logic). In ‘just’ a small buff to X, we don’t break Y. Eastgate have been politely studious in preserving balance even if at times it appears to be pushing back against ‘simple’ requests (disclaimer: I’m just a long-term user of the app & Tinderbox community helper, I don’t work for or speak for the app’s maker).

For better or worse, for 18 years Tinderbox has had uni-directional links. To better understand see (here) the data stored for a link, in the links table which is kept discrete from individual note data. In that context, to have a single (screen) link standing for two opposing-direction links (marked via the screen affordance of arrows both ends) would imply a fair bit of re-engineering of various under-pinnings. IOW, not ‘just’ a small tweak to how links are drawn. Maps are a rendering of the underlying hypertext, not a Illustrator or Omnigraffle drawing.

If the desire is to show visually a two-way relationship then, short of a do-over of the link design, a visual drawing object (double ended arrow shape?) is a solution less likely to mess other forms of use of the app.

In case it is not obvious, I use link-based action code quite a lot and with some complexity, thus my concern. In implementing something for maps (and which do think would be useful in that context) we need to be careful not to break some of the under-the-hood features that gives Tinderbox is its power and flexibility … and which some of us rely on in our daily use of this wonderful app.

More on Tinderbox links.

(Paul Walters) #18


Good point. Since the request under discussion has nothing to do with semantic links and is just a drawing feature with no meaning, some drawing / adornment features would be appropriate. And possibly expensive of Eastgate time, so I would understand if they never occur.