I think what you want is possible but you describe a UI metaphor that doesn't sound like TB, so I guess you're thinking in terms of some other app you looked at.

You describe a note title like "O'Neil _The Hairy Ape_" where I assume you're trying to combine 2 different pieces of information, like:



I've hardly scratched the surface of TB's features but we're already pulling out features.  This isn't Word! Rather than try to describe things within the constraints of past tools, put data in and then look for ways to link it, for instance by adding attribute. 'Books' don't have to be placed in a note whose title is their author.  They could be in a container called "Books" with an 'author' attribute. You could then, for instance sort by Author, DatePublished whatever.

Bottom line, stop worrying about what your last tool didn't let you do.

OK, (quick sketch later) go flat in Map view. Use links and the many affordances of Maps. Previously, I assumed you'd be using Outline view as your descriptions thus far implied a hierarchical point of view. albeit only a few notes, but it tok seconds to turn my earlier sketch around. Either way these were made from the same TBX - though I did promote the books out of the authors to make the Map.

I think - and I don't mean this combatively - you're encountering a common problem most of us share of feeling we have to see (an approximation of) the final structure before we feel confident to more ahead. My early TB use certainly suffered from this. Now I both start piling in information and structuring as I go. Sure, some TB experience teaches you more things you might do at outset but it doesn't stop one 'just' getting started. IOW, eventually one runs out of excuses for not starting or getting stuck in analysis-paralysis.

In parallel with the above in my day job I'm constanly being asked to "just build an example so I can tell you if it's what I had in mind". Problem is, building that is most of the task's work and invariably the answer is "no - do it differently", so pretty much everything is thrown away; wasteful in time if not money. I can't say it often enough, TB's flexibility allows you to side-step that trap and structure as you go avoiding either plan-everything-before-starting-paralysis or only-know-it-works-when-finished-and-then-do-over syndromes.

If an aliases Display expression was intrinsic to the alias and not the original, the 'list' - i.e. note titles - shown in an agent could show different information. Thus if the agent found all books but the $DisplayExpression for the aliases generated was $AuthorName you'd get a list of authors; OK, not a list of unique values but a lit nonetheless. To my mind this also touches on the fact we've now operators to make sets but sets aren't easily visualised in primary views as the information is an attribute - actually I'm wrong, the new v4.5 Maps tools offer a way around this. So, you make a set for $AuthorName of every Book-type note, then use the new map visualisation tools to display that set's data in Map view. By making a set we avoid the duplicates we'd see if using an agent.

In your agent make this the query

($Author!="")

which tells an agent to find notes where: $Author is not (!=) empty ("")

If you have multiple items with the same value for $Author, they will all show up in this agent. There is no simple way for an agent to include one note as a representative for all notes that have the same attribute.

There has been an advanced discussion whose principles might be applied to your problem.

But I don't think you need to go that way. It would be easier as in Mark A's example, to create an author prototype and create author notes from that prototype. Then link each author note to a particular book. Your agent could then simply search for any note whose

($Prototype="Author")

To offer a thought in a related direction: for a similar problem (portraying the structure of fictional works) I've used links and sets and had Tinderbox create Text based on the relationships defined by links.  In other words, to put characters into scenes (this could be books and authors), I created a link named AppearsIn.  I had a map with adornment actions that made AppearsIn links between the character notes and the scene notes.  I defined a custom attribute named SceneSet, which was of type "set".  Finally, Characters had a prototype whose rule was:

$SceneSet = links(this).outbound.AppearsIn.Name; $Text=$Name+" appears in " +format($SceneSet," "," ",", ","");

The rule creates Text, such as this example:

Martin appears in Day Three, After Dawn, Red Horizon


When the container for this note is of sufficient size on the map, the above appears.  With a series of agents, I was able to see on the map the entire Book - Chapter - Scene - Character structure.    The code example is old and doesn't take advantage of new Tinderbox features e.g., the link-making agents.

Mark B.,

I've created an agent with the code you've given me in the rule field. However, no links or files are showing. Do I need a separate command to display the results of the rule?

If so, could you please point me to the correct page of the manual?

Thanks,

~A

For example, the agent's $DisplayExpression might display the set of author names:

    DisplayExpression: $Name+": "+format($MySet,", ")