I’ve got a file for taking a face to face class online that includes a goal and lots of subtasks. One of those is meetings. I created a composite for meeting notes, but it seems to be behaving oddly and I don’t know how to fix it. Here is the composite in the composite container:
Notice that there is no “assignee” in the text box. I went to the outline version to check it out. Here is that result:
I thought I’d move “Question” to the top where it belongs and I must have assembled the composite in the wrong order somehow.
Notice that when I go back into the map container and select the composite, “assignee” is still not there. So I went back to the outline and “question” was at the bottom again. I once again rearrange it and then back to the map container. Here is the result:
Now “question” is missing from the text box. If I go check, “question” will be at the bottom of the outline list again. Also, if I create the composite using the right-click in the main body of the map, it shows the composite correctly, but when selected one item in the composite will be missing and one of the composite components names changes when I try to enter a title, such as Question 1. I must be doing something profoundly wrong and would appreciate your help.
FWIW, when selecting manually the items are in selection order and not outline order. It may be the case that the auto-selection of all composite items follows an order that is not $OutlineOrder.
Also, doing a simple check in of a non-composite selection in Map view, when selecting 3 items is the third item has text it is shown in the text pane listing. If it has no text it is omitted. It looks like some sort of bug, the exact nature of which I’ve not had time to explore.
Thanks Mark. I’ll try to make the best of it; the composite seemed like a good way to go for this particular task. I’m sure it will all get worked out. I will throw this one away again and rebuild it from scratch and make certain this time that I create them and stick them together in the sequence I want and see if that helps. I was using the outline view to organize them because I saw a post about that being the way to do it, but the post was a couple of years old and perhaps I misread it.
OK, well I’ve tried everything I can think of. Clicking on one of the notes in the composite sometimes selects the one you want (e.g., Question) so the question can be entered, but then things seem to get stuck where no matter which of the notes in the composite one selects it won’t change. For example, I click on Assignee and nothing changes, it stays as When. So if I’ve got one note header correct, I can’t get it to switch to another. I think I’d best find another TB approach for this for the time being.
Although I’ve documented and tested most (all?) of Tinderbox’s features, I’ve never had a real personal need for composites so I may be missing something. I think this is the sort of issue to raise with Eastgate via tech support, if only to bottom out your assumptions vs the design intent of the features. There may be a glitch or it may simply be the feature works in a slightly different way than imagined.
There’s definitely something wonky in the way this works.
I find if I add some text to the notes, then they all appear (with text) in the text panel. The order of the notes in the text panel depends on which of the composited notes I clicked when selecting the composite.
One possibility is to use the fact that prototypes can have children. So, when the prototype is assigned to a note, new child notes are created. So something like a “question & answer” prototype can have the five children listed in your post, and any note that gets that prototype will have five child notes generated.
Did anything ever come of your exchange with Mark B and your use of composites? Did you find they got better?
I’ve just finished two weeks of intensive TB work using composites and I have found them unreliable, by which I mean the source of confusing behavior. For example, I find they seem to change their order in the outline. I cannot say when, in general, but I seem always to find that things are not as I left them. Second, I’ve found them fragile. Quite often an element of the composite detaches from the composite when I have not sought to do this and do not seem to have done something that should cause this to happen. Again, I cannot generalise about the circumstances, but my guess would be that the necessity of command clicking a note to select it for editing may contribute to miscues.
It is bad enough that I am considering abandoning their use for their project despite the appeal of the promised functionality.
Unfortunately Paul’s sound suggestion for an alternative (viz. proto with children) will not work for me because I want the composites precisely as a visual affordance. I want any Map to show them always with the same spatial inter-relation; and any time I select one in any view I’d like the (three) constituent notes to show in the editor window in the same sequential order.
They are a useful way to clone a set of notes in a particular spatial configuration, but so is copy and paste.
It would be good to hear from someone who has made extensive use of composites.
Anyone who experiences issues is always very welcome to report them to tech support.
Outline Order: when an item joins a composite, it moves in outline order to be adjacent to other members. If it didn’t, selecting a composite element in outline or chart view might select items that are offscreen. We try to avoid that.
Composite fragility: this might in principle be an issue with changing scales or with extremes of scale. It’s never been reported at any time or with any version.
Well, as I said, I’m not able to generalise about the conditions when what I have observed occurs, so I’m not well placed to make a useful report yet. Perhaps others will recall seeing something, though not many people on the forum talk about their use of composites–it has seemed to me.
With regard to fragility, no changes of scale are part of what I have been observing because I have not been changing scale.
With regard to outline order, what I have found is that if elements of a composite are ordered in the outline a, b, c, then after many iterations of command clicking among them as part of editing their $Text(s), the outline order will change to something else, e.g. b, a, c. This is similar to what OP observed at the start of this thread, which is why I revived it.
I can’t say I’ve noticed changes in outline order, but i do notice composites are very prone to losing their arrangement. I’ve found touch interfaces (magic mouse/trackpad) overly sensitive. IOW the haptic interface is sending movements I’m not aware I’ve deliberately sent. Of course, the app can’t tell if an input sent is intended or note. Also we ‘see’ in our mind’s eye the obvious outcome of our intend. But, there are a lot of other things. fo instance, should a note, whilst part of a composite allow me to accidentally nest a note rather than it join a composite:
Yet, what if I join a an existing container to a composite (or nest into a composite note using a different view. Here, I feel for the maker and for those trying to test. Just how obvious (in intent) are our movements of items and where is the balance between flexibility for the user vs locking down the range of things the user can do to fence off edge-case behaviours.
Even without that factor, I’ve been surprised composites have no ‘lock’ on them. IOW, once the layout is correct, the composite elements can be locked can’t be re-arranged unless unless the composite is unlocked.
FWIW, I never got back to checking this issue and I’ve only used composites in a slightly different way lately (compiling notes from several different emails, for example). I’ve been super busy with not much time to play and learn, unfortunately. I should be rich I’ve been so busy, rather than just a poor college teacher, LOL.
Testing in v8.2.3, I observe that editing a composite consistently/reproducibly moves that note to the end of the composite set within the outline. Here’s the map (I’m editing the upper compsite):
I’ve no idea as to the cause but I guess there’s some sort of update happening within the composite context that’s moving the note I can reproduce this consistently and in more than one file.
I also notice that Cmd+Clicking a composite always make the whole map move slightly. I assume a redraw is occurring, if only to show the selection marks on the note, but why the overall map should scroll slightly I’ve no idea (it happens no matter how carefully I click to avoid unintended map movement.
Some other UI oddities I’ve seen:
if I click on a composite member in outline, the composite is selected. Switch to map, nothing is selected.
if I command-click on a single composite item in outline view, it is selected. But in map view I’d expect that note to be selected but it isn’t.
It’s not at all obvious how you move a composite. Logical is to click on the composite frame (bounding box) but that does nothing. Instead you click on a note and drag. Now the whole thing moves but given comments about how easy it is to mess up a composite layout, I submit dragging by the bounding box is a more sensible approach.
Band-boxing around a composite doesn’t select the notes. It does cause the composite to be ‘selected’ in that the border is bolded, though it’s a simple and easily missed change. This feeds back to the difficulty in guessing how to move a composite. Note to self aTbRef needs more info on this.
@eastgate, I hope this—or rather the opening part—give you a reproducible scenario to track down the underlying issue. I totally understand how hard it is for you to resolve something you can’t generate.
Thank you very much Mark. The editing behaviour you describe as reproducible I can now see is what I was observing–I had just failed to see it as general. Many of the UI oddities you’ve mentioned I’ve also noticed. Some of this seems to be a kind of inconsistency in how the interactions with composites were thought through. In some cases overloading some ways of interacting is a cause of confusion, for example you cannot command-click to discontinuously select notes in different composites (I’ve flagged this to Mark B). You’re right also about the map redraws. Because I have three notes in my composite, two with a large amount of text visible, and one with a large (decompressed?) JPEG, it could be that a costly redraw is what I have perceived as a lack of snappiness.
At any rate, thanks for adding some observations to the mix.
Part of this stems, in my opinion, from the fact that composites are not first class objects, yet not quite pure visual entities/affordances of the map. It reminds me of the remark in another thread about the similar shortcoming in links. Of course, one difference is that the links are almost always used to convey semantic content, whereas I have sought solely to use composites as “presentational” elements rather than as semantic contributors.
I’m not sure those assumptions quite match the original design intent. If I recall the germ of the idea was simply a quick way of adding a set of notes of pre-defined purpose to a map for rapid data entry - e.g. taking notes in a conference. I don’t think more esoteric issues like being first-class objects was part of the mix.
A general chick-and-egg issue here is we all tend to start from our intuited use/purpose and work backwards, but that’s understandable when the original design intent is not always clearly known. At the same time, over-emphasising the original concept might stop people finding new and productive ways of using features.
As a general point, Tinderbox doesn’t fit comfortably within a single product category. If one’s presumptive benchmark app type is any of the following, confusion is likely to follow: coding environment, mind-mapper, concept networks, drawing tool, ‘everything bucket’. There are plenty more but those are misconceptions that come up tiresomely often.
I agree with you, Mark. I hadn’t meant to say that composites should be first class objects, but the fact that they are not means that interactions that are familiar (or coherent) in relation to notes do not always apply.
I like the idea of composites as extra “sugar” for taking clumps of notes. There is a bit more there though since notes in composites have some attributes that inter-relate them in virtue of being in that composite. I have no ideas or opinion about whether this should be more or less developed.
On your last point, one reason Tinderbox is hard to grok is that it fails to fall into a familiar category. I looked at Tinderbox many times over the years and never went anywhere with it because I could not tell from the website or the demo version what the hell it was. When I finally broke through with it, I realised that for me it was most like Hypercard which was itself sui generis. And Hypercard–now being more than 30 years old I think–is not a familiar frame of reference for many. But Tinderbox has that magical toolbox-for-anything feel that Hypercard had.
Another composite inconsistency in Map view with Text: a composite with two notes, one with text and the other without may show the titles of both notes in the Text view or just one, depending on which is clicked.
If neither note has text, it seems to show only one note title.
