DEVONthink Replicates vs. Tinderbox Aliases

When Tinderbox watches a DEVONthink group it does not import groups that are children of the watched group, it only imports the documents in the top level of the group. So (given the current implementation of watching) you cannot replicate a hierarchy by watching its DEVONthink parent.

Ty so much for the input, however, these agent actions are strictly in list format. Not Helpful!

Ty Mark, and just to be more obnoxious, Keep it has bundles :nerd_face:
I now clearly understand that there is no way for Tinderbox to have a collection of items (deeper than 1 level) and for that item to live in more than one collection at a time, got it.

Sure. I looked up Keep It which is described as ā€œKeep It is a notebook, scrapbook and organizer,ā€¦ā€. I donā€™t think thatā€™s what Tinderbox is but something more like Curio. Itā€™s comparing apples and pears: I donā€™t think Tinderbox should necessarily have these things. If you really need features that arenā€™t in Tinderbox, then email info@eastgate.com explaining your needed feature and your use case for it.

Meanwhile, if there something we can help with using Tinderbox current features, ask away. :grinning:

will you guide me on how to create an agent that would automatically duplicate the original note to live in other container instances and every time I make a change to the original all of the other instances update automatically?

basically the same idea as before but throwing out the idea of tags & aliases; just basic olā€™ duplicates.

Weā€™re going round the same loop againā€¦ I really donā€™t know how else to say it: Tinderbox doesnā€™t have DEVONthink-style replicant features. Youā€™re basically asking for what Iā€™ve already explained isnā€™t possible.

Action code canā€™t create new notes or duplicate existing ones. You can create new aliases by using an action to set the $Container of an alias in an agent so as to move the alias out of the agent (meanwhile next update the alias re-creates the now missing alias). But the latter doesnā€™t solve your problem.

Perhaps if you can step back and explain to us why you need this specific feature (other than being used to having it in some other app) and in order to do what. There is invariably more than one way to do a task and replicants regardless Tinderbox has lots of features some of which should let us get a resolution by a different route.

What is the task youā€™re trying to perform? What is the problem youā€™re trying to solve?

Sorry if not helpful! Just trying to explain what the program does.

It sounds as if youā€™re looking for something that I donā€™t think Iā€™ve seen from any other program

Exactly replicating from one program to another; with full hierarchy intact; automatically updating; and with 0mb size footprint.

If you find a program that does that, Iā€™m sure weā€™d be interested in hearing about it!

No No, I have no idea what I would have done without your input or anyone elseā€™s input. I guess I forgot to sympathize with your intention and I simply answered the result of your input. Shame on me

Iā€™ve spent the better part of the last four years classifying and acquiring all sorts of information Iā€™ve jumped from tool to tool and I am really exhausted, Iā€™m really just trying to find something that works. I donā€™t see anything close to Tinderbox and the idea of readjusting what I have trained my brain for so many years gets me exhausted even more, thatā€™s all. Tinderbox is also a tool that has gained me at every step of the way, it gets really tiring. But by no means, I am very grateful for everyones input here, I really am.

Maybe I really donā€™t know what I am talking about, I did some testing and replicated a some media in a specific Devon database and the external drive showed the same amount data size. Im probably getting the knocks for my ā€œsense of gratefulnessā€ on the other question. Probably even those tags take some amount of data :slight_smile:

As Iā€™ve observed elsewhere, I think that one of the most significant barriers for newcomers to Tinderbox is that they have used other programs. In my experience it can take a while to see past oneā€™s preconceptions and start using Tinderbox for what it actually offers (which is unique) rather than trying to use it like another program. I found this observation by James Fallows was crucial in my adjustment of my thinking. Tinderbox is best used for associations not hierarchies.

1 Like

I revisited out conversation and it seems you did leave out a bit of informationā€¦I thought I was coming across a bit off. But I see the conversation with Mark A. where I did mention something along the ā€œthey have missed the pointā€¦ā€ only to realize that you really were my only solution :slight_smile:

Mark has made it very clear to me that what I was trying to accomplish was simply not feasible in Tinderbox. So I went, up the menu, a bit to the right, crawled down to edit ā†’ duplicates and the result was strange and unstructure so that confused me even more. However, youā€™ve had the holly ā€œCC/CVā€ grail all along. Iā€™ve been so preoccupied with the idea of ā€œā€¦away from dupesā€, Iā€™ve completely forgot your prior advice.

Can I get rococo with Actions on Regular containers?

Letā€™s assume I have a collection of lists of articles and a collection of categories of articles. Letā€™s name the article ā€˜Musical Formā€™; The original article will live in the, by article title list container collection and the duplicate will sit in two other containers within some classification scheme - well use an academic one in this instance . The top container for the categories collection will be 'Performing arts" and the original (ā€œcurrently living in the list collection containerā€) will be duplicated twice, once in the ā€˜elements of musicā€™ container and once more in the ā€œmusical compositionā€ container. total of 3 articles, 3 living containers in the structural and titling sense. Can this process (CC/CV) be automated without manually having to keep track and go back for updates ?

I think one has to get away from the idea of putting things in permanent containers. It is much better to create associations between items using metadata. We had a whole discussion about this sort of thing in another thread:

http://forum.eastgate.com/t/user-attributes-or-tags/1069

The point is that agents allow you have have dynamic ā€œcontainersā€ that present you with the stuff you want, and it doesnā€™t matter where the original is. You donā€™t need to spend time and effort ā€œfiling thingsā€ in the ā€œproper placeā€ because you find it according to what the note says, or its metadata, not according to where it resides in the structure of the database.

Not sure if that is clear, but perhaps the other thread will make it clearer. As I said, you have to get away from the idea of using this program like other programs.

I wasnā€™t referring to replicants/tags here. I did mention agents which in effect they can only ever hold aliases, hence the misunderstanding. You need to understand that although I might keep doing do the circle, I am trying the best I can. I canā€™t quite connect everything just yet, no matter how obvious it might be for you. Itā€™s all slowly starting to click, but sometimes trivial questions do answer ones you never meant to ask. Therefore, may I ask, can this be done?

Can I get rococo with Actions on Regular containers?

Letā€™s assume I have a collection of lists of articles and a collection of categories of articles. Letā€™s name the article ā€˜Musical Formā€™; The original article will live in the, by article title list container collection and the duplicate will sit in two other containers within some classification scheme - well use an academic one in this instance . The top container for the categories collection will be 'Performing arts" and the original (ā€œcurrently living in the list collection containerā€) will be duplicated twice, once in the ā€˜elements of musicā€™ container and once more in the ā€œmusical compositionā€ container. total of 3 articles, 3 living containers in the structural and titling sense. Can this process (CC/CV) be automated without manually having to keep track and go back for updates ?

I will give you and Mark an explanation shortly. I did briefly touch on the topic here:

see, Iā€™m really trying very hardā€¦

[/quote]

I would just give the original two tags: ā€œelements_of_musicā€ and ā€œmusical_compositionā€ ā€“ and of course you can give it as many other tags as you like. Then you just have an Agent that looks for the tag, either alone or in combination with other tags or content. That results in a ā€œcontainerā€ that updates automatically whenever you add the tag to a new note, or remove it, or whatever.

If I understand this is just an outline of notes consisting only of a title and no text or other metadata. If so, why do this in Tinderbox rather than a simple outliner?

Thank you Martin for your input, I really appreciate it.

To see if I understand this discussion correctly, I believe you are looking for an app or command that will:

  • Exactly replicate a multi-level hierarchy;
  • Keep that hierarchy constantly and automatically refreshed, across different apps (so not just within OmniOutliner or Scrivener);
  • And, as a bonus, do so with no appreciable memory or storage penalty.

I do not believe any such app or system exists.

In my other life, as aviator, this is like the standard wish list for oneā€™s ā€œnextā€ airplane. One that would be:

  • Faster
  • Cheaper
  • Carry more freight or passengers
  • Burn less fuel

New planes in the real world improve on one or two of those factors, in exchange for penalties on the others. Thus the aviation-world saying: ā€œFaster, cheaper, more payload: choose two.ā€

I know there are analogies in computer coding, writing, research, construction, medicine, many other realms. There was a well known Brit writer, now dead, of whom it was said: He can write better than anyone who can write faster, and faster than anyone who can write better. The point is the tradeoffs.

In Tinderbox I think you have these options, each explained in detail in posts above:

  1. You can create an exact copy of an existing hierarchy, through the Cmd-C / Cmd-V sequence once youā€™ve selected that hierarchy. But the copy you create that way will not automatically update itself to reflect changes in the original.
  2. You can see all notes at all levels of a hierarchy if youā€™re working in outline view, as illustrated above. But that is just one hierarchy (not its duplicate), and itā€™s within just one app.
  3. You can see all notes from all levels if you create an agent to find them (descendedFrom(), as explained above), and then display their aliases, in a map or outline view. And those aliases will automatically update to reflect changes in the original outline. But that will not show you their hierarchical relationship (though there are workaround ways to show the info, in columns etc), and again it is within one app.

So your initial goal, as I understood your explanation, is not feasible in any app Iā€™m aware of. Itā€™s like the faster-cheaper-bigger payload airplane that everyone would like to have and that no one can design.
If any of these workarounds suits your purposes, great! Otherwise, the faster-cheaper-bigger payload search goes on.

1 Like

I thought J. Michael Straczinsky said this of Neil Gaiman. I hadnā€™t realized itā€™s an allusion. Wikipedia attributes it to A. J. Liebling, but Liebling wasnā€™t British. Who is the source?

It was said of Alex Cockburn back in the day. Improbably (given our respective politics and general approaches to life) he was a friend.

It always had the feel of a chestnut that been originally said about someone else, but I never bothered to track it down, so thanks for the extra info!