I am back to writing my dissertation after a long pause. A lot of it needs serious editing or re-writing. The text was written in Mellel, which is an excellent editor with outline function. But all my notes and ideas for the thesis are in Tbx. So I thought that maybe it would be more logical to fill the text with data from notes, structure the text, and edit it in Tinderbox.
But there are still concerns about the feasibility of moving text from one application to another.
In old forum threads, the ability of Tbx to handle long texts was already discussed. The short answer then was - yes, it can, but there are better software tools for this purpose. However, after these discussions Dr. Michael Becker defended his dissertation, which he said he wrote entirely in Tinderbox. And Tbx itself has been seriously updated. So it seems not unreasonable to bring up this topic again.
In this regard, I would like to ask a few questions to those who have used Tinderbox to write/edit long academic texts.
Have you divided the text into component parts in Tbx? If yes, how did you divide the text into component parts in Tbx (by chapters, by parts of a chapter, by paragraphs, or otherwise)?
Did you somehow link the text to existing notes, attribute system, use agents?
Have you used Tbx to edit text that has already been written? How can this be done? (Open multiple pieces of text in the Text window at the same time?)
What did you do with footnotes in Tbx?
Have you organized a bibliography and a list of sources in Tbx?
In general, how would you rate the experience of working with long texts in Tinderbox?
Iâve been using Tinderbox for years and used it to write a lot of texts, from the outline of my first thesis to texts of different lenghts such as articles or forewords. In my experience, Tinderbox excels at facilitating the way you draw connections between topics or key-ideas. In that purpose, the Attribute Browser is an outstanding tool. But, and itâs not a default I attribute to Tinderbox, soon or later, I felt that I needed to gather a lot of materials â pdf, pieces of writings, and so on â and that, for that purpose, there were, as you said, many other tools, Scrivener in particular. But, if you donât care about it, I think that you can easily separate your draft using Containers and, in OutlineView, Separators.
I didnât find useful to complicate my process by linking notes to notes or to agents.
In OutlineView, itâs easy to rewrite and correct an already written text.
I use Markdown markups to write my footnotes. Once again, I donât want to use Tinderbox in a complicated way. I just need to use its basic functionalities.
Iâve been using Bookends and BibTex to handle my references. There are some useful interactions between Bookends and Tinderbox.
For a long walk, I prefer running rather than walking.
I didnât use Tinderbox for my thesis, but Tinderbox has moved on a lot. Note too that @satiskusalaâs method was driven in part by a non-negotiable final from of a Word docâugh. He also leveraged Markdown for footnotes and citation, though I think recent app feature improvements might allow doing that. Either way, he then used the free command line utility pandoc (see) to pump HTML/Markdown output to a final single Word doc.
The above is worth noting as if writing for publishing/assessment, it does well to take downstream fixed constraints in mind in planning oneâs workflow.
Iâââ answer the notes in a number of replies so they donât get too long. That notedâŚ
Yesterdayâs meet-up (30 Nov 24 - recording not yet posted) dived into use of Explode, with a user new to the feature being talked through its use.
Quite a lot packed in here. Perhaps the easiest way to answer the question is to rephrase it as two: âwhat is the most useful granularity of text?â and âare you exporting to final form once done?â.
What is the most useful granularity of text?
We know that Tinderbox design concept if for shorter sections of text. The length (character/line count) of a noteâs $Text is a perhaps a red herring (seeâ ). Due to the hypertextual aspect of Tinderbox people may assume the granularity is all about linking, but by being in discrete notes, the ordering of the content (in the narrative output) can easily be varied. For items that canât, or might better not be, placed directly in $Text (e.g. images, tables etc.) notes can act as proxy for such data, pulling in the content from the hard drive (e.g. images) or assembling tables from notes else where in the TBX.
With the latter in mind, Iâd suggest you want to import to reflect the granularity that best aids you for writing and (re-)assembling the eventual output.
The 30 Nov meet-up also looked at how Explode could be used to break larger input texts down into the desired note granularity.
Iâd not recommending keeping the source text of explodes once done, though you could also ways copy them to an archival TBX. Otherwise having the text twice means searching for a phrase will at minimum match twice, once in the overall source text and once in the note parsed out if the former.
Also, Iâd generally suggest not importing styled text. If copy/pasting use 'Paste And Match Style (see Edit menu) as this essentially pastes in plain text and saves you having to clean out broken styles in the pasted material.
I can be useful sometimes to paste the source text into a tool like BBEdit and remove any weird characters coming along for the ride and set custom Explode delimiters before pasting the text into Tinderbox.
Donât fall into the trap of assuming tools like Word Processors have any real notion of output other than print (even if they âprintâ to PDF, HTML, etc.). They focus on the style/render and donât have a notion of passing text in useful form to other editing apps. So taking data out of a word processor generally means scrubbing out the cr*p in the text stream before settling down to editing in Tinderbox. But, YMMV, different people with different (text) needs and different styles/domains will have differing experiences.
If using tools like Explode as part of your import, it pays to experiment with them before starting the the real import. That way you wonât trip over unexpected outcomes due to not understanding properly how the feature works.
Summary. Think backwards: what you want in Tinderbox and then figure out. Donât assume a single ârightâ way. Borrow techniques as opposed to the complete process unless you know the latter is an exact fit.
Are you exporting to final form once done?
Embrace constraints implicit in the final output. In short, donât create structure in Tinderbox that is inimical to the output you know will eventually need.
â . Sort of! If you are writing to a fixed length or max page count you can only assess that in the final submission format. For instance, line height/spacing affects the amount of âpaperâ a given paragraph of text will need. So, if you need that count, you might want to build out your export method (templates, etc.) early in the process, if only so as to monitor rough page count. All notes have a $WordCount and you can use String.paragraphCount() to get the paragraph count (IIRC each bullet in a bulleted/numbered counts as a paragraph.
A very open-ended question which implicitly suggests a right/best way to use things. But Tinderbox is a toolbox. The better question to ask yourself is what information is immediately discoverable in the existing source text and where/how best to map it (to attributes and links) or fin it (via agents).
If you find yourself planning a discrete agent for every discrete thing you might want to find, it is possible, but is a naĂŻve approach that scales poorly: eventually youâre mired in turning agents off and fighting edge cases where action code is acting on aliases in agents rather then the intended notes, etc. Also much meaning in human-written text is implicit so queries like $Text.contains() canât find what is not there in the literal text.
So think, too, about things you might want to explore via metadata: how to find it and how to store it. Again, experiment on small sets of text text before the main import so youâve less do-over due to bad prior guesses. Still, Tinderbox is very forgiving of such re-structuring: the main cost is your time/sense of humour.
You can store data about notes in attribute value and/or (typed) links. Why might you use both for some things? Links help you see relationships. Donât fall into the trap of putting all notes in one container just so you can use map view to see the linkss. To let you use the power of the outline for structuring your pieces of text (i.e. notes) via nesting to indicate structure, read up about hyperbolic viewâand experiment with itâas it allows you to look at links across the whole document.
Whilst storing metadata in attributes that you then display as Displayed Attributes, do also make friends with the Attribute Browser view which will leverage that data even more.
Bottom line. Whereas most text writing apps appear to do more for you, you still have to colour inside their (design) lines. Tinderbox offers much more freedom, but you the user are an active part in choosing which tools in the toolbox to use. Beware the lazy shortcut of âjustâ re-using someone elseâs method as generally it puts you back in that same constrained process of doing what makes sense for somebody else but not necessarily you.
The unstated part here is edit to what end? If the final output is from Tinderbox then previous answer apply. If this is an out-and-back edit, Iâd assess whether it is worth moving text to Tinderbox just for some textual editing/proofing.
Ideally, start the work in Tinderbox, or migrate to Tinderbox so you donât need those apps that arenât delivering for you.
Yes. But there are 2 questions here. You can use Tinderbox to organise info on sources, or simply link out to your reference manager app (Bookends, Zotero, Menderley).
The gnarly part is if you want to render the references for a long form article, such as a thesis, including the bibliography and with the latter in whatever style the end target format requires. @satikusala has experience of this.
Good! But from a standing start, there is some configuration to do. So for a novice user, some learning of the app. The learning burden, if such it is, is enlightened self-interest. Work? Sure, but the payoff is much more control.
I have only a moment just now, and there are excellent discussions already upthread. (Perhaps we should turn this discussion into a web page? a video?). But, very briefly:
Much depends on the nature of the edit you have in mind.
Structural Edit
Iâve recently taken the raw material for a projected book chapter and edited it to form a conference paper in a different discipline. The paper makes an argument thatâs completely different from the book chapter, so from the start I knew Iâd need an extensive structural edit.
In this case, both genre convention and the proposed argument for the short paper meant that the topic headings of the paper could be sketched at the outset. (OK: I got them wrong and introduced three subheadings, but the gist was obvious from the start). So this was a fairly straightforward, if extensive, exercise in finding evidence that slots into each heading. In some ways, very Luhmannesque!
Exploration
In the past, Iâve had rewrites where I really needed to return to raw notes and completely reorganize them, commune with them, and figure out what they were trying to tell me. Thatâs a map view task (now supplemented by GaudĂ and Hyperbolic view).
Line Editing
Line edit in place. Donât do acrobatics if you donât need them.
Links
Essential links include
connecting sources to notes based on or taken from the sources. This should be a Tinderbox link unless every source will be represented in your reference manager â that includes interviews and correspondence â and you completely trust your reference manager. If you do, a reference manager URL or citation will suffice.
Connecting multiple versions of the same note, whether duplicates or separate uses of the same extract. Again, if you donât keep these connected, you will end up using the same quote in two places, and readers will notice. (This is less a concern in a dissertation, where the audience is small and might not object to a bit of repetition, but itâs poor form in a book.)
Thank you very much for such thoughtful and informative answers!
I should have written right away that I plan to change the structure and partly the text itself, add some things, delete some things.
I plan to do the final formatting of the text and bibliography in Mellel using Bookends.
@dominiquerenauld Thanks for sharing your workflow. It is very interesting, there is a lot that can be taken into account.
Could you please clarify how you use attributes with topics and ideas?
This is a bit off topic, but I saw a Map container in your picture. Is it correct that you are using a separate container for the Outline view? Iâm trying to do the same thing, but I donât know whether itâs better to duplicate notes for Map view and Outline view or to create alias for one of the views.
Thank you very much, Mark!
Incredibly detailed response!
I should have said right away that I was not going to finalize the thesis text for printing in Tinderbox.
Yes! Thatâs what prompted my post. I was very impressed with how cleverly Bruceâs book text was transferred to Tinderbox.
My dissertation is in history field. So it wonât have pictures (just some number of tables and graphs). And I wasnât planning on putting them in the Tinderbox. But your line about âproxiesâ made me think about the possibility of using such notes for ideas and building text structureâŚ
Thank you so much for this advice!
The ability to track daily writing progress is one of the very valuable features of Scrivener. And I was very happy to learn that this can be seamlessly implemented in Tinderbox using Dashboards!
Yes, I plan to finalize the text and links in Mellel and Bookends.
Note, as Iâve demonstrated in past meets, using Obsidian as a bridge between Word and TBX is great; e.g., Obsidian will cover work formation (e.g., headings, bullets, and URLs) into Markdown. When you pass the markdown into Tinderbox, youâll have a much easier time with exploding the text.
My dissertation has well over 50 graphs and figures. As Iâve shown in past meetups, Iâve found the best approach is to keep the images (.png, .jpg, etc.) on the hard drive. I have a note for each image with populated attributes ($MediaPath, $MediaDir, $MediaCapation, etc.) that give the Tinderbox template Iâve developed the location of the image and how I want it formatted in the output.
That is a fine idea, I do that too but I use Zotero. Iâve found Pandoc (i.e., the of a .csl file) the best way to handle citations. You put anchor tags in your text, e.g., [@Becker2024;@Bernstein2017;@Anderson2019], and Pandoc will tale your Tinderbox output and automatically create your references for you. From there, you will likely need to do manual editing to clean up the formatting to align with the required style, e.g., APA.
Here is an example from my dissertation; all of it is automated. NOTE: for the entire document, I created a âDraftsâ strategy that I can share one day in a meetup to show how to do this across the entire document as opposed to on individual TBX notes.
The literature on PIMS has primarily focused on a general overview of a PIMS [@Bakir2021;@Janssen2022a;@Janssen2020;@Janssen2022], regulations, and infrastructure development, especially in the areas of data collection, storage, exchange, protocols, privacy-preserving search, standards, security, and user interfaces [@Jha2022;@Pawar2022;@Tzolov2020;@Bergman2008;@Zichichi2020;@DoleraTormo2014;@Sung2021;@Tzolov2019;@Wang2012;@Piasecki2022;@Tran-Van2017;@Florea2023;@Piasecki2021;@Loffreto2016;@TrustOverIPFoundation2021]. Below, I discuss 12 studies, published since 2010, that have specifically investigated the factors that might influence PIMS adoption.
Have you divided the text into component parts in Tbx? If yes, how did you divide the text into component parts in Tbx (by chapters, by parts of a chapter, by paragraphs, or otherwise)?
MJB>> This all depends on your intended output and writing style; as you see above, I spring it upon sections and paragraphs.
The real trick is to think about the following:
What do you want you final output to look like
How might you want to use/re-use elements of your work, .e.g., references, figures, tables, terms (e.g., for a glossary)âŚthe answer to this will help you with how much you want to atomize your work into smaller bits.
Once you answer the two question above, you can then determine your templating strategy and how youâll use Tinderbox to help you tease the the final pieces together.
Did you somehow link the text to existing notes, attribute system, use agents?
MJB>>I use links to help me think and work with elements of text in context, e.g., I can define linked terms in real time as I write them in text. I can then use the terms Iâve associated with a note via the $Term attribute in a glossary or indexes I create using agents as part of my output. Iâd be happy to demonstrate this again in a future meetup.
Have you used Tbx to edit text that has already been written? How can this be done? (Open multiple pieces of text in the Text window at the same time?)
MJB>>>What do you mean by âalready writtenâ? Do you mean in another app? Do you mean in other places within TBX?
What did you do with footnotes in Tbx?
There are several ways to do this; it all depends on your structure and writing style. Weâd need to dig into this further.
Have you organized a bibliography and a list of sources in Tbx?
MJB>> Absolutely; see above.
In general, how would you rate the experience of working with long texts in Tinderbox?
MJB>>For me, I spend 90% of my day in Tinderbox. I only go to other apps for final publishing and fit an polish.
Regarding the Attribute Browser, there are some useful resources.
Is it correct that you are using a separate container for the Outline view? Iâm trying to do the same thing, but I donât know whether itâs better to duplicate notes for Map view and Outline view or to create alias for one of the views.
You donât have to duplicate your notes. Tinderbox can open simultaneously several views of your notes using multiple windows:
This is not really off topic as understanding Tinderboxâs views is important for getting good use out of Tinderbox. At time of writing, Tinderbox has no less than ten different âviewsâ. All view the same underlying data, they simply present it in different ways and with differing scope/context.
For perspective, in a Word document, look at that appâs âViewâ menu:
Most people never change that selection but, in context, there are 3 other views of the document available. Tinderbox has more, and more varied views, but all drawing on the same underlying data in the TBX file (or âtextâ, if you prefer). Note that not everyone needs every view, and different views suit different user tasks. So do not worry if you donât use all the views. However, if you only ever use either/both Map and Outline views, youâre likely missing out on useful affordances of Tinderbox.
Tip: Read up on all the Tinderbox views and what they do (see), if only so you know you arenât using some of them for a reason rather than because you didnât realise what they offer.
Your use of the term container here possibly seems confused. The underlying storage of your TBX file is an outline (yes, if you view the fileâitâs just XMLâthe outline is there). A container can be thought of akin to a folder in the Macâs Finder. A Tinderbox map shows the contents of a single folder, by default an outline shows the whole document. If still unsure, Iâd suggest reading this section of aTbRef. Although it only discusses Map and Outline views, it should also give you some general concepts when looking at other views.
Map is most useful at an early stage for doing spatial hypertext: linking and using (map) position it indicate purpose. But, a map only shows one folder. As you move to long-form output, you will likely want to start nesting content and here the outline is more useful as once an note moves outside the parent container (even if as a child of a note on the map) it is no longer visible in the map.
Think: map more at early stage, outline as you move to a single (multiple) narrative.
App windows vs tabs. As you will note, a new TBX opens with two tabs:
The tab with a blue fill is the active view, the others are a click away, configured but not computing their view (i.e. not adding CPU load) when not active. Each tab can hold a different context/view and a different selection.
A document can also show more than one window (each with its own tabs). An important limitation is that the active tabs of all windows share the same selection (active note) even if that note may be outside the contexts of one of the windowâs active tabs (e.g. the selected note might not be on an active tabâs map in one window).
You can also open notes as standalone windows, to more easily view/read the text of two notes at the same time, one stand-alone, one in the text pane of the front window.
If multiple notes are selected in the view page, you get a composite listing of the titles/texts of those notes. Be aware (for those whoâve used Scrivener) that note $Text is not editable in this multi-note selection state.
I apologise if this seems tangential. However, understanding the above is important as all too often 'how do IâŚ?" questions arise because we simply assumeâbased on experience of other toolsâan outcome or method that is not directly possible. Yet, the goal is normally achievable, but just by a slightly different method.
Bottom line. There is a trade-off between flexibility and user engagement with the tool. a word processorâs offering is that you type in the white rectangle. Tinderboxâs offers much greater abstraction. You could write a very long piece in one $Text, but I think all here would counsel against that.
Thanks for the post @Maximus. I see that you and @dominiquerenauld use Bookends. Iâve attached a Bookends .fmt bibliography format export file, âTinderbox Abs. & Notesâ that I developed, with the extensive help of @satikusala and @mwra almost year back. This allows some useful things, including adapting the title and putting the contents of the Abstract and the Notes field into the Tinderbox text panel. Iâve also attached an example Bookends Library and the end result of copying the four references into a Tinderbox flle.
To do all this, the .fmt file first needs to be placed in your Library> Applications Support> Bookends> Custom Formats folder, and you then need to check it in Bookends> Biblio> Formats Manager⌠as per the screenshot:
and then press option command together and then, while still doing this, drag a reference onto Tinderbox. This creates a Tinderbox note, and fills the relevant Tinderbox attributes as anticipated for a reference, but also puts whatever is in the Abstract and Notes fields in Bookends into the text field - this last bit is what the Tinderbox Abs. & Note.fmt sets up for the handover to Tinderbox.
This is useful for various things, one of which is that I make extensive use of the Notes field in Bookends as I work through a pdf, making narrative notes and putting in clickable links to various highlighted text notes. I use # and § markers for sections and sub-section markers.
Then in Tinderbox I can use âExplode noteâ sequentially operating on the # and § characters to break down the Tinderbox note into sub-notes and sub-sub notes. In the uploaded Tinderbox file you can see that I have done that with the Mintzberg Strategic Safari book reference.
You can also see (the MacGarvin ref) how I broke up a Chat GPT conversation about Business Strategy into the responses to each question. In that case as I was reading and thinking about the responses when doing this, it wasnât a great issue to do it manually as I went along. Tinderbox, so used, makes it a lot easier to navigate long AI dialogues! But I expect it would have been possible to get Chat GPT to put in e.g. # and § markers, so I could have proceeded as with the Mintzberg ref. (This and the Mintzberg exploded example are best viewed in Outline mode)
Files. I have attached a zipped file with the âTinderbox Abs. & Notes.fmtâ file, a Bookends library file with four references in it, including the examples just given, âTinderbox Demo.bdbâ and the Tinderbox file with the four refs imported along with the various bits and pieces to make it work, âTbx Bookend Ref and Explode exampleâ.
Notes and caveats. 1) Anyone can download a Bookends demo that works with up to 50 refs, so it should be possible for anyone to play around with this. 2) Any links within Bookends wonât work, and the attached refs wonât be there 3) Iâve drastically pruned a Tinderbox file to include the minimum for this example, but the prototypes and other things to make it work all seem to be there when I tested it. 4) Itâs ok to test this by dragging a ref from the example Bookends Library onto the Example Tinderbox file. It wonât work with other Tinderbox files until you have copied across the adapted Reference prototype and just possibly made some other adjustments (a lot of stuff has happened since we made this!). 5) Importantly, it takes a while for the Tinderbox notes to update to display what you see here. Toggle between Map and Outline View and back to accelerate this process, and if that doesnât work, select File> Update Agents Now. 6) Formatting in Bookends is lost in the transfer, as are any internal links in your Bookend Note (However, as an entirely seperate action it is easy to copy eg a Bookend highlighted note in Bookends and paste it into the text panel of any note in Tinderbox - this has revolutionised my research!)
Hopefully I havenât forgotten anything vital. (Anyone) Let me know if it works, or doesnât work, for you!
