The Teaching Tinderbox thread branched into a discussion of my specific tasks. I’m continuing here both to put “my” stuff where I can find it easily and to avoid (further) hijacking the original thread.
The articles I’m working on are usually in the range of 1500-2500 words, often as part of a series that might total four or five such articles, so figure 10,000 words or so as the narrative unit that I’m trying to manage. Usually pretty technical material. DevonThink is my main reference database. No more than 2 or three images per article – we can ignore images for the sake of this workflow.
Not to put words in the OP’s mouth – she is a professional writer, so “how do I write an article” is not the query here. I believe we’re looking to discuss how best to use to use Tinderbox in these ways:
That last item is often perceived as difficult – how to get stuff out of Tinderbox. I suspect it’s the point where folks blanch most often because they encounter “coding” involved with export templates.
Thanks, @kderbyshire I’ve got to say, I’m having a good time with this. Your “challenge” has helped me codify a lot of the learning that I’ve been collecting over the last few months, and now I get to share it.
I’ve already started developing my base Tinderbox file (step 1 for me for new projects). From this process I’ve already developed some new insight, namely I can see derivative and unexpected learning and asset development from just the hour or so I’ve put into this effort this far (I’ll explain this in my video and review on Saturday… @Sylvaticus: Let’s be sure to have this thread on the Saturday agenda).
THIS. This is the single biggest hurdle stopping me from using Tinderbox more than I do. Working within Tinderbox is great…but eventually I need to get stuff out. That is where I would love to see the next Help file come in.
That’s the critical piece I often forget to do. And as you say that, I suddenly realize that a really simple way to do this is to create a boolean User Attribute titled $DraftArticle and then an agent to collect them. Rearrange and then do as you say.
Still, what I’d really like to do is learn how to use export code such that each note’s Name becomes a Markdown heading and the text within is nested under that heading to create a nice little structured note.
As a total aside, I tend to think in Markdown these days so if I had one additional Tinderbox wish, it would be even more support for Markdown (which at this point is probably something like auto-formatting as you type from MD to Rich Text, given that I don’t think a Markdown syntax highlighter is on the TBX roadmap!).
As @Jake_Bernstein notes, the “or aliases” part is the key. I don’t want to put the original notes in separate containers because doing so breaks the Map View I created back in steps 3 and 4.
All and only the notes that are destined for the final article have the same prototype. If necessary, I add a number to the front of the $Name of the note to indicate the sequence. I then sort in Attribute Browser view by prototype. You can then shift select the whole run of notes under the relevant prototype. Then I use File:Export to export, exporting only the selected notes in the subsequent dialog box.
Just wrapped up a new article using essentially (my variation of) this workflow. Went much more smoothly than the last time I tried this, which is good.
The magic step that I was missing: you can use an Agent to create aliases of all notes with a given attribute. You can’t rearrange the contents of an Agent in Outline view, but you can copy all of the aliases to a new container, and go from there.
Yup…that is a great insight. So glad you made progress.
I know I promised you a recording today. I’m nearly there. I figured out a really cool Devonthink issue I’ve been trying to tackle as part fo the collect process.
True. But you can sort them. That’s nice when your agent needs to collect hundreds of notes; for modest numbers of notes, it’s easier to arrange them by hand.
Part of my research process involves subdividing material into smaller and smaller buckets. I might start out with 50 or 60 papers in a given topic area. Of that, I’ll want maybe 10 or 15 as the basis for a reasonably sized article. (Hence the need for a series of articles to cover a general area in detail.)
Each of those (plus any relevant interviews) will yield between 10 and 15 individual Tinderbox notes, sorted into no more than about five idea clusters. (I know I’ve done enough research when the material starts to get repetitive. If I have more than about five clusters, there’s too much to cover in a single article.) I’ll ultimately pull out about ten notes as my outline, then add more back in as I write.
So automatic sorting is useful, but in most cases contemplated by this thread a large group of notes to sort is a sign that I need smaller buckets.
You can also use the agents to run actions, e.g. add refining tags and attributes to contribute to your clusters.
Another approach would be to use adornments to help you categorize and cluster your notes as you write them. This is that I’m planning on explaining next week.