Hi there, nice ideas posted here. What is your end goal? To produce a latex or markdown file? Is your end goal a Word doc? If so, you can go straight from Tinderbox markdown to HTML and then have Pandoc process the HTML and convert it to a Word doc. In this workflow you would NOT have the YAML in Tinderbox, rather, it would be an external file you’d have pandoc pull during the conversation phase. I do this all the time. Have developed a variety of attributes and stamps to modify the output depending on the goal. There are some other tricks I’ve learned, like custom divs so that word recognizes the heading in the file ad maps them to the Word template headings, Finally, what citation format do you use? I’ve recently refined an APA compliance stylesheet, pretty cool! It formats citations and headings, e.g. 4 and 5, properly. Once I have a better idea of what the ultimate end goal is I can explain further.
Of course you could create the external file from the Tinderbox document. You’d just need to ensure it was properly exported before exporting the main HTML document. Not suggesting one must, but it’s possible and it gives you the same flexible authoring for both exports.
I saw a video you made about converting a Tinderbox document written with markdown and HTML templates to a Word doc. It’s very useful. The reason why I don’t use this workflow is that, in my own writing experience, document appearance matters a lot and I really don’t feel comfortable with Word document look. For a moment, I tried to use Mellel, great word processor, but it was to mimic Latex look and as I examine every last detail, I found too many esthetics inconsistencies in the final layout that it was easier to use Latex directly. That’s why I wrote my doctoral thesis using a completely rewritten Latex template. But, recently, I found markdown much more simple to use than Latex markups, even if I can’t do without those ones as the process to output a Latex layout requires to set a Latex front and backmatter.
That said, ideally, I’d like to use a Tinderbox file with a markdown structure combined to a YAML template. In this way, I could write down my notes in Tinderbox and output a Latex document. As @mwra said, it would require to use an external file as I do, I suppose, with external Latex files.
In the screenshot below, certainly clearer than I can say in English, my current writing workflow:
Love the diagram! It shows how many ‘moving parts’ there are even if with time/practice we no longer ‘see’ them all as we work.
I noticed reference to R.studio. Althuogh you are using it for word generation, R.Studio’s genesis in the R (what is R?†) community offers a link to use of Markdown for content generation. See:
That might offer some new avenue/simplifications. I can’t give more detail as I’ve not yet had time to do more that look at the ever views.
For those whose domain is more sciences/empirical (i.e. not as above) the R Markdown/Bookdown approach draws on the approach of active documents as seen in thinks like Jupyter Notobooks (Python-based).
As proof positive, I belive some (all) of the Routledge/CRC Press ‘R Series’ books are generated—i.e. formally published—using the above methods, e.g. here.
Mark, thanks for the hints!
And: Have you come across Quart Pub (https://quartopub.com), which is even more de rigeur these days, offering a further avenue to publishable format?
Interesting. How does one research this as home page offers a sign up only! I’m a bit wary of providing info to an unknown/undescribed service. Ah, not linked from home page is this Quarto - Publishing Basics which gives:
So, force sign-up for a free service. What is the web coming to?
But, that aside this looks like a web-based doc production service built atop pandoc. so, yes, interesting!
Sad I can’t make this weekend’s meet. But if you get into discussing export/publishing of Tinderbox material, I think there’s an aspect overlooks as it is too easily wrapped into tool features rather than being seen as structure. This relates to pretty much all non-body text and in no particular order:
- Table of Contents
- Running headers/footers
- Pagination page numbering (and ToC page links)
- Screen vs hidden (link) targets, i.e. can the URL of the target be seen without following it (or at all in on actual paper)
- Endnotes, chapter endnotes, or (page) footnotes plus forward/back interlinks.
The above are often hard, impossible in many workflows, to implement. Though many have a root in print era practice many still have a use. Some need change, endnotes/footnotes being the most obvious candidate.
If we’re just writing a shopping list for a sandwich or blogging about the sandwich we made, the permalink is the one thing we generally need. But moving up the formality chain to the likes of academic publication, the above affordances are mandatory so need to be available those who use them. Whilst I don’t think Tinderbox should be trying to service them all, methods like
pandoc and sidecar info (e.g. YAML data etc.) can probably provide the necessary info, i.e meaning we can create the source data all in the same TBX.
For general web use I’d love a means for (template-based?) transforms to insert a pages’ footnote notes as an include but _importantly with in-page links/return links and a means to shifting to-footnote anchors in $Text to superscript numbers. This implies on-export number generation. I think @satikusala emerging report creation method offers one such way, though it’s embedded in a system so perhaps overkill for some users.
I too love this workflow, very powerful. For clarity’s sake, my workflow is similar, but I try to keep as much of it in Tinderbox as possible (see image below). This way I keep all the rewrites in one place (I have had an automated drafts method setup that gives me over 50 verbal affirmations each time I create a draft).
One day I’d love to learn the other tools and play with Latex, but so far I’ve been able to centralize nearly all my work through Tinderbox.
The key success factor in this, for me, was to focus on the separation of content, structure, and appearance; to learn markdown, HTML, CSS, and regex. It has taken a while, but I’ve boiled my temples down to just a few cascading templates that can create outlines, pull pictures, and tables, adjust to apa and Chicago styling, import in citations, and more.
I would really love to see your process in action. Maybe we could schedule a call in a few weeks or you could share it during one of the meetups.
Cab you explain these a bit more?
Darned auto-correct! What I meant was the same TBX can both store/export the data needed to inform
pandoc how to process your main export, and do the main export. The term ‘sidecar’ is general usage—perhaps too informal†—to refer to metadata stored outside a file either because the format doesn’t allow embedding or because the user doesn’t have the tools/permissions to insert such data into a file.
Hopefully that clarifies things and sort of the auto-correct typo, now fixed.
†. Sorry, didn’t meant to be obscure. IIRC the term gained traction in the early days of image/photo metadata such as EXIF when most image editing tools could not embed/process such data (or do so correctly, as ‘standards’ were in flux).
Thank you for this suggestion. Interesting meetup today as usual!
Video on Vimeo: Tinderbox Meetup 20 August 2022 on Vimeo
The video can also be found here, along with an intricate sample template: Tinderbox Meetup - Sat. Aug. 20, 2022: Importing, Explode, and RegEx - #4 by satikusala
Great session, sorry to miss it.
Regarding the Setapp service - I’ve made a quick reference googlesheet for those interested in perusing their current offerings without having to wade through screenfuls of iconery…
Edit: made the link shareable.
Cool. Which ones are your favorite? I use Bartender, Downie, TextSniper, Houda, REcord audio from the past, Trac CPU, CleanMyMax, Expression and a hand full of others.
I already have† lots of these, or else SetApp would be appealing. But if I didn’t, it seems a sweet deal if you can predict you’ll use the service for a few years.
†. Of these those bold are used regularly: Aeon Timeline, Dash, Expressions, Hype, iStat Menus, RapidWeaver, Screens, TextSniper, TextSoap. If I hadn’t got proficient with regex in BBEdit or
sed on the CL, I’d still use TextSoap more - so no shade there. If I didn’t WFH so much I’d also go for Bartender as I have a 2010 MBPro 14" with a notch and the single-screen menubar UX … sucks; luckily I’ve 2x24" externals when WFH.
I use (or have used) the following:
BetterTouchTool - BetterZip - ChronoSync (full ver) - Commander One - Disk Drill - ForkLift - Gemini (switched to Duplicate Detective as I found it more reliable) - PDF Squeezer - Renamer (now mostly using Finder) - TaskPaper (less so these days) - Ulysses (moved to Scrivener)
For me as well, setapp is not practical at this time. I’m committed to lessening my app footprint; moreover I don’t really need most of their current collection. Exception - I’m intrigued by apps like Coherence X and Unite, as there are a couple useful applets and customized calculators that would be good to have locally as apps for when I’m on the road and offline.
A point on TextSniper - nice little app! However, Monterey offers this functionality now built into MacOS, as I’ve mentioned elsewhere on the forum; using it involves an additional step or two over TextSniper, but I don’t really mind as it helps me fix the inevitable typos and also gain muscle/mental memory in the process.
Interesting re TextSniper. Agree the OS option is there. I struggled to make it work smoothly† to the point where dropping $ in the AppStore on TextSniper made sense. Driving that was the fact I’d been asked to ‘just’ quickly get the text from this monster, a mural poster in JPG format:
TextSniper got me there—822 discrete text line/paras—faster than expected‡. I liked the additive clipboard option (not used above as the individual ‘box’ texts needed to remain discrete), but I did use the ‘ignore line breaks option’ and that worked well. Fidelity of capture from slightly blurry JPEG text was impressive. Rotated text, if clear of other text was detected OK. Only challenge points: text wrapped around images with ragged margins and tables/charts. The letter I didn’t expect to just work as the layout is not like linear text so would need extra processing. Still I reckoned I only needed to correct less than 1% of the capture in the above dataset, and only then for such reasons as being from tables, etc.
I’d agree that the free OS (Monterey, Apple CPU) LiveText-like option makes sense unless one will do this sort of thing often/in volume. In that case, if you have setApp or if $7.99 isn’t out of budget, I’d consider TextSniper
†. Probably that’s me misreading the instructions so my failure to use the OS method quickly/efficiently must not be read as negative against the feature.
‡. Great for a first real use, but underlining the point we often make here re utilities vs toolsets like Tinderbox. The utility does a (set of) very narrowly defined task(s) with few if any optimisations. The user does what the utility allows and as long as that aligns with the user’s need, happiness occurs. With a toolset, the user chooses the right tool to fit their task as—counterintuitively for the new user—we don’t all do the same thing or do the same thing in the same manner. Thus the differing, disparate and sometimes complex workflows we see here and the fact some user thought is usually needed.
Good points and review; I don’t do that much capture yet, but would definitely invest in TextSniper were that to change.
My thought at the time, too. The things we volunteer for