Interview with Keith Blount, Creator of Scrivener

Once you get it set up how you want it, DevonThink absolutely rocks and it’s perfect for non-fiction research. It’s got a 100 hour free trial so worth testing it out. I’d recommend downloading the manual and/or the Take Control Of DevonThink book from the devontechnologies website (both free) and giving both a speed read so you get an idea of some of the stuff it can do.

Obviously depends on your own individual workflow but here are some of my recommendations:

Make sure you get the iOS companion app DevonThink To Go. This allows you to capture everything on your mobile devices. Also make sure you install all add ons via DevonThink3 -> Install Add Ons.

These add ons will:

- Allow you to view and import emails in DevonThink (you will also have to enable the plugin in the Apple Mail app for this functionality to work)

- Put an ‘Inbox’ folder in your finder which allows you to just drag any file or folder into it and have it appear in your DevonThink Inbox

- Give you OCR for all your pdfs

- Give you extensions for browsers to clip things easily.

When you clip something that you want to read or refer back to you’re generally best off using the ‘paginated PDF’ setting as not all the available options support annotation and the AI search. It’s not a big deal if you don’t because you can just right click something and click ‘Convert To’ to turn it into a pdf later.

Two settings I’d recommend trying out:

DevonThink3 -> Preferences -> General -> Interface -> Retain View (On)

Stops the view from changing every time you switch databases which can be annoying.

DevonThink3 -> Preferences -> General -> Interface -> Unify Inboxes (Off)

Moves each databases inbox into the database itself which makes it a lot easier for when you’re filing things.

Four Shortcuts I Find Useful:

Cmd-Ctrl-M - brings up a panel for moving a file quickly

Cmd-7 for wide screen view (makes more sense than the standard view)

Cmd-option-i - toggle right inspector

Cmd-option-1 - toggle left sidebar

Try UpNote, I find it more usable than Obsidian,

So it will, to varying extents, reduce time you spend in any of the following programs:

- Finder

- Preview/PDF Viewer

- Bookmark Manager

- Hazel (automated file management)

- RSS Feed Reader

- iBooks

Everyone's needs are different and my workflow is still a work in progress but I basically run it on similar principles to Inbox Zero. It goes like this:

You have your Global Inbox database which is your in-tray. You then have a bunch of different databases which you could just view as being top level folders. Each of these has its own Inbox. Within each database you can create groups which are basically just folders. You can nest as many as you like. I hope from this description you can see that it is basically like a fancy Finder.

Anytime I find anything on the internet I find interesting I clip it into the global inbox as one of two things (there are more options but these are the ones I use):

- A bookmark if it is just a general site I may want to check out again. I particularly use this for my Coding database for storing useful tools I may want to use at some point.

- A clutter free PDF if it is something I want to read, extract quotes from, annotate etc

DevonThink also adds an Inbox folder to your Finder sidebar so if you have any other files you've acquired throughout the day e.g screenshots you can quickly add these in by dragging them into this folder.

I then make sure that each day I go into the global inbox and move each item into the respective database's inbox (or just the top level of the database if I'm feeling lazy). This way nothing ever gets too out of hand. I have databases for all my broad interests e.g Maths, Coding, Fitness etc as well as a Misc database for everything that doesn't warrant having it's own database.

All of these category databases gets at least an Other folder. This means that, when organising each database I can keep the top level free of files: if I don't know where to file something it goes in the Other folder. Think of it as the junk drawer of the database. Periodically I will go through a database's Other folder and will look to see if any new natural groupings have emerged in which case I will select the files and create a new group for them. Otherwise I'll just leave them as is.

You can also set up smart rules to automate a lot of this for you so you can just tag things as you are clipping or importing them and then DevonThink will automatically move them into the right places for you. You can do much more than this with Smart Rules: for example you could have a script triggered when a file was added that met certain conditions but I don't have any need for this advanced level of control right now. You can also set it up to track RSS feeds and import them if that's your thing.

As well as the category databases I also have a few databases that are organised differently. I have Non-Fiction and Fiction databases which house pdfs organised in folders by author name. These are for general non-fiction and fiction only. Specific subject textbooks and papers live in their respective databases instead. I have an encrypted database for important stuff like insurance details, and I have an Articles database.

The Articles database currently has no groups inside it just PDFs at the top level of any articles I found interesting. By article, I primarily mean blog posts and journalism not (normally) academic papers which tend to be specific enough to add to a category database. If something is obviously category specific like a coding tutorial I will move it into the relevant database. I do not bother with folders in the articles database as I have found it is almost impossible to organise general articles due to them often touching on several different themes and you never know what is going to trigger remembering an article or when or what you will use it for. This is also probably why newspapers resort to using such broad categories like Culture, General Affairs etc. I thus find it better to just have them all at the same level. I may group them into folders by the year I added them into the database if I feel it getting too messy.

I'm not being disciplined enough with myself currently, but what I want to start doing is skimming the newspapers each morning, clip the articles I want to read, and then do all my reading in DevonThink. This way I can highlight and annotate as I go rather than reading in the web browser and going back after the fact and highlighting the parts I found interesting. As you're highlighting you can choose to right click and Add Note to add a specific comment you may want to make. You can then click Tool -> Summarise Highlights -> As Markdown which will generate a file with all the quotations and comments you've made along with app url links to both the general file and the page number of each highlight for easy access.

As well as dragging files and folders into DevonThink you can also index files and folders stored elsewhere, allowing you to keep them in their original location but access them from within DevonThink. You can take advantage of this by creating a new database for your notes and then clicking File -> Index Files and Folders... and selecting your Obsidian vault. You can then use the Quick Finder (cmd-ctrl-m) to move your newly created highlight note into Obsidian for further editing. Or you can just edit the note from within DevonThink using the built in markdown editor. If you add it into Obsidian you can take advantage of block inclusion which allows you to keep the quote in its original file but use it in many other files without having to copy paste and potentially losing track of where something came from. To do this you just start typing ![[filename^ at which point Obsidian will give you a drop down list of sections in the file. You select the part you wish to use and Obsidian will then automatically generate and add a hash that represents that block e.g your markdown text will now look like ![[filename^1js3cj]] and the WYSIWYG editor will display the quote in your new file as if you just typed or pasted it there.

I think by now you can see how you could use this kind of workflow to not only organise all your different bookmarks, files, highlights in one relevant place, but to also use that information to generate your own new content quickly and easily. Having everything organised so well has even more benefits when you want to retrieve something as you have a much better idea of where it is going to be located and can then take advantage of that by doing a search on that specific database rather than having to do an entire file system search from Spotlight/Alfred. DevonThink's search is much better than either of these as it will search the contents of your file not just the filename so if you have no idea about the name of the article but can remember some of the topics it contained you can pretty much always find what you're looking for. You can also use search operators such as AND, OR etc as if you were doing an SQL query if you want to get find grained. There is also an Alfred workflow available so you can type your search queries out lightning fast.

One final thing worth mentioning is another program devontechnologies makes called DevonAgent. This allows you to do searches of many different search engines simultaneously (including academic ones), including deep searches where it will follow links on the pages found to as many levels deep as you set. You can search with boolean and advanced operators such as term1 near\n term2 which will search for term1 with n words of term2. You can also set the program to do these searches periodically e.g every 24 hours if there's a particular field you want to keep up on. You can even get it to send you the results as an email. Once you've got your search results, you can then either open them in the app, in the browser or import any interesting looking articles directly into devonthink.

I should have my new blog finished soon and I may create a post and video or something at some point to go into even more detail but hope this helps in the mean time!

> make a […] web app

I’d make a different recommendation. If you want it on Windows, make the best Windows app you can. Take advantage of all the APIs and features the platform has to offer. If you do a good job, you will make something faster and smaller that follows the platform conventions and is comfortable to use by other people like you.

Right now I think the Mac has the healthiest ecosystem of interesting software from small developers. I’m excluding games - Windows and the three big consoles rule the games space. For personal software (like the titles you mentioned) Windows feels like it’s fading and Linux is doing what Linux has always done.

I build marketing software for authors (called StoryOrigin if you want to look it up for credibility purposes), so I'm quite familiar with the space.

Exactly what you've described is already out there: https://www.atticus.io/ (released last year - I have no affiliation.)

Ah, I'd actually forgotten the RTF files at the bottom of the folder hierarchy.

I put together a set of Org-mode key bindings, config variables (file-local) and tag color codes for a screenplay project. Org-mode's HTML output was way off from what I wanted, so I hacked up a CLI filter, in a language I know better than Elisp. From HTML it's easy to get to PDF, an industry-standard format.

> I've not used Scrivener myself but as I've heard it described it is as a tool to organise chapters / parts of a book and sort out their relations, how are you using LaTeX for that?

It doesn't really do that though. It's not a relational database. You can't say, create a MainCharacterProfile document that then gets "attached" to certain scenes. The closest you can get is manually typing out "characters in this scene" and then manually hyperlinking to that document. Which won't update automatically if you were to remove that character from the scene. It's kind of a mess to use it like the marketing says you should.

If it was relational, that'd be fuckin' dope. It's best to think about it as a combination of a filing cabinet, a typewriter, and a cork board. You can manually sort your filing cabinet, rearranging your scenes by page or by folder. You can type up new scenes and edit old ones. You can drag and drop things on your cork board. You can add labels and colored flags for filtering and sorting. But that's functionally it. The only things it writes to disk are rtf docs. No library or structured metadata or anything; just more rtf docs.

It's honestly closer to "wordpad with tabs and folders" than anything else. I've been doing NaNoWriMo on most years since ~2011, and I think it's total garbage. Plus my editor always wants a word doc, so revisions after the draft just happen in Word anyway. I think there's a good chance I could convince them to read a LaTeX document though, maybe even teach them how to add comments. As a bonus, the word document that scrivener generates doesn't have all of my background and notes, but the LaTeX doc has them right in-line as comments, so I think the feedback would be more constructive in a lot of places.

This has triggered a memory, whereby I had spent probably a similar amount of time getting a custom template done - then, by the time my book was finished, there was an upgrade and the template vanished/became unusable - and then the whole templating system seemed to have changed dramatically, and I didn’t have the energy to burn another 2-3 days re-learning it all… Anyway.

I actually think the higher price is an asset, not a liability (unless you can't afford it, of course). It means they have a business model that doesn't require a huge mass market to survive, so they won't go broke or get acquired and watered down by some big company.

If you’re interested in screenplays you might be interested in the fountain file format. It’s like markdown but for screenplays. It allows you to write screenplays in any text editor and I really like the simplicity of it; keeps me focussed on the task at hand. There’s a very cool app created by screenwriter John August (Big Fish, Charlie and The Chocolate Factory) called Highlander2 on the Mac App Store that is purpose built for it (technically it uses a custom file format called highland but it’s basically just a folder with a fountain file inside it and a few other files for editor settings, metadata and handling images if you have any). There’s also fountain extensions for VS Code, Sublime Text, Obsidian and probably other editors out there I don’t know of.

Links:

https://fountain.io/

https://highland2.app/index.php

Vellum is expensive in comparison to Scrivener, but the output is very nice and you soon see the increased value of using it.

Scrivener and Vellum do different jobs. At a simplistic level the former is for creating the content and the latter for producing the final outputs (Scrivener can do both but really only excels at the first).

"writing books" vs. "publishing" ? Is that a sharp distinction?

I recently wasted three days trying to get Scrivener to do something very basic - autonumber sections in a book without including part numbers.

It should have been easy, but the autonumbering features didn't work as advertised and I ended up with something that is probably a hack that relies on some bugs.

It's good VFM, but considering its audience it's also one of the most user-hostile pieces of software I've ever used. [1]

There's far too much "Yes it does that but it's not designed for it so keep your expectations low" for comfort.

Either include features and make them professional, or don't include them at all. I'd happily pay two or three times as much for something that does all the things it sorta kinda implies it does but you know actually not really.

Vellum does almost nothing in comparison except produce a limited range of beautiful books. But perhaps that may be of some interest to writers?

[1] TBF it's far better than Calibre. But that has the excuse of being free.

skribh is a cutty-sounding syllable for sure.

Swing your vorpal sword, skribby skrab.