Looking back at Scrivener and OS X

Posted by Chris Rosser on Mon 30 November 2015
Hello! This site is archived and no longer maintained. For Chris' main site go to chrisrosser.net

I finished NanoWriMo yesterday with a wordcount a pip over 50,050 words. It feels good; quite a chunk of my planned re-write is now drafted and I'm probably only a month away from finishing the first draft.

For reasons I'm not entirely sure about, I decided to go back to Scrivener for the next stage of my drafting process. Maybe, it is the carrot of the iOS version dangling in front of my iPad mini, maybe it's the promise of integration with Aeon Timeline to help me fix a chronology issue I'm having. Or perhaps it's because the original manuscript is already in Scrivener format and I was missing the built in Concise OED in OS X.

Any of these reasons are a good excuse to try and put Scrivener to work once again so I spent yesterday afternoon importing the outline of my redrafted novel and my NanoWriMo 2015 file. Both of these are in markdown format (even the outline) and to its credit, Scrivener handled the import well and faithfully created the structure in a nice hierarchy of folders and documents.

Firstly, I'll mention the things I liked.

Outlining and Corkboards

Scrivener's organisational tools are pretty damned good and I greatly appreciate the ability to drag and drop chunks of the story around. As I got writing NanoWriMo, I started to deviate from my outline as the creative process led me in different directions to my initial plan. That happens, writing in an imperfect practice but Scrivener has your back. The outline and corkboard is also really nice for getting a visual overview of your story.


Another great feature that allows you to merge two or more documents into a single buffer. It does come at a performance cost though, especially when doing a big Find and Replace operation like changing your quotes from straight to smart.

Apple's text framework

This is actually far and above anything you find on competing operating systems and is one of the strengths of the platform. Its word count was much more accurate than Gedit and it's built-in spelling system is really good. Also being able to look up a word instantly in the OED is fabulous and much better than the online approach that Linux apps use.

Now for the things I didn't like.

Rich text

I hate working in rich text with a passion. It's distracting and limits what you can edit it with. I spent far too much time configuring the editor because I didn't like the defaults (hopefully that's a one-off event). Scrivener also didn't convert all the MultiMarkdown and CriticMarkup syntax into it's rich text equivalents either.

Bugs, instability and performance

I have to be pretty honest, I was quite frankly pissed off at the bugs, instability and performance issues. I suspect the main culprit was El Capitan but I also got repeated crashes of Scrivener and Aeon Timeline when I attempted to use them together. Moreover, performance was a real dog; I've been so used to how fast and efficient Linux is that it was a real shock to me how slow OS X was at opening files and processing large chunks of text. My Mac mini has twice the memory of my Linux laptop (4GB vs 2GB) and much faster multicore processor (2.5ghz i5 vs 1.4Ghz Celeron) and OS X was giving me beachballs at every turn. To rub salt into OS X's wounds, I'm running Unity on my laptop and it's considered one of the hungriest and most bloated desktop environments for Linux.

Flat folder output

Because there's no iOS version, to work on my iPad (or a computer without Scrivener) I'm forced to export either a Collection to IndexCard or a flat folder containing all the files (and folders) with sequential file numbering. Doing so was an afterthought late at night, so I opted to export to files instead of IndexCard. That left me with nearly two hundred text files, ordered but not structured. Not fun.

Separated notes, metadata and synopses

Visually, Scrivener separates information about your documents into different panels in the UI. At one time, I used to like that, but I now find it a jarring disconnect. With MultiMarkdown and AsciiDoc, I've become used to including metadata in the file's header and notes and comments directly in the text. It's also how I work when I write code. Not only do I find it more convenient but it's also better for portability.

I've changed, Scrivener hasn't and OS X has gotten worse

Most of my issues with Scrivener are my fault because I've changed my personal preferences in how I write and how my computing needs have changed.

I prefer the simplicity and portability of plain text over rich text. Scrivener at its heart is a rich text word processor and it wants me to make a lot of compromises in the process that are too close to vendor lock-in for my comfort.

I prefer Linux to OS X, and Scrivener is first and foremost a Mac app. I'm even starting to prefer iOS to OS X. OS X is really starting to feel like Apple's redheaded step child and I don't think they care any more. So... Linux and iOS... a contradiction in terms? Not really, I want my desktop system to be powerful, stable and friendly to power users. At the same time, I want my tablet operating system to be simple, have a rich eco system and go easy on the battery.

Using the iPad as an accessory device to my Linux rigs seems like a sacrilege but it actually worked quite well. The iPad provided a convenient second screen to access my task lists, outlines, reference material and Apple Music.

So will I stick with Scrivener?

I'm having a hard time justifying it to be honest. The Mac and OS X are rapidly losing their appeal to me. I feel with the Mac being something I'll only use when I have to access a legacy project or a proprietary file format. My relationship with Scrivener will largely depend on what the iOS version is like.

Scrivener's unique proposition early in its life was its organisational tools and it's unified inteface, creating what is essentially a writer's Integrated Development Environment (IDE).

Honestly, I prefer the approach I took last year for my draft of The Florentine Conspiracy. For that I stuck closer to the Snowflake method. I planned in advance and outlined with a spread sheet, which I also used to track my word count and metadata much like Scrivener's outline mode. The chapters were written in a single markdown file with scenes separated with asterisks.

As for dictionaries, I've got them coming out of my ears, including a copy of the two-volume Shorter OED and the equivalent Macquarie dictionary for Australian English. If I want the convenience of an electronic dictionary, I can buy them for iOS. As already noted, the iPad makes a pretty good accessory device for consuming reference material both mine and third-party and it makes good, portable and idiot proof markdown typewriter in it's own right.

The truth is that I don't need Scrivener and I don't need OS X, especially for new projects or even older ones less heavily invested in the platform. Weaver of Dreams on the otherhand is a legacy project and it comes with a lot of baggage. Its original Scrivener project file is enormous, filled with several drafts and countless notes.

Over the next few days, I'll need to think long and hard about if I stick with scrivener or make that clean break at last.

Wow you read this far! This site is archived and no longer maintained. For Chris' main site go to chrisrosser.net