Tag: process

how I write

Tracking Nonlinear Work

Writing a novel is something with many phases to it. Some of those phases are short and atomic and don’t need to be broken down into smaller pieces…

Ha ha ha ha! That’s a good one.

No parts of a novel are like that.

It’s All About the Words

Writing is an activity that is most often measured in words or pages: “I wrote 500 words today.”

I use this metric myself during National Novel Writing Month, an approach which is largely required by the goals of the challenge: write 50,000 draft words in a month. Indeed, whenever I am making a rough draft I will use the raw word count as a measure of my progress.

Word count is simple, objective, and linear. It’s very clear how many words you have made at the end of the writing session.

… Except When It Isn’t

How good are those words, though? Are you going to use all of them?

One of the most common objections to the NaNoWriMo process is that the words written are the roughest of the rough. I accept that in my work: a NaNovel is a zeroth draft. It will be filled with inconsistency and bad writing.

Many authors, when they talk about their 500 words for the day, are talking about 500 finished words. My NaNoWriMo word count is raw.

Another metric that I use during NaNoWriMo is chapter count. This also acts as a proxy for plot consumption rate. The way I develop outlines makes this a practical measure for how quickly I am moving through the story I have planned out, and it’s a helpful indicator of how likely I am to finish the narrative arc during the month.

Nonlinear Processes Need Metrics Too

Word generation is, at least for me, a linear process. I don’t delete anything during drafting (my Rule #2 of NaNoWriMo) and so the word count for the book is monotonic.

Revision is much slipperier. Some parts of it are straightforward (copy edits, consistency checks, etc) whereas other parts are iterative processes that require going over the same text multiple times.

What this reveals is that revision is not one process but several. Some of those processes have very clear metrics that can be applied (chapters per day) but others are fractal discovery bug hunts that are as long as a piece of string and twice as knotty.

How should you track non-linear revision progress in a linearly-focussed world?

Track the Work, Not the Goal

My approach is to track the work done rather than progress towards an end point.

Such a tactic is not especially unusual. Events like NaNoEdMo use an hours-of-work goal for the editing you do, and in the past I’ve applied a conversion factor of a thousand words to the hour to make that time look like word counts. This way you’re mapping the time to the words you would have written in that time.

This requires time tracking though, and I am not always consistent enough with that. I like a more objective counter.

What I did for Camp NaNoWriMo this year was to track words using two different strategies:

  1. revision plan — I was finishing my revision plan before I started on the revision itself, so I counted the new words I wrote in that plan.
  2. revision — once I started the editing itself, I counted the words I had revised in a session. So when I finished a chapter I would count the words in that chapter.

    This method has risks, in particular inflation of the work and double-counting.

    • inflation — revising words is arguably not as hard as generating them in the first place, but in revising something you are immersing yourself in the text. Going through a chapter for a particular point, I find I end up reading every sentence, often multiple times. I think it’s fair to count all those reviewed words.
    • double-counting — my current revision plan consists of nine sections for different thematic and textual components: a particular character’s story, a certain plot element that needs to change. When I am working on a chapter, I can’t hold all of those in my head at once, so I work on one or two components at a time, trying to finish a particular section for all chapters.

      This means that I will come back to some chapters multiple times, and that I will count that chapter each time I revise it. Hence I am double-counting.

      I’m fine with this.

      Firstly, I think this accurately reflects the effort put into a revision pass. If I’m working for two or three months on something like this then I want my word count to reflect two or three months of output.

      Secondly, I don’t revise every chapter for every plan section. If a chapter doesn’t feature a particular character then I don’t need to revise it for the character’s back story, for example. In other words, there are critical chapters that will be touched several times, but there are others that might only get a single pass.

I like this approach because it represents pretty clearly the effort put into a revision effort. It didn’t produce outstanding results for Camp NaNoWriMo this April, but then April wasn’t an especially outstanding month so that’s alright.

I’m going to carry on doing this. I think it’s useful.

Leave a Comment

The Failings of Todo Lists

When I was writing the March things post, it occurred that there is an emotional component to todo lists that can be discouraging.

When I am writing a todo list, I often put something like “write todo list” on the list, so that when I have finished making the list I have something I can cross off immediately. This seems healthy to me, a way to get a quick win and boost your confidence about accomplishing the rest of the tasks.

Those kinds of daily todo lists are also healthy in that you can see the whole thing at once, usually. A daily todo for me will be a single small sheet of paper, not usually as small as a Post-It but maybe a 3″x5″ index card, or a quarter sheet of Letter or A4 paper.

But the weakness with the task lists I use in my writing is that they are either too big to apprehend in one go (eg revision plans) or they are too verbose to fit into the window I have for them on screen. That means that I cannot get a clear visual sense of what the state of the tasks is, which means I have no immediate feedback when I do make progress.

An added complication with Three Things as a methodology is that the Things themselves above the daily level tend to be fairly coarse: they will include discovery, design, implementation and verification of something, for example. If I were putting them through an Agile methodology I would break them down into smaller, more bounded tasks with definite completion criteria.

In my day job task lists, I have tried to to mitigate this by listing my sprint tasks (which are bounded and measurable) in a window which fits them. They are always visible, and they can be marked complete easily so I can see my progress.

I need to find a way to do that with my writing tasks too.

Leave a Comment

NaNoPreMo, part 4: Outline

Feel free to ignore this post.

If you are a confirmed pantser who writes without wires or safety net and produces fine stories despite that, then this post is not for you.

I am not that kind of writer. The one year I did not have an outline all the way to the end of the story was the closest I came to not finishing NaNoWriMo. I need an outline, and the outline structure I use is fairly loose. I still want drafting to be fun.

Types of Outline

Outlines come in all shapes and sizes and levels of detail. A list of chapter subjects is a basic outline, as is a paragraph-by-paragraph breakdown of the action, but outlines have the following properties in common:

  • an ordered list of elements forming a narrative
  • the sequence of elements is the way the story is going to be told

So, your story might be performing some structural acrobatics as in Christopher Nolan’s Memento or Iain M Banks’ Use of Weapons, and you might have a full layout of the plot in chronological sequence, but the outline will be in the order the events will be shown to the reader.

Scenes as Use Cases

There are many levels of detail you can work at in an outline, and if you follow something like the Snowflake Method you would start with top level headings and then decompose those into smaller and smaller elements, but I like to work at the scene summary level.

I have a particular format I use for my scene summaries derived from my experience in software: I treat each scene as a use case.

What this means in practice is that for each scene I need the following:

  • who’s in the scene — major characters only. Side characters may be invented as I go and even reoccur, but this captures just the major characters who have a part to play.
  • where the scene starts — “where” means both the location and the emotional state of the narrative.
  • where the scene ends — similarly, this is as much about end state as end location.
  • things that need to happen — there might be particular elements of foreshadowing or character development needed, or just making sure that the MC hets the gun now that they need to use in four scenes, but there are likely to be particular things that have to be in the scene.

And now I need to go and work on my own outline for November.

Good luck in the wars to come.

Leave a Comment

NaNoPreMo 2018, part 3: Structure

Once upon a time there was a writer who wanted to break into the romance market. They researched what was in print, looked at what had been printed historically, what was selling at the time, and then wrote something which tried to avoid repeating the tropes of existing stories — they wanted to produce something fresh, something that everyone would be eager to read because it was new.

But no one wanted it because it didn’t fit: in trying to develop something different, they had abandoned the form that the readership was looking for.

The point here is that people love stories that are, to some extent, like the stories they already know. They want a twist, but only a twist, and people have a strong sense of what makes a satisfying narrative and will drop a book which doesn’t pay that off.

This post is not going to be an exhaustive treatise on structure, but I want to talk about two structural components that I have struggled with and continue to work on.

But first…

Why Structure At All?

Every story has a structure. Readers are looking for structure to help them navigate the narrative, to guide them in understanding the events being related.

Even if you set out without a plan, your story will have a structure. It might be a messy and incomprehensible structure, but one will be there.

If you want to write stories that people want to read, then you should think about a structure you want to use and be intentional about what that structure is.

Because (as I know from years of experience) trying to add structure later rarely works well.

Three Acts

The three act structure has been used as a baseline format on television for decades. It’s not the only structure available, but variations on it deliver stories that make sense to readers.

The version of these three acts that I use is:

  1. Act 1
    1. inciting incident — the thing that happens that starts the story.
    2. first attempt to deal — the protagonist tries to make things better, but makes things worse. Or they do something which should fix the problem, but the antagonist is revealed to be doing things that thwart those efforts.
  2. Act 2
    1. protagonist realise that something is up and gathers resources to try again
    2. second attempt to deal — another go at fixing things, another thing which (mostly) fails. This can include a small victory for the protagonist, but overall the protagonist should understand that there is a lot more to be done.
  3. Act 3
    1. darkest hour — the antagonist takes something from the protagonist thinks they can not do without: their mentor, their lover, their shiny new sword. But in their defeat lie the seeds of victory.
    2. final challenge — the protagonist and antagonist face off and the antagonist is defeatad… for now.
    3. denoument — aftermath of the final challenge, where the protagonist has to accept the changes that have happened.

There are lots of variations on this including pinch points, structures with seven acts and so on, but this is a serviceable form which will produce a sufficiently compelling story.

To give a specific example, Livia and the Corpuscles was built around this form.

Point of View

When we talk about points of view (POV), there are two things we could discuss: voice, and multiple characters. They are related but distinct elements.

The voice choice is usually between one of first person, close third person, distant third person, and some variety of omniscient third person.

  • first person — the words are those of the character. Good to make characters sympathetic because self-image tends to be positive, but very difficult for relating events that don’t happen to the character directly.
  • close third person — a narrator is describing the character, but there is direct insight into the character’s thoughts. Focusses on a single character at a time.
  • distant third person — the narrator describes the character but in a more remote way: unlikely to have visibility of the inside of anyone’s head, allows for more broad descriptions, opportunities for dramatic irony, and so on. The focus is on a single character, but can zoom out to describe things which are a little more distant if need be.
  • omniscient — the narrator knows all. This is where head-hopping can pop up, but basically the narrator relates everything that is relevant to the story.

Changing voice is entirely possible, but always fraught. It’s been two years since I changed Song from close third person to first and I still find incorrect pronouns.

Narrative distance from the character dictates things about how the story is told, but another element of POV which has given me trouble over the years is multiple protagonists. I have tripped over two significant issues:

  1. giving each POV character a meaningful narrative arc. POV characters need to learn and grow, or at least change over the course of the story, so it is important for each of the characters telling the overall story to have their own narrative structure. What I do here is to consider each character independently of the overall story and to prepare three act structures for each.

    This has been particularly valuable as an approach with Song. That is in fact a single POV story, but I found that describing the narrative from supporting characters’ point of view helped me to sharpen the narrative when those characters were getting page time.

  2. handing off from character to character. My early work structures multiple POVs as: each character does a complete thing, then jump to the next character. I am telling multiple independent stories, effectively time-slicing between them*.

    Unfortunately, that isn’t compelling story-telling.

    Switching POV characters should be timed to answer a question raised in another POV. The idea is that the other POV might have information that the first doesn’t have, or might learn something that might be useful if only the other one knew…

    This is something I still struggle with. I have successful short stories with multiple POVs, but no functional novels yet.

Final Encouraging Words

Structure is hard, and as a new writer it felt like too much work, but believe me when I say it is much more work to write an elaborate narrative without a structural plan and then have to try and corral the story into something that a reader might want to actually read.

Go forth and struct your stuff!

[*] consolidating narratives per character is how I will try to rescue that story.

Leave a Comment

NaNoPreMo 2018, part 2: Characters

Last time I discussed building a setting. Now let’s look at characters.

I’ve written before about the Fiasco method for character creation, where you begin by developing the relationships between the characters before deciding any of their identifying properties. In a world of over-elaborate character sheets this might seem the wrong way round, but the point of working on relationships first is that this is a more direct way of uncovering conflict, and conflict means story.

Fiasco Method Summary

This is for developing central characters for your story. Supporting cast don’t need the full treatment, although considering those relationships with the primary cast would be useful.

  1. pick a number of characters. One of these is likely to be the protagonist while one is likely to be the antagonist, but these roles do not have to be defined yet. Identify them with neutral labels like letters, colours, or numbers. You’re going to want at least three, but more than five may become unmanageable.
    You may have well-defined ideas for character properties going in to this process. Note those if they are relevant to the story.
  2. write the labels out in a circle on a piece of paper. You have choices here: if you already know the protag will have relationships with every other primary character then maybe put the protag in the centre with the others in a circle around them. However, in a formless void I recommend a circle.
  3. consider the relationships between each pair of characters at the start of the story. What you’re looking for are relationships that would exist in your setting which explore aspects of the story you are interested in exploring. This may be a place to do some brainstorming to find appropriate terms.
    Tip: the relationships do not have to be symmetrical.
    Tip: there could be multiple relationships between two characters.
  4. assign appropriate identifying properties to the characters.

Example 1: Livia

The idea gathering for Livia started with three characters. These were labelled A, B and C. Then we brainstormed potential relationships and threw them up on the board. We ended up with an eternal triangle with some complexities around accidental incest.

For this process, none of the relationships we came up with were specific to the story, although the character properties were.

Example 2: Spores

Let’s start here with some actual character ideas, because this doesn’t always start from a blank piece of paper.

I’m going to keep this story within the village I posited last time. So, let’s have three characters: X, Y, Z. Imagine these are in a triangle, but I am going to represent these relationships in a table instead. I am going to take advantage of this tabular format to talk about the characters’ self-image.





self: there is a path for X in the village – become a farmer and till the land his family owns. He’s not interested in that, but he is well liked despite that uncertainty.
X is also going to transform pretty early in the story.
X is Y’s younger brother. He has always been admiring of Y, but is appalled by Y taking on a purity patrol roleX knows Z because everyone knows everyone in the village, but Z is a lot older and has never beene tolerant of children.


Y is X’s older sister, but has never been protective of him in the way some kids are. There is about five years between them. Y is proself: Y is proud to have joined the purity patrol.Z is Y’s leader in the purity patrol. She doesn’t trust him because of how he looks at her, but she is loyal to his leadership of the patrol.


X is just another kid taking up spacefinds Y very attractive. Will not make a move on her for any number of reasons (he’s married, she’s a subordinate, she’s too young), but he’s not very good at hiding his feelingsself: Z is secretly one of the transformed that it is his job to remove from the village. His power is subtle (heightened sense of emotional states) but he quickly realised that the best way to avoid detection was to be the detective.

So, that’s a nice web of conflicts. Let’s assign the characters some identifiers:

  • X — Xero (this is future, so going to use some weird names). Late teens, say 17, which means he thinks he’s safe from transformation (most of the transformed do so before the age of 15). He has just started working on the farm after his schooling.
    Xero’s transformation happens after a sporeswarm, and he starts to exhibit matter phasing when he drops a glass during dinner: it passed right through his hand.
  • Y — Yasmine (but some normal names). Early twenties. She has been working in the purity patrol for about a year, and we first meet her as she is ejecting a transformee, one of the kids she used to babysit. She is not callous, but she is remorseless.
  • Z — Seth. Has led the purity patrol for ten years, and served in it for fifteen years before that. He transformed late (even later than Xero) which helped him remain hidden. That was when he resolved to join the purity patrol.

Next time I’ll talk a bit about structure.

Leave a Comment

NaNoPreMo 2018, part 1: Setting

This series is about writing speculative fiction, so I’m going to think about setting first. Literary fiction is led by characters, and characters are enormously important in any storytelling, but the “speculative” part of spec fic tends to be around the setting.

Developing A Setting

It’s quite possible that you have an idea for a setting that you like. It could be as simple as “our world, but vampires are real” or as elaborate as Middle Earth*, but if you don’t have a firm idea then here is an exercise which might help.

First of all, make sure you time, a quiet space, and some mental energy. Doing this last thing at night after a skinful is not recommended. You want to be alert and rested.

Sit down with a piece of paper and the writing implement of your choice and then think about the kind of stories you want to tell and the kinds of things that make those stories unique. Do that for a few minutes.

Then write down words that describe the places those stories will happen.

This is a brainstorming exercise, and the trick with brainstorming is to not censor yourself. Write down whatever comes to mind and don’t criticise what you’ve written. It’s OK to read the words because they will probably spark more ideas to put on the page, but thinking negative thoughts about what you’ve written now won’t help you.

(think of this also as good practice for November. You’ll be writing freely without self-critical thoughts then, as well)

Once you’ve got some words down (however many you think is enough, but a dozen or two would be my target) then review them. Which ones pique your interest? Which ones inspire more ideas than the others? Which ones seem most awesome when combined?

Pull out the ones that speak to you most, and there’s your setting.

I would recommend spending a bit of time asking more questions about the elements you’ve picked. If you have vampires, are they Christian, Buddhist, or Shinto? Can they go out in daylight? How many of the Dracula myths apply? If you have starships, are they faster than light? How is that achieved? Can they be manually piloted? If you have a modern setting, is it the same as our world or different in significant ways?

And so on.

Example Brainstorm: Livia and the Corpuscles

Livia and the Corpuscles started as a brainstorming exercise on a whiteboard. The setting portion of this consisted of throwing a number of setting words up on the board:

  • The Future: in space? post-apocalyptic? dystopian?
  • The Present: Portland? Britain? the Amazon?
  • The Past: steampunk? Arthurian? Roman?
  • Somewhere Else: an alien world? alternate history? elves and goblins?

In this case, of course, I just picked Roman steampunk because it was too awesome to ignore.

Then I picked a few salient facts about the setting: it was a continuation of our Rome on another timeline; human slavery was abolished; the Republic had outposts on other continents.

Example Setting: Spores

Im not going to use Livia as a worked example, but I do want to have an example I can use in future pieces.

A few years ago I was preparing a roleplaying campaign called A New Dawn. When I was writing about that prep I invented a setting I never used for anything, a hot, post-human Earth where plants have awakened to predate upon large animal life:

humanity is in reduced circumstances: climate changes triggered by profligate fossil fuel combustion and misguided efforts to recover methane from deep water methyl hydrates have made temperatures climb; shifting water mass (melting ice and deeper seas) has changed the pressures on continental plates and triggered increased volcanic activity; the seas have risen and weather patterns have thrown agriculture into chaos. People live in sealed cities, or high in the mountains away from the plants.

Ah yes, the plants.

With the increased temperatures, plants have run rampant. Long-suppressed genes for ambulatory motion and other predatory behaviours have expressed, and the herbivorous biosphere is generally in the business of eliminating large animal life. Humans are still high on the food chain, but the top spots are taken by plants.

From this seething, super-evolving biomass emerges superhumans, people who through weird genetic accidents exhibit abnormal abilities: some are expressing long-suspected genes in human DNA, some are mixes of humans with animal or plant.

That setting was intended to provide a large scale playground for half a dozen player characters to romp around in. There needed to be scope for big fights but also intrigue.

For this example, I want to use this setting in a slightly smaller way, make it more personal.

Let’s start with a mountain village in this setting. It used to be bigger than it is: the population is not growing, and people are being lost to the plants and the transformation induced by sporeswarms.

This village does not accept the transformed. Those who change are shamed into leaving, or exiled.

This tells us several things:

  1. there must be some pretty strong social structures in place to force the changed to leave. Maybe there is a charismatic leader, or a strong religious tradition. The former makes the village more isolated, the latter makes it more part of a wider social structure where those exiled will not find sanctuary easily.
  2. there must be a barrier to keep the changed out, so maybe the village is set on a mountain-top with a drawbridge over a deep ravine, or has a substantial wall.
  3. the abilities the changed possess are not generally strong enough to get past this barrier.

I have more ideas here, as I’m sure you do too, but I will save more of those for the plot section.

This is Spores.

Next time I’ll look at constructing characters, because a setting on its own isn’t a story.

[*] which is cheating a little as an example because of course Tolkien developed Middle Earth before he had any stories in mind, but it’s a very elaborate world.

Leave a Comment

NaNoPreMo 2018

October is a month effulgent with promise, for National Novel Writing Month is only a few short weeks away and it is time to think about what to write. October is National Novel Prep Month.

I have a short series of posts to inspire your own preparations for the rigours of November. There are many approaches to prepping for the literary frenzy, but this series is going to follow a plan tuned for speculative fiction:

  1. setting – spec fic is usually grounded in the setting where the story happens.
  2. characters – every story needs characters that people want to read about.
  3. structure – call it plot, call it narrative arc: stories that engage need some kind of structure.
  4. outline – this is a more contentious topic because some writers object to the very idea of outlining their work, and that is their right. I need an outline, though.

You may gather from these topics that I am not really a pantser when it comes to NaNoWriMo: this is the kind of planning I find I need to make the writing function.

By all means follow along as I apply these steps to my own book.

A Note on Content

In this series I’m going to be developing a story which I am not currently planning to write because I already have a story for November this year, but having a concrete example is more helpful I think than talking abstractly about Character A and Character B gallivanting about in Setting Y.

Welcome, therefore, to the world of Spores.

Leave a Comment

Song Completion: Prioritising What’s Left

I am still working on the third draft of Song, and I realised that I was floundering a bit in finishing off the remaining tasks.

My experience as a software developer means I tend to approach writing with a software mindset, and the same applies to process. When you’re not sure what to work on next in software, you look at all known issues and you prioritise them.

For the remaining Song problems, I did the following*:

  1. assigned a reference number to each open issue.
  2. copied the issue numbers and descriptions to another file.
  3. classified the issues. The classes I used were “broken”, “hidden”, “silly”, and “rewrite”. These classes would undoubtedly be different for a different project.
  4. ordered the issues by priority. The classes help with this because anything “broken” is likely to be higher priority than the merely “silly”, but there priorities within a class and I did find some broken things which were less urgent than would be expected.
  5. copied the ordered issues into a new section of the third pass tracking file called “::PRIORITY::” so I can find them again later.

The point of this is to give me a list of things to work on over the next few weeks with the highest impact, because I am going to put this down soon. Novels, they say, are abandoned rather than finished and there will always be things to be improved if I keep looking for them. I need to set an end point, a finishing time. This needs to be done with by the end of October.

Having a prioritised list of things to do means I will get the best value I can out of the time remaining.

How do you plan your completion of a project? How do you finish things?

[*] I edit my tracking files in vim, which makes automating parts of this process quite easy. I started documenting the vim commands I used, but this isn’t a vim tutorial.

Leave a Comment

Song Punch List 3: Punched Out

That took longer than I expected, but I have finally finished working through the punch list comments. All the remaining remarks (the ones with wider scope) are pulled into the third pass edit file.

I’m pretty happy with the consistency of the story, now. What the PLR exposed were niggling things like visual intrusions when the MC is wearing earplugs and other timing issues. There are also a few notes to say that whole scenes need to be rewritten or removed, mostly where I have written things that are just too silly, but definitely closing in on having this draft done soon.

Draft complete this quarter? Possible but unlikely, especially given that I have another story due in a month. The real goal is to finish the draft before November.

Leave a Comment

Song Punch List 2: The Punchening

Last week I said that I had finished the punch list read on Song. Yesterday I finished collecting the comments.

“Collecting the comments” here means walking through the book looking for the remarks I made and doing one of three things:

  1. applying the comment — if it’s a typo or a bit of local rewriting which doesn’t require wide-ranging review of the text then I would just fix it in the text.
  2. recording the comment — larger thematic remarks, substantial new content, or things requiring review and change over several sections of the book were written in the punch list file.
  3. ignore the comment — not common, but sometimes I looked at a comment and thought “this is wrong”, in which case I ignored it.

There is a hidden zeroth step to this process which is to figure out what the comment means. The usual culprit here is autocorrect: I performing punch list reads on a tablet, so the comments are going into Kindle notes. Sometimes autocorrect decides that it knows better and renders the comment incomprehensible. And, sometimes, my comment is just weird and it takes me a bit to reconstruct the meaning.

Once a comment has been processed, I change its colour in the Kindle reader. The default colour is yellow, and I use blue for processed remarks. This is a change: I used to delete the comments entirely, but I’ve realised I am too paranoid about losing information if I destroy the comments.

After all of that, I have 93 comments in the punch list file. I will be working my way through those next.

Leave a Comment