Updated: • These notes explain how and why I built the field guide the way I did. If you’re curious about digital books, take a peek.

Map and compass

Every project needs a map, a compass and a soundtrack.

All of my decisions are made using that map and compass. Reader feedback, and input from comrades who’ve generously helped me edit the copy and pictures in the book, have helped me refine those tools.

Soundtrack? I’m one of those people who puts a song on loop during a project, and I listened to Halsey’s cover of “The Sound” on BBC Radio1. While updating v2, I played Rosalia’s “Pienso en tu mira” like a ritual.


The long way

From research, to draft and design, to marketing and shipping the final product, I wanted to be hands-on for the whole process. Trading efficiency for a fully DIY approach would give me important insight for the next book… Wouldn’t it?


Edit tight and compromise.

Each photo included in the Field Guide must:

  1. Tell a good story.
  2. Add context.
  3. Avoid redundancy with the usual stuff (Instagram, tourist guidebooks).
  4. Reinforce the tone and message of the Field Guide.

When I started the book, I asked myself a question: Does my reader even need to see the author’s photos? After all, she’s headed to Beijing to make her own.

Test it! In the beta that went out to a hardy group of testers (readers?), the only photo included in the Field Guide was on the cover. Reader response was unanimous: “Where the hell are the photos, man?”

Version 2

All photos are 1000px with a 1:1 ratio.

This update includes 3x more photos compared to the v1 field guide. The choice seemed straight forward: nearly everyone complained about the lack of pictures in the previous version.

While v1 only included monochrome images (which were chosen to reinforce the vibe of the book), I decided to ditch that approach this time because it wasn’t working for my readers.

Kindle v1 Kindle v2 EPUB v2
11 photos 37 photos 37 photos
7978 KB* 13537 KB* 9360 KB†

*According to the Amazon product pages.
†According to my Github repo.

As you can see in the table above, the EPUB weighs less than the Amazon version, which is interesting because Amazon starts with that same EPUB file.

Amazon, iBooks and other platforms request authors upload “original, high resolution” files. (There are some good reasons to request this. There are also some sloppy reasons to request this.) Because I’m now distributing the field guide on Amazon and also directly as an ePub, I optimized image files myself before before packing them. I need to keep my workflow streamlined:

  1. Keep photo dimensions “small,” at 1000px 1:1. Again, navigating with my map and compass.
  2. Compress image files using ImageOptim.
  3. Pack these files and build the EPUB file. (That EPUB is uploaded to my Github repo, and also converted to MOBI and uploaded to Amazon.)

Notes from V1: picture perfect?

Kindle is not a charitable platform for any kind of images – and delivering high-quality photography to readers is especially troublesome.1

Compare Kindle to Apple’s iBooks: Amazon charges authors by file size, for one thing. Meanwhile, the platform’s code limits the responsive options we all take for granted with modern web browsers. And readers’ devices run the gamut. (No more colorspace puns, I promise!)

Research told me many photographers are very choosy about their displays – my images would need to meet the demands of top-shelf phones and tablets.2

But device storage space is truly at a premium during international trips. An ebook like mine would be competing for space alongside movies for the flight, other ebooks and travel apps, and, most importantly for photographers, copies of their own images. It is easy to imagine a Reader backing up her photos to an iPad. She might also proof those images with an app like Lightroom, too.

So, I was much less concerned about showing off high-quality images in the book than I was with the risk of taking up space on Readers’ devices. Then there is that irritable warning from Kindle to customers: “Due to its large file size, this book may take longer to download.”

According to the Amazon product page, the v1 field guide weighed 7978 KB. This is roughly the same file size as a hefty novel or nonfiction book – like Michael Meyer’s great book, “Last Days of Old Beijing.”

Book Published file size
Photographers Field Guide v1 7978 KB
Lonely Planet Beijing 86787 KB
Last Days of Old Beijing 7559 KB

Maps and illustrations

If you can’t get there, you probably can’t photograph it.

Offline maps weren’t part of the plan, at first. If most readers will be reading the field guide on their phone, won’t they just reach for their favorite maps app?

As research progressed I realized that I’d need to include custom maps for 3 reasons. Feedback from early readers convinced me that:

  1. Few travelers are willing to endure the extra hassles caused by the GFW to get to that favorite maps app.
  2. Although there are very good English-language maps designed for tourists, there were still large info gaps.
  3. Maps are an efficient, quick way to organize and explain complex ideas – making them an ideal way to deliver info to my reader.

In the v1 field guide, I included 14 high-level, thumbnail maps alongside the 17 illustrations I’d already made. (Some of these were removed from the v2 field guide.)

Both the maps and other illustrations were created in SVG.

Joni Trythall’s Pocket Guide to Writing SVG is brilliant, and I only wish I’d found it sooner.

Maps were illustrated using Maperative on top of open-source OSM data. Some hand-polishing of the raw XML and the output SVG files was needed. Maperative is a brilliant tool although it takes some extra work to get it up and running on MacOS.

Anyway, none of the other options I found provided cost-effective licensing for ebooks.

Other illustrations were just coded by hand in BBEdit.


Reality is not spec. Face palm.

Originally, all the maps and illustrations in the v1 field guide were to be published in SVG. According to both the ePub spec and Amazon’s spec for their own formats, SVG should’ve worked. Reality is not spec. Early tests (2016) revealed that, in fact, Kindle’s SVG support was too limited.

At that point I also learned that, apparently, some of Amazon’s e-ink hardware couldn’t render PNG. In the v1 field guide, it looked like my only options were JPG for photos, and GIF3 for everything else.

Again, ImageOptim was used to shrink down file size.

Creating these illustrations was the most challenging part of the v1 field guide. Because it was the first time I’d done something like this, the workflow and technical challenges were beastly. It took significantly more time to complete than actually writing and editing the copy.

Worth it? Absolutely. So far nobody has said a thing – which means the visuals are doing their job. You tell me: [email protected].


Version 2

Everything was written in Markdown, using TextEdit. Then, using Pandoc, I built a rough EPUB. This was hand polished with Calibre and BBEdit.

Calibre is very useful for editing and debugging EPUB files. Pandoc can be used for all sorts of things. Both have active communities. If you’re trying to build a book, try them out.

Gitbook changed their product offering, and wasn’t used in this version.

Version 1

My earliest drafts were puked out in MacOS’ TextEdit. Inevitably my best ideas were noted on-the-go with whatever was handy on my phone, including Evernote and video selfies. This mess was then assembled into a working draft.

As the project developed, I needed more structure. Gitbook allowed me to pull my disparate TXT files together. Writing in Markdown is great, so that’s what I did. Being able to easily add my visuals, and edit captions and other bits alongside the rest of the copy in Gitbook was very helpful.

From Gitbook, it was easy to export the whole thing as an EPUB. The output was chatty but good. Anyway, I still needed to get some things worked out in the XHTML files specifically for Kindle. For that, Paul Salvette’s books were an awesome resource for understanding what needed to be done. BBEdit and Calibre got me the rest of the way there.

Writing for mobile

Presentation width is important to the readability with any text.

Assuming that most Readers would be paging through the field guide on their phones and tablets, I wrote for that. And that took some practice. I reread the usual sources. I spent a lot of time working on the flow of information across sections, paragraphs and sentences. I refined my tone, syntax and word choice for the environment I assume my Reader will inhabit4 while reading the field guide.

The AP Stylebook might be Religion, but the good book provides no advice on typefaces, line length or other UI issues.

Butterick and the BBC did have some ideas.

For the v2 field guide, I needed to step back and reorganize the book’s structure once again. Abby Covert’s book, “How to make sense of any mess” and Craig Mod’s essays were very helpful.

Lessons learned

If you’re making a product you will have to experiment, test and revise. Recruit your team early on.

The Photographers Field Guide to Beijing is the first book I’ve written. It’s not the first book I’ve helped produce, though. It’s also not my first big, hairy project.

There were a few false summits.

Again, I’d like to emphasize the importance of having a compass – you need to create a reader persona and learn to consult that document as part of your process. Mine is here.

  1. Find your beta testers (beta readers?) as soon as you can. Treat their time with respect. This feedback is invaluable.
  2. Find a good editor. If you’re including maps, illustrations and photos, pull in a photo editor or a designer, too. These eagle-eyed folks will catch mistakes and help you through the 11th-hour gales.
  3. I was told to divide my time equally between making and marketing. And I should’ve listened.

From the start I knew I wanted to do the whole thing myself. I wanted to understand how it works. It really was worth the extra effort, confusion and truly sickening number of late nights.

Go make a book
You’ll be happy you did

Your feedback on methods, tools and – of course – the field guide itself is always welcome. Let me know what you think: [email protected]


Note 1

There are several challenges with photos and visual content – and Amazon itself isn’t always the limiting factor:

Compression. Kindle automatically compresses images and the rest of the content you upload. This is probably in everyone’s best interest. The scant documentation on images doesn’t serve readers or authors well.

Resolution and colorspace. Mobile devices can be among the most demanding ways to view an image. That said, a 500 ppi mobile display and an e-ink reader with limited processing power want very different things from a photograph. Both Kindle gurus and the platform’s own documentation advise authors to consider their readers’ needs. Good advice!

Grayscale, btw. Kindle documention doesn’t mention which grayscale conversion standard they utilize (or if it’s proprietary, which some authors have asserted). I can only say that it’s worth testing your color illustrations on Amazon’s own e-ink hardware. ↩

Note 2 helped me get a better idea of my reader’s resolution needs. Then it was time to experiment. Like the rest of the content, photos were tested on half a dozen different devices. Then I solicited feedback (and screenshots) from beta testers. ↩

Note 3

ImageMagick’s great convert command let me quickly transform SVGs to GIFs. However, it did have some trouble with the bilingual copy on my maps.

I needed all the visuals to look consistent, and styled the text with a generic font-family="sans-serif". But ImageMagick stumbled with Chinese fonts. The program outputted Chinese text in serifed type, even though it outputted English text in the sans-serif I expected. As I couldn’t find a remedy, the SVG maps were converted to GIF using Inkscape instead. ↩

Note 4

Some apps/platforms do help Readers understand they can tweak UI options (e.g. line height, columns, font choice). It isn’t always so obvious, though: Some of my Readers had trouble with this, which startled me – I’d assumed their meme fluency and Reddit expeditions meant it’d be obvious. ↩