• Member Since 30th Jun, 2014
  • offline last seen 5 hours ago

Chicago Ted


"Friendship" is a magical-class noun.

More Blog Posts104

  • 6 weeks
    Every Page a Painting - Walls of Words

    Yup, hello, it's me, back on my typesetting binge again, with another "Every Page a Painting" to show you. And boy oh boy, do I have a real treat for you this time around: one of my favorite novels on this site, one that hasn't been typeset before. . . well, until now, of course.

    Read More

    2 comments · 72 views
  • 8 weeks
    Every Page a Painting - By Any Other Name

    First of March, it's clear to me
    There's something that's uncomforting. . .

    Here I am again, about a fortnight after the first "Every Page a Painting", locked and loaded with a second one, whether you wanted it or not. Enjoy.

    Read More

    4 comments · 55 views
  • 10 weeks
    Every Page a Painting - Click, Clack, Neigh

    I know, I know, it's quite bold of me to publish this on Valentine's Day of all days, but here it is all the same.

    If you don't like the timing, just come back tomorrow. I'll wait.

    If you're still here and you don't care about when you'd get this, all I can say is buckle up.

    (Disclaimer: everything you see here is work in progress and subject to change.)

    Read More

    3 comments · 75 views
  • 12 weeks
    The Art of Typesetting

    "Hey Ted, remember when you said you'd work on another blogpost right after your last one?"

    Read More

    2 comments · 118 views
  • 16 weeks

    Ah yes, my hundredth blogpost on Fimfiction.

    I know I should try to find one single topic to spend it on, but I've got several going through my head and only one milestone to do it in, so. . . what the hell, I'll just talk about all of them.

    Buckle up; this is a certified Anthology Blogpost.™

    Read More

    4 comments · 171 views
Jan
29th
2024

The Art of Typesetting · 6:39pm January 29th

"Hey Ted, remember when you said you'd work on another blogpost right after your last one?"

I did, thanks for nagging me. But as I'm sure you're well familiar with, Reality™ has a peculiar way of giving you something to do when you've already done that yourself---not to mention, for once, I'd woefully underestimated my wordcount. But after cooking for a month, I've finally done it.

This is a friendly reply to RBDash47; more specifically, to his self-publishing primer. Here, I'll go over this primer, clarifying here, rebutting there, and all the while delve into the art of typesetting.

A Signature Touch

While Dash does a good job breaking down some of the jargon for the lay author, he does miss a few things if you want to take your book a step further. Again, I get it. He points out that it's not meant to be an exhaustive guide to the process; all it does is get you on your feet. And all this supplement is meant to do is exhaust it just a bit more.

But he does miss one crucial part of bookprinting, and I want to set the record straight.

Books aren't laid out in leaves; they're laid out in signatures.

And I don't mean having the author sign it after the fact: a signature in this context is a small booklet from which the leaves and, in turn, pages are made. You might be surprised to learn that assembling a book from just individual leaves is a tedious, delicate, and more importantly unreliable process.

To make a signature, several pages are printed one one sheet of paper. It's then folded in half a few times, with three of the four sides then cut away so they can turn freely, a process known as trimming (which you might know from the term "trim size")---typically it's ⅛ in. (3mm) from these sides. This is done so that any graphics meant to run to the edge of the page actually do that.

But one signature a whole book does not make; otherwise it'd be impossible to hold properly. Instead, many signatures are made, dozens, scores, hundreds even for larger books---then they're sewn or glued together inside the cover. Assuming you assembled them in the right order, congratulations, you've got a book in your hands.

How many pages are in a signature? Good question! The short answer is "it depends." The longer answer is that it depends on:

  • book size (i.e. the final trim size---larger books have shorter signatures, since you can only make paper sheets and rolls so large)
  • paper thickness (you can get away with several folds with Bible paper, for instance)
  • format (books, magazines, and newspapers have different binding and, thus, signature requirements)

Typically with any large commercial printer, you'd want to use sixteen-page signatures only, to keep everything simple for their machines---they run on rolls, or webs, of paper that are cut to size on the fly before being folded. What that means is that you, the author, need to make sure your page count is a multiple of 16: 16, 32, 48, 64, 80, and so on, a practice known as "even working". If it's one or two pages short, the leftover pages are still bound, albeit left blank---a common occurrence even for the big dogs. If it's just over even working, however, that would mean an entire signature would be printed for just one or two pages, wasting paper. And while it's possible to mix signature lengths to create a precise page count, in reality that only adds to the complexity of the printing workflow and, in turn, printing costs. Trust me, it's easier just to stick to one length.

While sixteen-page signatures are the industry standard, many books are printed in thirty-two-page signatures. These are more useful at smaller sizes, but it requires wider webs and makes attaining even working much trickier. Mass market paperbacks (those really small paperback books you might find in bookstores or have read from in school) tend to be printed with twelve-page¹ signatures, but you'd be forgiven if you couldn't tell. Larger books like dictionaries typically use eight- or even four-page signatures. And yes, you can think of a leaf as a two-page signature.

Print-on-demand (POD) companies, however, don't work with webs. They just use regular sheets of paper, fed into several much smaller printers. Signature sizes vary depending on the manufacturer, but you might notice a common pattern emerge:

  • If you're going with Ingram, they use a couple different approaches, depending on the trim size:

    • If the trim size is 6.14×9.21 in. or smaller, they'll use either four- or six-page¹ signatures. This is determined by the total page count of the book, depending on if it's more evenly divisible by four or six (though the smaller trim sizes may have six-pagers). (If it's evenly divisible by twelve, it can go either way.)

      ¹These odd departures are due to a Z-fold.

    • If the trim size is larger than 6.14×9.21 in., they'll use four-page signatures only.
  • If you're going with Lulu or Amazon, both of these companies use four-page signatures across the board.

If you're going to print with any of them, please take note of these counts and lay out your book accordingly. The good news is that, because of their practices, if you optimize for any of the "professional" signature lengths, you're automatically good to go for POD!

In summary: books are printed in small groups of pages, cut and assembled together into a whole. How many in a group depends on where you go, but any author interested in self-publishing needs to abandon any notions of pages per book and start thinking in signatures.

Three Peas in a POD

You might've noticed that I gave only three examples of POD companies above:

This is because they're the only real POD companies, insofar as they actually own their print facilities. So if you're shopping around for other POD services and find a company that isn't one of these three, that means it's a mere subcontractor who may just be looking to take your money. Trust me, it's better to do the work yourself and cut out the middleman.

So that whole spiel on other POD companies Dash mentions? Forget Blurb and The Book Patch. Also, his speculation on BookBaby is correct: it's another subcontractor. Feel free to look at what services they provide, but treat them more like a checklist for yourself than reasons to do business with them.

But Before I Sign Off...

If you're just going with Lulu, everything I've said about signatures applies. But if you want to print with Ingram or Amazon, listen up: I'm about to attach the mother of all footnotes to that.

See, these two companies add what I've come to call a colophon as the last page of every book that rolls off their assembly lines. It contains information about where and when it was printed, meant for their inventorying purposes, not for the end reader.


On the left is Ingram's colophon, in a copy of Starscribe's Fine Print. On the right, Amazon's, in a copy of Jim Theis's "The Eye of Argon".

If you follow my advice on even workings and then take your book to Amazon or Ingram, you're going to get three blank pages plus their colophon at the end. So I'll amend my advice here: always stay one page short of even working. You'd end up with a blank page with Lulu, yes, but it'll mesh perfectly with the other two.

The Out-of-Body Experience

And speaking of pages, let's go over page design.

Dash goes over what the body text is, and I can't refute anything he says there. He even explains display pages (e.g. titles, tables of contents (TOCs), copyright info). But he doesn't explain the other parts of a page; namely, the headers and footers, and the content that occupies them.

These contain the page numbers (duh), and some information about the book you're reading, either the author and book title or the book and chapter titles. Note the hierarchy of the information here; you can't just put the author and chapter title and neglect the book title altogether. This is especially important if you're compiling several novellas into a single volume. Also note that the more important information should be placed on the right pages, e.g. author on the right, book title on the left. But don't feel pressured to put them in there just because I said so; many books go without it altogether.

The page numbers tell you, surprise surprise, where you are in the given matter. You can number the front and even the back matter if you'd like, but it's mandatory for the body matter, simply out of its sheer length. If you do number the front and back matters, it's helpful if you use a different numbering scheme to the body matter, just to clarify which one the reader is looking at. Lowercase Roman numerals is a typical choice for those parts.

"But where on the page should I put them?" you might be asking. If you want to look "professional," then:

  • Page numbers can be placed:

    • Centered in the foot
    • On the outer edge in the foot
    • On the outer edge in the head
  • Author/book/chapter titles can be placed:

    • Centered in the head
    • On the outer edge in the head (but not further out than the page number if that's there too)
    • They can also be omitted entirely.

Just remember that consistency is the name of the game. Once you decide on where they go, that's the way it'll be throughout the entire work. If your story spills across volumes, they should be typeset the same way.

Notice that I keep using "outer" and not "left" or "right." Text tends to disappear into the inner edge, or "gutter." You don't want to print your page numbers outer-gutter-outer-gutter; they'd become unusable otherwise.

Good heads and feet do not call attention to themselves; the focus should always remain on the body text. Don't set them in any bold or unusual typefaces. Use italics or smallcaps for the titles; oldstyle numerals might also be a good idea. This of course depends on whether your typeface has these features, which nicely leads into my next topic.

Just My Type

Now let's talk about the elephant in the room: typefaces. You can't set a book well if you don't have the right type to set it with, after all. The right typeface can make or break a book---after all, it's the very thing your readers will be staring at for hours on end; you owe it to them (and yourself!) not to give them a headache.

Most of Dash's advice applies here, but he doesn't cover the specific tones typefaces carry. Even a novice typophile can tell what tone Helvetica carries (even if they can't name it off the top of their head), or that of Times New Roman---or, to reach for the low-hanging fruit, that of Comic Sans.

Ah, get the point now? Formal or informal, gentle or brutal, or simply as little as possible---a well-typeset book doesn't distract with its choice of type, while a masterpiece enhances the text with type.

You want some recommendations from me? Here are some (just some) to get you started:

  • Serif: these are the ones with the fancy ornaments on the letters, as Dash points out.

    • Garamond: you've heard of this typeface before; every major foundry from the late nineteenth century to today has their own cut; it's considered a timeless classic for a reason. It can be considered the magnum opus of French printer Claude Garamont, as he used it to typeset all his books. You can't go wrong with this one if you're trying to look old-fashioned or be serious.
    • Minion: Robert Slimbach, Adobe's golden boy, does a lot of revivals of oldstyle typefaces, but he also creates originals. His philosophy is that they're borne from handwritten calligraphy before they're cut into lead; while his digital workflow skips the lead, his style echoes a lot of that of the old masters. Since its release in the early 1990s, Minion has served as both Adobe's "default romans" for foreign-script fonts (e.g. Adobe Hebrew) and a testbed for future digital font development. It's a good default "ol' reliable" typeface, neutral in its expression, as used in Robert Birnghurt's The Elements of Typographic Style, and across Ponyfeather's entire catalog!
    • Adobe Jenson: another oldstyle typeface, based off the Venetian style from the early Renaissance. It's my go-to for dark and horror.
    • Arno: another Slimbach original, its italic swashes almost lend a handwritten quality to the text. If you're working with adventure, or maybe high fantasy, this would be an excellent choice.
    • Didone: just like Garamond, it's a classic typeface genre, characterized by its extreme stroke-to-serif proportions and its straight-vertical weight axis. I reach for this whenever I want to emulate a mid-to-late-nineteenth-century look. (The default typeface for TeX output is a Didone variant called Computer Modern, so it's also a good choice for mathematics.)
    • Palatino: originally designed by Hermann Zapf for printing on low-quality paper (a prevalent problem right after World War II), it quickly established itself as a go-to for setting body text, despite not initially being designed for this role. Even today, it's still a budget classic---it's preïnstalled on Windows and macOS, but it isn't Times New Roman. If you want to convey a strict, no-nonsense tone, whether it's academia or technical writing, this is ideal.
    • Vollkorn: now here's a typeface with some meat on its bones! Its cuts all have some boldness even at its lightest weight, which lends nicely to smaller type sizes, allowing a printer to fit more text on each page. It's also my personal choice for slice of life stories, especially in rural agrarian settings.
    • Times New Roman: a typeface that needs no introduction, it's arguably Eric Gill's masterpiece, first commissioned by The Times in 1932. Even after The Times allowed third-party use, it remained a respectable typeface, optimized for cheap paper---until Microsoft started including it with Windows as the default serif typeface, causing it to be overused to the point of losing its charm altogether. Use this if you want a sarcastic tone, maybe to emulate poor-quality works, or just as something to submit to your editor (who would most likely require it anyway).
  • Sans serif: these are the ones without the fancy ornaments. They deliberately eschew them in the name of simplicity, doing the bare minimum to represent each glyph.

    • Neue Haas Grotesk: better known as Helvetica, it's the epitome of the International Typgraphic Style: even while carrying many of the same elements as conventional serif fonts, it just does its job and then gets out of the way. Various foundries have this or some other clone of it; Arial is a spacing-identical substitute shipped with Windows. Use this if you don't want to have any subtext whatsoëver---it's that neutral.
    • Inter: another Helvetica lookalike, except this one's free and open source. Use it for the same purposes without breaking your budget.
    • Univers: on the surface, it looks like just another Helvetica clone, but it's actually an older typeface, first released in 1957 and available in more weights and widths than its competitors at the time. (Nowadays we're reëxploring the concept with variable fonts---more information below.) It can be used in all the same places as Helvetica, but with a more "consistent" form. (Fun fact: this is the typeface used in the Star Wars logo!)
    • Futura: instead of Swiss neutrality, this one is based on the Bauhaus art movement, using only geometric shapes for all its glyphs. It was first released in 1927, but its forms still lend a futuristic tone today. Whether your story is set in 2127, 3027, or even 1997, it'll always look like it's yet to happen.
    • Eurostile: you know that single typeface Hollywood uses for all their science-fiction films and TV shows? It's this one. Use it if you want to have that same tone, but be warned: because of its (over)use in cinema, it might come off as kitschy or schlocky.
    • DIN 1451: not a name you've heard of, but it's one of the more well-recognized things to come out of the Deutsches Institut für Normung (DIN)---right up there with ISO paper sizes, in my opinion. Good if you want to convey an industrial or scientific tone. Windows has its own implementation, called Bahnschrift.
    • Avant Garde: first released in 1970, it was actually first used in the logo for the short-lived Avant Garde magazine. It's rich in ligatures, letters intersecting with one another, allowing anyone to assemble a logo similar to that of the magazine. It pairs nicely with Futura if you're into that.
  • Monospace: with these, each and every glyph occupies the same horizontal space, a feature borne from the brutal mechanical limitations of the typewriter. They may or may not have serifs, but in their artificial, stilted shapes and spacing, that hardly matters.

    • Courier: the undisputed king of typewriter fonts, first created by IBM for their revolutionary Selectric series. They were so confident in its quality that it's now one of the few truly public-domain typefaces, so everybody and their mother has their own take on it. Obviously it's good for emulating typewritten documents, so you can use it to style yourself as a newspaper reporter or police detective.
    • Inconsolata: an open-source clone of Consolas, but in many ways improving upon its inspiration, it's widely used by programmers to do their work, especially since it's optimized for modern high-resolution displays. If you want to do some mockup source code for whatever reason, this is a slightly better choice than Courier.
    • Space Mono: commissioned by Google for their relaunch of Google Fonts, it takes inspiration from the grotesk details of 1960s headlines, combining it with the geometric precision of the Bauhaus movement. If you want to convey the same feeling you get from classic science fiction but don't actually want to stoop to Futura, this is a good alternative.
  • Other: handwritten, fantasy, dingbats, &c.: these are all the fonts that don't fit within the other three categories.

    • Biró Script: if you stare at a handwriting font for long enough, you'll start to notice it looks too regular to be real handwriting---every glyph looks the same. Not Biró. It features a plethora of ligatures and alternate glyphs to emulate sloppy handwriting, complete with rough strokes to make it look like it was scrawled with a throwaway ballpoint pen.
    • Caveat: like Biró, it also features alternate glyphs to avoid self-similarity, but unlike Biró, it features far fewer broken strokes, altogether conveying a more precise and conscious effort. It's also free and open source.
    • Comic Neue: okay, fine, you want to use Comic Sans, but you find it's too pixelated for your workflow. Here's an updated remake, which makes the letters more uniform while still retaining the ironic so-bad-it's-good charm of the original. Also free and open source.
    • Milkshake: this is a brush-stroke handwriting font, and it looks really good at that. It'll stand out against busy backgrounds, it's got a good assortment of ligatures and alternate glyphs, and personally I just find it cute.
    • Kinescope: this one's reminiscent of 1940s cartoons and films, especially those by the Fleischer Brothers. If you're looking for a retro "golden age" look, here you go.
    • Pacifico: similar to Milkshake, it's good for fancy signage, perfect for emulating the 1950s surf culture in California. (If it looks familiar, yes, it's used for Pillowfort's logo.)
    • Chomsky: this fraktur typeface, created by Fredrick Brennan (yes, that guy), was inspired by (and in fact had some of its glyphs traced from) the New York Times masthead. Many fraktur typefaces abound online, but this one is free and open source, has decent Unicode coverage, and in my opinion just looks classier than most.
    • Generation B: originating as an emulation of the handlettering in the title card to The Parent Trap (1961), you being a fan of My Little Pony will instantly recognize it as the "Friendship is Magic" font.
    • Woodrow: another honorable mention, this is the font used for the titles and intro credits in the show. For the first six episodes of Season 1, it was also used in the end credits.
    • Many fantasy fonts aim to emulate foreign scripts. I can't give you examples simply because there are far too many to name, but you know what I'm talking about and where best to use them.
    • Zapf Dingbats: another of Hermann Zapf's magnum opi, you most likely have this already installed, even if the font isn't called Zapf Dingbats. These symbols are so popular that the Unicode Consortium assigned a codespace just for them in 1991.
    • Other dingbat typefaces exist, with just about every theme you can think of.

But whichever one(s) you choose, remember above all else, these typefaces should be used consistently---a good book makes the most out of as few as possible.

(I must disagree with Dash's assertion that sans-serif fonts work better on the screen than in print. Serifs improve legibility at smaller point sizes, regardless of medium; the problem here is not the serif typefaces, but pixel density. Of course they'd become pixelated at 72 or 96 PPI, but on modern displays (in particular phones, tablets, and Kindles) with pixel densities measured in the three figures, the serifs fade out as intended.)

"But I don't have any money!" a few of you surely cry out. Good news: you don't need big budgets to get good typefaces! Around the world, many people create new fonts just for its own sake, and release them for free (quite often open-source too) licensed under the Open Font License (OFL)---not public domain, but it does allow commercial use. Everything Google Fonts offers is licensed that way, for instance.

You'll also notice I've listed a few default system fonts: Times New Roman, Palatino, Helvetica, and Courier. Dash tells you to avoid using them, but I disagree: Microsoft and Apple both throw away million of dollars in licensing so many high-quality typefaces; if they look right for your project and just so happen to be installed on your computer already, then by all means, just use them.

And if you're not satisfied with everything I've suggested (my list is absolutely not exhaustive), just Google around for more ideas. Or ask other typophiles. There are literally thousands of fonts out there, maybe even millions; you'll find the one(s) perfect for you.

This Makes for Good Optics

You might've noticed several font families have way more files than you're expecting, on a factor of four or even five. What's going on here?

A long time ago (and I mean a long, long time ago), all typefaces worked this way. They were all cut from lead, either by hand or by machine, but they always physically stamped ink onto the page. Consequently, only a handful of point sizes existed in those days, but each of those sizes were cut to have characteristics to match.

That all got defenestrated in the 1950s when phototypesetting became the vogue. Now, instead of cutting glyphs to a particular point size, printers had a lens to zoom characters in and out, making them as large or as small as they pleased. And swapping between size-specific "cuts"² would just slow them down, so the entire industry moved away from the practice and adopted a one-size-fits-all solution.

²Same terminology, even though they now existed on photographic film.

This persisted right into the digital age, and very much still exists today, but now, armed with the speed of modern computers, type designers can turn back the clock and restore this lost art.

I'm talking, if you haven't pieced it together yet, about optical sizes.

These are different cuts of the same typeface that are slightly tweaked for a range of point sizes (not any specific sizes anymore, thanks to digital typesetting). Typically for a typeface with optical sizes, you'd get the following:

  • The regular optical size, which is best used for body text. Typically it's optimized for 9-12 points.
  • The caption optical size, which is best used for, you guessed it, picture captions, but also footnotes and anything 8 points or smaller.
  • The display optical size, which is meant for larger text, typically 18-21 points or larger.
  • The subhead optical size, which fills the gap between regular and display.

Even with optical sizes, their actual designated point sizes will vary between them---even from the same foundry---so read its documentation thoroughly. There may be additional, more specialized optical sizes (e.g. Arno has a "small text" optic between regular and caption)---again, check the documentation.

Results May Vary

Although variable fonts are still partially experimental, they do show some promise in simplifying file distribution. Whereas you might need dozens of files to encompass the full weight and optical capabilities of the font (doubled if it has italics), now you'd need just one or two. This is a godsend for websites, which can serve just one variable font file and use some CSS wizardry to render everything with it---but it's also useful for graphic artists, who can define their own variables as needed for whatever projects they're working on.

Currently the OpenType file format defines the following variable axes:

  • Weight
  • Optical Size
  • Width
  • Italic
  • Slant

Typically with some typefaces, however (serifs especially), the italic and slant axes are conflated into one, or are even encoded in a second font file. The optical size axis is often automatic, scaling up or down depending on the rendered point size.

Some examples of custom axes include (but are not limited to) ascender size, descender size, and "casual"---as implemented in Recusive, it can convert a sans-serif typeface into a serif one!

Let's Get Back to the Point

This is perhaps the smallest possible detail to dedicate a section to, but on the right scale it can make a difference.

A "point" is a unit of measurement used for type and nowhere else. How big is a point? That's actually a good question:

  • 1 American point = 0.013837 in. ≈ 0.3515 mm
  • 1 Didot point = 0.375 mm ≈ 0.0148 in.
  • 1 PostScript point = ⅟₇₂ in. ≈ 0.353 mm
  • 1 TeX point = ⅟₇₂.₂₇ in. ≈ 0.351 mm

By and large, however, everything you do on a computer assumes you're talking about PostScript points. Some WYSIWYG programs like InDesign can be set to use TeX or PostScript points, but that's about it. If your workflow requires a different size of point, most likely you know how to define it for yourself.

You'll notice that both PostScript and TeX point definitions are based on ANSI measurements, and no metric equivalent exists. That's not necessarily true: DIN 16507-2 defines a digital metric point as .25 mm, and a hot-metal metric point as .375 mm at 20°C. The former, refered to as quarts (Q) is heavily used in Japanese typesetting, to the point where CSS had adopted quarts for displaying text in 2015.

Another unit of measure is the "pica," by definition twelve points. It's an older name for that size of type, but nowadays is often used for measuring both type and page size. For instance, 1p2 is read as "one pica, two points"---fourteen points, or (assuming PostScript points) about 0.194 inches.

Points are also used for measuring the amount of leading, or space between lines of text. However much leading you add is up to you; typical ratios include 5:6, 3:4, and 2:3. In other words, assuming you use 1p0 type, that would respectively mean 1p2.4, 1p4, and 1p6.

Smaller type and tighter leading lets you fit more words on a page, which can help drive down printing costs, but you also want to balance that with reader enjoyment. Anywhere from 0p8-1p0 will work fine for body text; you may want to go slightly smaller for headers and footers.

To the Edge of the Glyph

As the adage goes, "the Devil's in the details." Here's all the little details in formatting your text that'll just make it pop. And yes, this goes for body text too.

(Do note that what I'm about to show you is as universal as it gets. Not all typefaces are made the same way, and will have different capabilities; if in doubt, check the documentation.)

My Little Dashes

Hyphens, en-dashes, em-dashes: what are they for? How does one use them correctly? Dash (no pun intended) insisted there's a difference, and indeed there is, but he doesn't explain them. For shame, my guy.

  • Hyphens (-) are used for:

    • Compound adjectives (e.g. "one-way street")
    • Complex nouns (e.g. "mother-in-law")
    • In limited cases, splitting compound words to avoid confusion (e.g. "re-sign," "semi-independent,"³ "shell-like")

      ³Another option is to use a diaresis (e.g. "semiïndependent"), though this is far less common and far more limited in usefulness.

    • Attaching the pre-, self-, and all- prefixes; or any prefix to a proper noun
    • Attaching the -elect suffix
    • Adding singular letters to words (e.g. "T-shirt")
    • Combining words and numbers (e.g. "post-90s")
  • En-dashes (–) are used for:

    • Time ranges (e.g. "2010–2019")
    • Attaching affixes to a compound word (e.g. "Academy Award–nominated director," "Elvis Presley–eque fashion")
  • Em-dashes (—) are used for:

    • Introducing and emphasizing more detail⁵ (e.g. "His downfall came from the last person he expected—himself.") (Note: a colon or semicolon can perform the same function.)
    • Introducing and emphasizing a detail in the middle of a sentence⁵ (e.g. "Napoléon Bonaparte—the first French Emperor—placed his crown on his own head.") (Note: commas can perform the same function.)
    • Listing items⁵ (e.g. "These are the colors of the flag—red, white, and blue" or "Red, white, blue—these are the colors of the flag")
    • Marking a sudden change in tone⁵ (e.g. "It was a quiet night on Christmas Eve—but it would not stay that way")
    • Interrupting dialogue, whether by the speaker or something else⁴ (e.g. "I could—no, should go to the movies," "I feel as though something's about to—")

      ⁴In the Commonwealth, a spaced en-dash ( – ) carries out these functions.

    • Attributing a quote to its speaker (e.g. "An eye for an eye only makes the world blind." —Mahatma Gandhi)
    • In some languages, they're used for dialogue (e.g. ―O, miss Douce! miss Kennedy protested. You horrid thing!).
  • And I don't think I need to tell you how to use Rainbow Dashes.

Better Living through Kemistry

Kerning is the distance between two glyphs. Good kerning doesn't consider just the horizontal space; it also takes the slight verticals and diagonals into account. For instance, consider this:

AV

Put a straightedge up to those two glyphs. Notice how the top of the V and the base of the A overlap in the middle; if they didn't, your eyes would perceive a new word in the gap.

The art of kerning is knowing when to kern and when not to kern. Obviously you do not want letters to be too far apart, or else text might look weird; but by the same token, you do not want them to overlap with one another, or else it'll all turn into an illegible slurry of ink on the page. Bad kerning, regardless of which type, is typically termed "keming," a term that in itself provides an example: when the r and n glyphs are kerned too close, they turn into an m.

An example of the former is the "Fresh Avocado" Vine; several letters are spaced too far apart, leading to a Del Taco advertising "fr e sh a voca do". An example of the latter---and yes, Shinzakura, I see you back there---is this box with ten "fuckering" lights. Proper kerning makes efficient, but balanced, use of line space.

Putting It All Together

But kerning can only get you so far. At this point, you're finally, officially making an effect in the quality of your typography, or else you're floundering with some affliction.

Look more closely at that last sentence. Do some of those Fs look off to you?

Ligatures are an ancient practice, older than moveable type. The first ligatures were scribal abbreviations, typically for very common words. "And" is a common one: Latin & (et), Greek ϗ (kai), and Tyronian shorthand ⁊ (same as Latin, but more commonly for Irish agus) come to mind. A quick flourish, combining a few letters of a common word together, made scribal transcription easier, but too many of them made it harder to read.

When Gutenberg first invented the printing press, he noticed that a lowercase F frequently overlapped with several letters, necessitating a few combination type pieces---particularly with the lowercase I and L. Similarly, because the long S was in use at the time, ligatures also had to be made for that combination of letters. In fact, seven ligatures (, , , , , , ) have been encoded into Unicode, at the start of Alphabetic Presentation Forms.⁵

⁵The Consortium has declined to encode additional ligatures, stating that this is best resolved on a font-by-font basis.

Some ligatures have become letters in their own right. Classical Latin ae (/ai̯/) and oe (/oi̯/) had both evolved into /eː/ after the fall of the Roman Empire; consequently, Medieval scribes had to improvise a new glyph to reflect the new pronunciation. Thus, æ ("Ash")and œ ("Ethel") became the norm in Latin texts. These glyphs were later adopted into other languages as distinct letters rather than ligatures: æ is used in Danish, Norwegian, and Icelandic, and œ in French. (This is also the origin of German's (in)famous scharfes S or Eszett (ß).)

Modern digital typefaces (including digital revivals) typically include a Th ligature as well. This trend has garnered some controversy (detractors claim it's a solution looking for a problem; supporters claim it removes a clash between two commonly-adjacent glyphs, which is the whole point of a ligature), so I will leave it up to you to determine if it's right for your usecase.

Beyond the seven found in Unicode, ligatures are absolutely not standardized between typefaces, so seriously look over the documentation to see what's available to you.

Say What?

99% of you reading this have written their work entirely in English, no fancy letterplay at work. If that's the case with you, then this section is not meant for you, so feel free to keep scrolling. If you're part of the 1% who knows what I'm talking about, don't fret---you're in good company.

Multilingual typesetting isn't anything new: it shows up frequently in academia, though it's also used for artistic purposes. However, if you've read those works, more often than not different scripts, or even different glyphs of the same script, will clash with one another, often in blatant ways. Sometimes this is unavoidable, due to poor digital font support (or even from a lack of Unicode support), forcing typesetters to make do with whatever they have on hand.

Thus, achieving script harmony is a tricky process, made all the harder by its niche nature. Luckily, we have a few tricks up our sleeve:

  1. Most languages around the world are written in the Latin script, thanks to European colonization and Catholic evangelism, and most typefaces handle just the Latin script. It used to be that good support went up to Unicode's Latin-1 table, but nowadays most fonts support Latin Extended-A (and possibly parts of Latin Extended-B). Nowadays the gold standard is Vietnamese support (the second half of Latin Extended Additionals).
  2. Even if it's not written in Latin, most serif fonts are "pan-European," meaning they support Latin, Greek, and Cyrillic, the second most common script in the world. It's used primarily by Russian, but you'll also find it in Ukrainian, Belarusian, Bulgarian, and Serbian. The Russian mode of the script has also been adapted to the various languages spoken in Russia, the former Russian Empire/Soviet Union, and even Mongolian.

That said, once in a blue moon, you will get yourself into a situation where you need to use another font for an unusual script. While Latin (and to a lesser extent Cyrillic and Arabic) dominate the planet, you might need to typeset text in Hindi, Chinese, or something else.

The only solid recommendation I can give you is Google's Noto family of typefaces. These serve as the default typefaces for Android, but they've been extended to cover just about every script that you can think of, and many more that you've never even heard of. Its name comes from Google's aim at removing all the square-like placeholder glyphs you might find on the Internet, known as substitute characters or "tofu." They work well with each other due to their consistent stroke size and metrics, but they can work as a nigh-universal fallback for your projects. (Its main weakness is the CJK stuff, but to be fair, you can't possibly cover every single Chinese logogram in a font; you'd quickly hit the 65,536 glyph limit in the OpenType specification.)

If Noto's too bland for you, you can always keep looking around for other fonts for a given script. The more speakers a language has, the more typefaces you'll come across. You might be able to find some decent stuff for something obscure like Canadian Aboriginal syllabics, but if you're in that scenario, you most likely already have your own solution. And remember: as bland as it is, Noto is reliable.

(Here's a tip: if you're looking for CJK fonts, their serif is called ming or mincho, and their sans serif is called gothic.)

Tools of the Trade

I do have a few more options if you're looking for interior software:

  • QuarkXPress: once upon a time, this was the WYSIWYG layout program---then everything changed when the InDesign nation attacked. After being displaced in market share, nowadays it's making something of a comeback by the simple virtue of offering a nonsubscription pricing model. (Gee, it's almost like people only want to pay for something once.)
  • Vellum: if you're on a Mac (sorry, Windows/Linux users!), this is as close to an automated book designer as you can get. Import your text, graphics, set a few parameters, and it just does the work. It does, however, rob you of a lot of the flexibility that comes from doing it yourself, but if you don't care much about that, it's perfect for you.

One More Thing to Cover

Yes, you do need to lay out a cover independent of the interior, but how do you go about it? Most image editing programs, both raster and vector, have the ability to nest subdocuments inside a "master document," which I find helpful for my workflow. Namely, I split my covers into separate panels:

  • Front cover
  • Back cover
  • Spine
  • Front flap (on dust jackets)
  • Back flap (on dust jackets)

In these panels, I put my text (e.g. title, author, description, author bio) and other minor graphics (e.g. imprint logo, ISBN, smaller artwork), granting me the freedom to rearrange these without fear of overlapping other areas of the cover. Meanwhile, I keep the background artwork, if any, inside the master document, unless I have a specific aesthetic reason to do otherwise.

Every Which Way

It's easy to arrange text on the front and back covers, and even the front and back jacket flaps if you're going down that road. But what about the spine?

That actually depends on where you live and what language it's written in. English publishers (both American and Commonwealth) are used to having text run from top to bottom; Scandinavian and Dutch publishers agree with the Anglosphere in this regard. The main advantage of this is that it would be upright when laid face-up on a coffee table, but an even bigger one, in my opinion, is if several books are stacked in this manner, since you can't even see the cover that way.

On the European continent (and in Latin America), spines run in the opposite direction: they appear upright when the book is placed face down. The main advantage is (and please take this with a grain of salt) it's claimed to be easier to read this text when placed vertically, like on a bookshelf.

Even then, it's not perfectly consistent: countries like Slovakia traditionally use continental spines, but in recent years some books have English spines, depending on the publisher. In Russia (and the rest of the CIS), GOST 7.84--2002 prescribes English spines, but most publishers continue to use continental spines anyway.

At least, these rules are true in the European tradition. What about non-European scripts? In the Middle East, where scripts tend to be written right-to-left, continental spines are the norm, ironically for the same reason as English spines. East Asia is unique in that their scripts have traditionally been written top to bottom, so they're already perfectly adapted to book spines. There the rule is always top to bottom, keeping the characters upright, then right to left to the next line.

Of course, if your spine is wide enough, you can just run text horizontally.

What's Your Favourite Idea?

By this point, you've read Dash's primer (at least, I hope you have), applied my own supplements to it, and now you're ready to start printing your book. And if that's all you want to do, then neither of us are going to stand in your way. Have at it!

. . . what's that? Oh, that's right: you're here because you want to learn the art of typesetting, not the compliance to industrial conventions of typesetting. My mistake.

You're here to learn the tips and tricks that'll make your book pop out from the crowd. And I'm going to provide.

That's right, you maverick you, it's time you and I got creative. This is the part of the blog where I finally depart from Dash's advice. Gather 'round.

Being Creative

The Sincerest Form of Flattery

A large part of creativity, ironically enough, lies in imitation. Imitation of what, exactly? That's up to you to discover!

Ponyfeather did this for a few titles. Chris's The Purloined Pony is formatted to look like one of those old Choose-Your-Own-Adventure storybooks. (Initially the cover looked exactly like one, but an informal request from the relevant legal team put a stop to it. Still, the second reïnterpretation looks nice---transformative, not imitative.) And more recently, device heretic's Eternal's cover was made as a facsimile of the Everyman's Library's no-nonsense cover design.

But that's where my praise of Ponyfeather ends, regrettably. Beyond The Purloined Pony, each interior looks cookie-cutter identical to the next---makes for quick typesetting, but even the most brilliant design will become stale with overuse. Don't be like that. Really think about what it means to be creative, both in and out. To Ponyfeather's credit, they do decently with their cover designs---some cheesy, some high-class, but all unique with few copies of one another.

This is more than I can say about Ministry of Image, however. Comically enough, in my opinion they have the opposite problem: cookie-cutter covers, unique interiors. But what makes them a cut above Ponyfeather is that they don't have a POD company printing for them; instead, they've hired out a commercial printer that does things traditionally---after all, none of the three PODs offer sewn-in bookmarks (unnecessary for shorter books, but a nicety for longer ones). Their typography is pretty good too, faltering in some areas, but excelling in others: in particular, their Alicorn Adventures is typeset in Palatino---normally reserved for academia, but since it's in-universe academia written by Twilight Sparkle, it gets a pass from me. All their art, however, is commissioned from Ramiras, across their entire catalog, and I just have to say, I feel bad for the poor guy.

Absolutely Everything has both problems, but in their defense, they specialize in exactly one niche part of the fandom: Fallout: Equestria. As everything they publish falls in that continuum, so they must design to reflect this common setting. I'm happy to say the design they picked works marvelously, though no doubt you'll have some debates on its use in the other books they offer.

There's one book in their inventory that's not published by them, and that's TwistedSpectrum's Five Score, Divided by Four. That was done by Nonexistent Publications, though it was not the first title under their belt. Over the years they've printed several classic titles, many offered elsewhere (but always, somehow, inferior in quality). Nonex very much gets it. They know the design of the book needs to reflect the story printed within, and they've taken great pains to see this philosophy through. The typefaces match the tone and style down to the tittles. Artists are selected and commissioned on the same basis (assuming it gets art at all). They understand that there's no such thing as a single perfect layout, only perfect layouts for any given project. A personal highlight for me is Salvation, which was formatted to look like a late-nineteenth-century English book. Typeface, art, even the punctuation rules are on point.

If you're looking for a certain feel, like a time era or book series, sample actual books from those places and replicate them for yourself. It doesn't need to be a 1:1 copy, either; more often than not a reinterpretation will do a better job.

Go Big or Go Home

Another tip I should mention: all of these books I've mentioned in this post are typeset in a few very similar trim sizes:

  • Most publishers in the United States work with 6×9 in., known professionally as octavo.⁶ Ponyfeather mostly works in this format (with a single exception), as do most independent authors in this fandom.

    ⁶Lulu calls this trim size "US Trade."

  • Another trim size, though less common, is half-letter (variously called "statement" or "digest")---5.5×8.5 in. Absolutely Everything works with this trim size.
  • Non-American publishers, whether they're in Europe, Russia, China, South America, or what have you, will typically resort to A5 (148×210 mm). Both Ministry of Image and Nonex work in this format.
  • Smaller, cheaper books might be laid out on even smaller trim sizes. The smallest offered by the POD printers are:

    • Amazon: 5×8 in.
    • Lulu: 4.25×6.875 in. (same size as a mass market paperback!)
    • Ingram: 4×6 in.

That's all well and good if you need something small and portable for shorter texts, but what if you have the opposite problem? One common solution is to split your work into separate volumes---Nonex, MoI and Ponyfeather all do this. Absolutely Everything, uniquely, contracts a printer that stocks Bible paper, to cram more pages into a similar spine width. I, however, am about to suggest something truly unorthodox.

If you can't add more books, and can't fit more pages into your book, then try fitting bigger pages into your book. While unconventional, it's still compatible with print on demand.

The process is quite simple: you split your book into two columns (or three, if you're daring!). Pages are still numbered, but not individual columns; even so, you run your text left to right between columns, as you would pages. If you're working with newspapers and/or magazines, this is what you need to be doing anyway. Letter (8.5×11 in.) and A4 (210×297 mm) trim sizes are ideal for this task. The downside is the increased overall size (duh) and cost---again, not something you have to do for books per se, but that's another option available to you.

A Little Bit of Pizzazz

Remember when I talked about dingbat typefaces? I know, it's been several thousand words, but bear with me here. Dingbats are graphical elements that aren't text; they're basically types of illustrations, except they're usually repeated throughout a book. These might include arrows, crosses, stars, and . Some specific subcategories include:

  • Fleurons: stylized plants or leaves, these were among the earliest dingbats in the history of printmaking. Typically they're reserved for page decoration, though they often decorate covers and spines.
  • Emoji: you already know what these are. Normally these don't show up in printed literature at all, though I could see a case being made for replicating text messaging or breaking scenes, though it's still a pretty good way to date a book. Remember: a little bit goes a long way.

Dingbat use can be as simple as improvising shorthand for a certain phrase (e.g. using a pointing finger for "See here for more information") or as lavish as decorating the title page. Feel free to experiment and see what works for you, but remember: the absence of dingbats is also a valid design choice, and a very frequent one for a reason.

Completing the Trifecta

Authors work with editors to make the story come out right. Editors work with book designers to make it shine. Book designers typically do not work with authors; I'm about to make a case to the contrary.

Take Background Pony, for example. Anyone's who's read any amount of it will remember the distinct colored text, signifying Lyra Heartstrings, among other things. (Note that I'm trying to be vague here.) Two printings of this story exist. Ministry of Image keeps inventory in stock, but to cut down on printing costs, the colored text were simply retypeset in different fonts---a fact that may upset potential buyers, who long for the colored text, just as SS&E intended. Nonex, meanwhile, preserves the colored text the whole way through.

If SS&E were to rewrite it to remove all the pony stuff (Heaven forbid) and take it to a professional publisher, more likely than not they'd strip out all the color formatting. Color ink is expensive, and they'd cut the print cost in half if they removed it---but in the process, they'd get rid of part of what gave the original its charm.

As it stands, book designers largely take orders from editors, and if they and the author disagree on something, the editor very usually wins. Don't you think that's a bit unfair? While their job is to spot errors in a story, they have a tendency to "over-correct" and in so doing strip out unique voices, storytelling events, and the like, rendering the prose stiff as cardboard and sterile as an operating theater.

Luckily, since you're doing your own typesetting, you get to dodge all that nonsense, since in the world of self-publishing, the author and the typesetter are very frequently the same person (who may also be the editor!). You the typesetter know what you the author want; there's no miscommunication whatsoëver.

Woah There Friend, You Might Need to Slow Down

I've just showed you how to break typesetting convention, so that gives you carte blanche to do as you please. . . right?

Well, not quite.

Conventions exist because they apply to the lowest common denominator for storytelling. Most printed stories don't really play around with the medium, simply because it's a really hard thing to do. The reasons people take notice when the rules are broken are:

  1. It's become a garbled, unreadable mess; the story beneath it all might be good, excellent even, but they had done themselves (or worse, a paying client who had written that story and entrusted you to present it beautifully in print) a disservice in, perhaps permanently, tarnishing it in the eyes of the readers. In short, they don't get it.
  2. The effect is subtle, somewhat distracting, but not enough to detract from the reading experience. But in due course, it becomes clear that the typesetter had used some clever element in their toolkit to enhance the storytelling process, either by their own invention or by the author's, and which might be repeated in future stories. In short, they get it.

What I'm trying to say is, don't just break the rules willy-nilly. Understand what rules you're breaking, and more than that, why you're breaking said rules. Don't break them just for its own sake. Even most "experimental" books still contain many familiar conventions, because they just work.

For instance, just off the top of my head: if you want to add some acknowledgements, normally I would put these at the back of the book. Some authors put them in the front; this is acceptable too. But don't insert them in the middle of the narrative; it just comes off as a quasi-sponsorship: "This narrative brought to you by. . . ."

That's just one instance, but think about what I've just written. What are you trying to do with your book? Need it be produced a certain way, or will any old format work? And how can it be made to suit its story (not vice versa)? These questions should inform the direction you take in your typesetting process.


And there you go, those were all my additions to RBDash47's self-publishing primer. This isn't exhaustive either (I'd be naïve to think so), but hopefully you'll find some part of this useful for your endeavors. If you have any questions about the process, or just want to leave a random "lol ur a nerd," there's the comments section below.

Good night, and good luck.

Comments ( 2 )

And before you go: I'll be starting on my next blogpost soon, showcasing one fanfic I've laid out (but not yet printed per se). Despite some hurdles in the process, I found it was a lot of fun, and the end result looks incredible---though probably not the kind of "incredible" you'd expect.

Stay tuned for that!

I took way too long to finally read this but it was just as fascinating and enjoyable as I was sure it would be!

Thank you for writing this and here's to any future writer/typesetters reading this. You're in for a treat. n_n

Login or register to comment