Category Archives: Using Type

Fine Print: Fonts for Small Sizes

Type designed specifically for use at small sizes attempts to solve the problem of too much to say in too little space. When conserving resources (ink, paper, etc.) comes at the cost of legibility, it’s best to plan your type palette around faces made to perform below conventional text sizes. How will you know which faces hold up small? Here are a few principles to consider:

Picture 1

Choose something generally robust.

Picture 3

The x-height should be relatively large. Counterforms need to be accentuated. Spacing should be pretty generous. And careful with fine lines—they tend to disappear. Notice above how ITC Bodoni Six differs from a larger optical size Bodoni Seventy Two. This brings up another point. Sometimes a font’s name includes a reference to its optical size, such as a number, or the terms agate, pearl, or micro. Any others come to mind?

Size-specific optical corrections

Picture 2

As if reproduction concerns weren’t enough on their own (managing inevitabilities such as ink spread and so forth), at small sizes your eyes tend to further distort the letterforms. For example, when strokes intersect, (especially at acute angles) the negative space closes up near the intersection. This is remedied by forcing white up into the crotches and joins by way of light wells or ink traps, as you can see above in Matthew Carter’s Bell Centennial. Other phenomena, such as the flattening and squaring off of curves are counteracted by enlarging the gestures of each form. These are easy to compare in faces that have a range of optical sizes. Without, it just takes testing and practice recognizing what works and why. To get you started in the right direction, here’s a FontList of type that works at small sizes curated by a friend and fellow typographer, Nick Sherman.

Picture 4

Take these principles into consideration and you’ll find they apply not only to print, but also to screen media that’s meant to be viewed from a distance. You may also find, as many have, that the unusual texture of faces designed for small print, such as  Joshua Darden’s Freight Micro above, can be quite appealing at larger sizes. That’s it. Using Type continues here Thursday. Thanks to Nick Shinn’s Scotch Modern for setting the opening title.

Two Spaces After a Period—The Definitive Guide

When you’re typing, you may out of habit put in an extra space at the end of a sentence; That’s two spaces between the closing punctuation of one sentence and the beginning of the next. Or it may not be out of habit, but rather, on principle. Let’s talk about that principle.


Before common use of the space, early Latin writing ran in a continuous script, a stream of letters, line-breaking often in the middle of (what we call) a word. With the addition of spaces, and later, punctuation, words as we now know them developed, and along with them, sentences, grammar. As to the amount of space between sentences, that would vary by hand and circumstance until moveable type.

Screen Shot 2013-11-14 at 6.17.33 PMNicolas Jenson, Venice, 1475.

And after. You see, spaces come in a variety of standard sizes. The earliest works printed from moveable type were all set in justified columns, meaning that the width of each word space expanded or contracted slightly from the normal word space to fit different numbers of characters per line as needed. In these early samples, the amount of space between sentences ranged from quite wide, to slightly narrower than the space between words. The most obvious reason for using a narrower space here is to maintain an even typographic color, since the period traps quite a lot of negative space on its own. By the late 18th century, conventions favored more generous sentence spacing. My take from all this early letterpress stuff—counting the number of spaces after a period is as pointless in samples of early printing as it is in handwritten manuscripts. This holds up until of course, typewriters.

Two-Spaces-1Published by Anson D. F. Randolph, New York, 1864. The text is a Caslon.

With typewriters, we get three things—unjustified type, fixed-width typefaces, and typists. The first two pertain to our discussion because we can now count precisely the number of spaces after the period. The third, because our usual scheme of submitting something to be published just got another layer. The typewritten page became both an interim format and a final format for papers and books. Scholarly papers, for instance, were submitted as typewritten manuscripts which then did the rounds between editors, and on to typographers, and ultimately, typesetters. In contrast, for some publications, the typewritten page (or a copy of it) was the final. Whether two spaces after a period arose as a standard under the first typing instructors’ imitation of the printed page, or the practice began independently of what came before, is unclear. Ultimately however the question of how to compose the input should be settled by its output.

The common convention for published works today is a single space between paragraphs.

If you’re submitting your writing to be published, save the typographer the trouble of finding and removing all those additional spaces by not keying them in in the first place. If you’re producing final output on your own (writing for a blog, etc.) the same rule applies.


Keep putting two spaces in if:

  • You’re typing on a manual typewriter, or your final output will emulate the look of that of one.
  • If you’re specifically pursuing a naïve style.
  • If your style guide requires it.

This last one you should be careful to do begrudgingly, like I do when occasionally required to capitalize the word ‘Internet.’ Which arcane style guide requires two spaces after a period, you ask? APA Style does. MLA, Chicago, and AP have all questioned the usefulness of that second space and decided it’s best to go without.

And that’s it. Thanks for reading. Thanks also goes to Evert Bloemsma’s FF Legato for setting this week’s nameplate. Using Type continues here Thursday.

Doing Responsive Typography

Responsive typography makes it possible to serve typographic compositions that adapt to fit their various environments, resizing, reflowing as necessary to best serve the reader, whether they’re viewing the content on a phone, a cathode ray tube, a large display, in print, or something in between. With this piece, we’ll take apart a simple example of responsive typography I made, and see how it works. Adapting other aspects of typography beyond size, measure, etc. is certainly open to further exploration, but I’m keeping this first one simple.

Responsive typography is just typography.

All normal principles apply. The new principles involved deal mainly with being mindful of in-between states—the areas just outside your control. Ready? Let’s go.


Here’s the demo I made.

I started by setting up a development environment on my computer (an Apache web server), and tested on real devices as I went. I wrote the markup and styling (HTML and CSS) in a text editor (BBEdit). And I did a lot of fiddling with the styles in the browser (using Firebug, an add-on to Firefox).

I built out the structure of a two-column layout with the idea that as the width of the page contracts, the right column is hidden. (I could have just as easily turned the right column into a footnote or something.) I styled the page following my own CSS advice: specify as little as possible. In the HTML header, I added this important meta tag:

<meta name="viewport" content="width=device-width"/>

This prevents smaller-screened devices from displaying a scaled-down desktop version of the page. It keeps the text from getting too small. Don’t underestimate the importance of this step.

Set the body font-size to 100%.

Type that’s sized smaller or larger from here should be specified in em units. Since em units change depending on the font-size specified in a given element, you may want at times to refer to the original em you specified. You can. They’re called root ems, or rem units. In the right column, I give the line-height in rem so that I’m able to maintain the same baseline rhythm as the normal body copy. (Since the body’s line-height is in rems, I could have just as easily set the right sidebar’s line-height to inherit.)

Just as an aside, reorienting from portrait to landscape has become commonly understood as a simple way to make the text a bit larger. Some (those with poorer eyesight) are in fact counting on this. So just because you can maintain a consistent text size across different orientations doesn’t mean you necessarily should.

Screen Shot 2013-11-07 at 7.55.06 PM

Give yourself margins.

One common problem with typography on the web generally is a lack of margins. Setting margins to auto works fine until the window or viewport causes that auto value to go down to zero. So in this example, I put an additional 2em body margin outside the content wrapper’s auto margins.


It’s precisely at this spot where the auto margins hit zero that I place my first breakpoint. More like a transition/squish-point, really. All that changes when the viewport goes down to 52 ems is that the content wrapper goes from a specified 45 ems in width, to taking up 100% width of its parent element, the body. In CSS, that looks like this, just below everything else in the file:

@media only screen and (max-width:52em) {
 #wrapper {
    width: 100%;

Screen Shot 2013-11-07 at 7.56.15 PM

As the two columns shrink in size, their text reflows as needed until the right sidebar column gets too narrow. At this point, for simplicity’s sake, I just take it out, and let the main column take up the full width.

@media only screen and (max-width:45em) {  
  #sidebar {
     display: none;
} Screen Shot 2013-11-07 at 7.57.38 PM And it's that simple.

What’s left to improve here? Well, for starters, I think I could show more variety in terms of type size and webfont specification. Maybe the right sidebar could go up a touch in weight. I could get those baselines to reconcile.  I could demonstrate how to target different screen resolutions differently. Seems like the measure on the iPhone is a bit narrow. And I could definitely do something more interesting with the layout at larger sizes. Maybe add a nav area to show how to scale and selectively show/hide links. There’s really a lot I’ve got left to explore.

I’ll return to this series on responsive design from time to time with demos of stuff I find interesting.

Using Type continues here each Thursday. Thanks to Nicole Dotin’s Elena for setting the nameplate, and of course Michael Abbink’s FF Milo Serif Web for the live demo text.

Understanding Responsive Typography

Forget what I said last time. Opening my mind to all the ways I can think up for how responsive type can work has generated a great delay in me actually coming up with a coherent message on how it’s done. The fact is — even if we limit the scope of our interest to technologies that are currently accessible and mainstream — we still live well under our privilege when it comes to what’s possible with responsive typography. So rather than a show and tell, where I say ‘here’s how it’s done,’ I’m instead focusing on pushing the boundaries of what’s considered responsive design just a little, and giving a few pointers along the way.


What responsive design means now

To most in the field, responsive design means controlling the presentation of a webpage through some carefully written CSS that checks the minimum width of the viewport and rearranges the layout, reflowing the copy and bumping the size of the text up or down accordingly. Often the design transforms along a small set of three or four base layouts, separated by discrete breakpoints.

What responsive design can become

Oliver Reichenstein touched on this in a talk he gave at TYPO San Francisco. He was discussing how for his app, Ai Writer, Nitti Light worked perfectly; but after Apple introduced its higher resolution Retina display, the type was all wrong. Something had changed. It was lighter, but it had also subtly lost its character in the new environment. This problem is something that can be helped by responsive typography. And this is a good starting point, but really only one avenue to explore among the many I alone see. Here’s how I break it down: There are conditions that can be measured, and there are responses that the design can make. And there’s a step in between that deals with how capable and trustworthy the data that measures the conditions is, and how reliably an appropriate response can be formulated and made.

Conditions that may elicit one or more appropriate responses

Size of output, intended purpose of the design, number of persons viewing the content, one Nick Sherman pointed out that I mentioned last week: distance from eyes to the reading surface, are all huge factors in how best to show content. Also, things like facial expressions, or the part of the screen on which the the viewer’s eyes are focused may be particularly useful. If projected, the position, color, texture, and motion of the object(s) onto which the content is projected could open interesting possibilities. Back closer to the here and now, things like region, location, position in space, current speed/momentum, ambient light (value and color temperature) and proximity to other devices are additional conditions to consider.

Ways the typography may respond

The obvious ones are things like font size, length of measure, line height, margins, etc.. Others within present reach are the font’s weight, color, style or family. And just outside of easy reach are movement, perspective, and direct manipulation of a font’s masters, (meaning that the type would respond fluidly across changes in weight, width, or any other arbitrarily defined axis, such as serious to playful, futurist to humanist, Persian to Arabic).

Just to conclude these thoughts for now, I know we’ve taken a hard road to get where we are with webfonts actually working as well as they do, but I wonder if it isn’t time to move on. I hope we don’t get too comfortable with type and typography in its present form.

Using Type will pick this back up on Thursday. Thanks Telefon from Monokrom for setting the title.

Responsive Typography

Responsive typography extends naturally from the idea of responsive design, a term first appropriated to the field of graphic design by Ethan Marcotte, which he took from his observations in a different field, responsive architecture. The main idea is simple: a single webpage being displayed on multiple devices, each with a different-sized viewport, must adapt to function in its different environments. The previous answer to the question was to maintain a separate mobile friendly site, possibly further distinguishing it with a .mobi domain. (I can’t think of anyone offhand who presently does this.)

Picture 1

The current method uses CSS media queries to get a sense for how large a screen the site visitor is seeing. If it’s large, the code styles the page to show everything you’d expect browsing the web with a desktop computer. If it’s small, the styling changes the body text to take up the full column and sizes it to be readable on the small screen. It likely also pares back the user interface to its essential or most useful elements given the context.

Picture 2

Such a set of media queries only takes into account the size (and perhaps tangentially, the resolution or orientation) of the screen, but as new thought on the subject is pointing out, there are a number of other factors to which screen media can and indeed ought to respond. Nick Sherman gave one example, the distance from the reader’s face to the reading surface, in a talk last week at ATypI. The proof of concept was done by Marko Dugonjić. Also, the way a layout or composition responds to its environment or viewing conditions is something still completely underexplored.

We’ll pick up this discussion next week, and I’ll get a into the practice of how responsive typography is done. That’s it for now. Using Type continues here Thursday. Thanks to Alda for setting the title.

Making Your First Font—A Little Guidance

Last time on Using Type, I opened the subject of designing your own fonts. Whether or not you pursue type design, as a hobby or professionally, starting the process of creating a typeface on your own can offer insight into how type works, how to use it well, (and certainly) how much trouble type designers go to to make truly great type. In bringing up the subject last time, I left most of the particulars up to you, but if you’d like a little more direction than that, here it is. By the way, this is just how I approach type design, there is no singular ‘right’ way to go about it.


Mess around a lot first.

Get your ideas in front of you on paper. Your hands will provide special abilities as well as impose specific constraints that the computer won’t. Keep it analog during the exploratory phase. There’s no need to seek out special materials. Whatever you’ve got will do. When you’ve got a concept you feel like pursuing, tighten up your work a bit, and move on to the computer (unless of course you’re producing purely analog type).


If your design is modular in nature, (something to produce in FontStruct or some pixel-based font editor), get right to it. You’ll likely be able to see and test your work right in front of you and probably don’t need my help getting going. Otherwise, there are a number of considerations left to touch on.

Is this type or lettering?

What’s the difference, you ask? Designing type is making a cohesive set of interchangeable letterforms (or characters) that work in every possible combination. (Or at least all the most probable ones.) In lettering, the forms only have to work in a single, specified order. Because the constraints on lettering are less rigid, lettering can do things that type—generally speaking—can’t, such as stray from its baseline or fill some arbitrary shape or dimension. You should experiment with lettering too, but this piece is about designing type.

Start small; Test as you go.


Start with whatever will be used most. If you’re creating a text face, I’d recommend you begin with and spend most of your time on the lowercase. If it’s an all-caps display face, start there. The first thing I do is choose a word or short phrase that includes a good representative sample of the characters, and also one that demonstrates well the overall feel of the face, and get started creating just those characters. Make sure the concept is working well with these before moving on. Begin with your control characters: in the case of a Roman uppercase, that’s H and O. (Lowercase is similarly n and o.)

Picture 1

This is where I recommend you slow down and take your time. Getting the spacing of the straight-sided and round characters solidified gives you the standard by which you space and fit the rest of the characters as you draw them. Remember to export new font files often, print out test files and check your prints.

Keep your alphabets separated.

The capitals, lowercase, and figures (numbers) are all separate things. When drawing the characters in your typeface, fit and space caps with caps, lowercase with lowercase, figures with figures. Then kern like with like, and then kern caps to lowercase, etc.. My own philosophy on kerning: treat it as a last resort. Many spacing problems can be resolved with proper fitting and spacing.

Seek professional help.

Attend workshops, read books, participate in forums such as Typophile’s and TypeDrawers’. Send your favorite type designer a letter. They love letters.

That’s it for today. Using Type continues here Thursday. Special thanks to Gnosis by Gábor Kóthay for labeling the series up top.

Fonts: Making Your Own

We get requests from time to time by young designers wondering where one begins when they want to design their own font. I could suggest that they leave this to the professionals, but that would be a missed opportunity, since it was specifically because I picked up a pencil and drew the type around me that I developed a sense for how different typefaces do what they do, and how type genres relate to one another. So inasmuch as learning to talk about type anatomy and dabbling in type design offers insight into using type well, I’m including this in our series on using type.

Comma,-Apostrophe-1There are plenty of different approaches to type design. Some designers do very tight drawings on paper, others sketch just enough to get the main ideas worked out. Some, from this point (myself included) skip the step of digitizing hand work and compose directly on screen in a font editor. Some draw only on screen (I discourage a beginner from following this method, by the way). Figuring out what your shapes should look like, knowing which to begin with, drawing them, fitting them, testing and adjusting them to work properly in whatever context is required for target output is largely what type design is. And by necessity, the process varies with the requirements of each design.

IMG_1476To the graphic designers interested in type design who are now reading this: If you want to design type as a hobby, or for a one-off project, great! (And if you want to take it further, great!) It’s never been easier to start drawing type. I would recommend beginning with something simple you can complete in a few days’ or weeks’ time, just to get the hang of the entire process. Certain constraints and genres of type naturally lend themselves well to a quick project, such as a pixel-based design, or a constructivist face. I recommend FontStruct for someone who just wants to jump in and start making something: Arrange “bricks” on a modular grid, fill out your character set, and download your font.

Picture 1For designs that don’t work on a modular grid, you’ll need a font editor that allows you to draw and space your own vector shapes. These include Glyphs app, RoboFont Editor, Fontlab, Fontographer, and the open source editor FontForge, among others.

Picture 2An alternative to drawing your own vector shapes is to autotrace work you’ve scanned using a relatively inexpensive program called ScanFont. (There are others.) Depending on the level of quality you’re pursuing, the nature of the design, and your means of reproduction, this might be precisely the solution you’re looking for.

Follow this advice and chances are your first typeface will be a great learning experience (and honestly, I wouldn’t expect much more from it beyond that.) Good luck, remember to move on, and let us know how it goes!

Also, reach out to members of the type community, online and in person. They’re generally very helpful since every one of them has been where you are now. Once you’ve completed your first face, process stories like Tal Leming’s reflection of his new sans, Balto, will be even more meaningful to you.

Using Type continues here Thursday.

Making Do When Accented Characters are Missing

You flow in your text, and then you notice something odd. Not all the characters made it. In their place, a disruptive, boxy, Not Defined character.
Yes, it’s a problem you generally get only when the fonts you’re working with contain limited standard (non-Pro) character sets. The best way to deal with it is often always to upgrade to a font with a Pro character set, but, when you’ve got to work with what you have, you make do. In case you didn’t notice the blips in the text above, they’re marked below:Making-Do-When-Missing-Accented-Characters-3
In this example, the accented character that’s missing up top is a lowercase a with macron. If we look in the Glyph palette, there’s a macron waiting there for us, all by itself. We type in the a, then double click the macron to insert it after.
Put the cursor in between the two, and kern them closer until the macron is in place. Simple. Now do it for all accented characters that are missing. Now replace all instances of the newly fabricated characters with their replacements.

When is this a bad solution?

Almost always. If, for example, it’s a print piece and you’ve got control of the entire process, and the substitutions are relatively few, go for it. Otherwise, you run the risk of having to make and remake these fixes with every revision’s reflow of the text. And on the web, in a PDF or whatever, indexing engines will choke on your fake characters.

Say it’s for print only, isn’t there a smarter workflow than the above?

Potentially, yes. You could flow all your text in, set it in a system font so that all characters are represented, make that into a paragraph style, and carefully set up some GREP styles that substitute the missing characters with their tightly-tracked replacement characters (which each have character styles applied to them). Yeah, or something like that. It gets kind of dicey applying styles on top of styles. And your spacing is bound to be thrown off slightly on the right side of the substituted character.

When will I most likely need to pull a stunt like this?

Proper nouns, such as author names, are often the sneakiest kinds of information that come bearing requirements for glyphs that even your Pro fonts may not have in stock.

We’ll wrap here. Thanks for reading everyone. What questions am I missing? And thanks also to Century Expanded Std for its lead role in today’s story, as well as a thanks to Peter Verheul’s Versa Sans. Using Type continues here Thursday.

Kerning Defaults: Metrics (or Auto), Not Optical

Just to make it extra clear while we wrap the subject of kerning, I do have a preference on kerning defaults, and you should too.

Picture 3

The main message of last week’s piece on kerning is that you should only kern what you have to. Take advantage of the kerning that comes built into your fonts. Do this by setting your default kern settings to Metrics in InDesign, and Auto in Illustrator. Don’t set your default to zero, and don’t set it to Optical.

But Optical sounds so nice.

It does. It’s alluringly named. Here’s what zero does, and what Optical does: Setting the kerning value to zero ignores the kerns in the font data. Setting it to Optical ignores the kerns in the font data, and essentially makes a spacing exception between every character, meaning that it’s more than the troublesome pairs who need it that get kerned, everything gets kerned by a robot that’s not very good at kerning.

Picture 2

The above is InDesign kerning FF Legato based on—respectively—the font’s metrics set by its designer, the font’s spacing values only (no kerning), and based on an algorithm that discards both spacing and kerning values and comes up with its own. Note how Optical sets WA a little loose, and LT and ER tight?

For display work, you’ll probably be able to spot this stuff and fix it. But what I really worry about is when Optical kerning is applied unknowingly and nobody catches it—in body copy. At text sizes, Optical kerning leaves things kind of tight overall and otherwise just slightly off.

How do I set the kerning default to Metrics?

Picture 1

Open InDesign without any documents open. In the Character palette, set the kerning field to Metrics. All done.

Aren’t we making a lot of assumptions here?

Only a couple. One—that the fonts you use most are good. Two, that when you could potentially benefit from Optical kerning, that you’ll know to switch over from your default.

That’s it. Catch Using Type here Thursdays.

Basic Kerning

When type designers draw type, the amount of space each character (or glyph) takes up is carefully set. This includes not only how wide the character is inside its box, but how wide the box is, or how much space is required on either side of the character. Done well, a typeface’s spacing maintains a consistent rhythm between positive and negative forms and along with the letterforms themselves gives the face its texture and distinct color. Its spacing values are intended to work under general conditions, like with like, (meaning lowercase with lowercase, figures with figures, etc.) and within a given size range. Of course, general conditions often don’t apply.


So what’s kerning?

A kern is a spacing exception. Kerning is determining which specific pairs of letters (or glyphs) need adjustment to their spacing, and either tightening or loosening the spacing by adding a negative or positive spacing value, respectively. Generally speaking, all professional fonts have kerning built in. Some designs rely very heavily on kerning, such as script faces or tightly spaced display faces. Some, such as monospaced faces, require no kerning. Most text faces, I’d say, employ kerning only to fix the major problems, such as the large gaps in To, Ta, Vo, Va, etc., and to cause punctuation characters such as dashes and commas to land in the correct spot. See below that when the cursor is placed between the T and o in Valentin Brustaux’s Tiina that the kern value is -64 units. The value is in parentheses because it’s supplied by the font’s Metrics, which should be your default in this field. I mention all this just to reiterate that in most cases, the kerning has been done for you. From there, whether it’s suited to your taste or up to the task you give it is largely a matter of your, the typographer’s, discretion.

Screen Shot 2013-09-05 at 5.55.24 PM

When do graphic designers kern?

Kerning is commonly required when a typeface is being used at a size or scale outside the range for which it was designed and tested. It’s also advisable to check and perform kerning when the tracking or letterspacing of the type you’re working with has been adjusted. Lastly, when you’re stuck with it, you’ll need to kern stuff that wasn’t done properly or at all.

Wait, what’s tracking?

Tracking, or letterspacing is the uniform addition or subtraction of space between all characters in question. In the Character dialog above, it’s the zero value just to the right of the (-64) kerning value. Track type open by highlighting it and using the key command Alt + Right Arrow.

Is it time to kern something yet?

Yes. We’re taking this word ‘WOULD’ set in Mark van Bronkhorst’s Sweet Gothic and kerning it so that it works better large. Note that at the small size it works great, but in the large size immediately following, it begins to fall apart.


To remedy this I track it tighter (shown in the second large sample above) but there’s still something wrong with it. See if you come to the same conclusions by first observing with your own eyes and then highlighting to reveal mine in the hidden paragraph below.

Making it tighter created some new problems. The space between W and O needs to tighten up some. Also the LD combination now feels kind of tight, and needs a little extra space.


There. The top part of the above image remains the same for reference, but the bottom portion is now kerned consistently with my findings above. The word is still somewhat loosely spaced, but unified throughout.

How did we get from there to here?

There are differing techniques, but the main idea is to look at the composition the word or line creates as a whole. If that’s too much to focus on at once, try looking at three characters at a time. When you find a pair that needs adjusting, just put your cursor between the two characters, and Alt + Left Arrow to kern tighter, Alt + Right Arrow to kern looser. Remember to refresh your eyes occasionally by printing out your work or changing the distance between you and your screen. Sometimes it helps to focus on just the form by decoupling form from its meaning with a shift in orientation. Look at it upside down or backwards for a second opinion. You can, by the way, overdo it. The 1970s were marked by a generation of overwrought ad comps. Not to suggest that that particular approach was wrong, only that it’s now tied to a specific period, and, in retrospect it seems to have sacrificed purpose for style. I’m of the more pragmatic school of thought that function should take precedence, and that kerning, when done right, should remain invisible.

What else do I need to kern?

If you’re using multiple fonts on the same line, chances are good you’ll eventually run into a combination that needs kerning. That goes even for a Roman and italic of the same typeface; there’s still no way to kern between fonts. (Though there are some crazy ways to put multiple typefaces into a single font and kern all those with each other.)

Okay, that’s it. Please add your questions. I’m also taking requests. Using Type continues here Thursday.

CSS font-size Property

Just so you know, we’re not all print all the time here at FontShop. In fact a great majority of my work here has been conveyed solely over the internet. In terms of graphic design, print’s my first love, but I distinctly remember some time in late 2009 that I heard something about webfonts, the canvas element, pseudo-selectors, transitions, and a number of other experimental and not-widely-supported web technologies, and thought, ‘I need to really dig down into this stuff.’ One ambitious autoinitiated web project later (it was something similar to Readability, before I knew it existed, indeed, perhaps before it did exist), and I was happily up to my nose in markup, CSS, JavaScript, and preprocessing, feeling good about the direction of the internet and electronic media in general.

Picture 1

Since then I’ve tried to keep up with at least the aspects that touch web typography, but I’m not a current practicing web designer, or a specialist in web composition. That distinction goes to people like Eric Meyer. I am though, like most typographers, a natural fiddler and figure-outer, and so briefly I’d like to share one tiny but important practice regarding sizing type on the web.

body { font-size: 100%; }

This is it. Start by specifying the font-size as 100%. Assuming that you are capable of controlling all other viewing conditions, this is the equivalent to specifying 12pt, 16px, 1em, the keyword medium, or its associated number value, 3. Actually, after testing it, I would never recommend using those old html number values, but the rest hold up. From here, all other font-size specifications should be in em units, e.g. this: h1 { font-size: 3em; }

So why percent? And why 100%?

First, percent holds up where the others don’t, namely in not interfering with font scaling, user overrides, and not wrecking everything in older browsers. About 100: It behaves predictably, and it’s large enough to comfortably read on just about any screen. (The text you’re reading now is set to 100% or 16px.) A caveat: There’s a strain of thought regarding font-size that says 16px is an unusual reference size, so why not shift the scale so that 1em is the same as 10px, 1.4em the same as 14px, and so on? To do so, divide 10 by 16, and set your body font-size at this percent, 62.5%. Then set subsequent elements in relation to this, in em units. This is easily achievable and works reliably in most browsers, but still gives you problems in the older browsers. So weigh your options, and stick with what you know works in the majority of cases.

Speaking of sticking with what works, you should also consider doing the opposite: abandoning older browsers for newer ones, taking advantage of newer CSS functionality, and letting a service such as Modernizr or Sass deliver more primitive CSS to the old browsers.

Hang on, what kind of new CSS functionality?

One that directly relates to this subject is calc(), which lets you do simple math in your CSS. It’s really handy for specifying fractional units simply, and for mixed-unit calculations. Another is a new unit of measure, rem, which unlike em, functions as a constant, referencing the root em rather than the em specified in the parent element.

So what size is it going to be, really?

We’ve arrived at the saddest part of this piece on sizing. At this point in time there’s no definitive answer. With screen resolutions all over the place, a pixel is not a pixel is not a pixel, and while absolute units such as the point, (1 in = 72 pt) ought to stay absolute, apparently doing so would break old designs, and we’re stuck with them being reduced to the task of stand-in multipliers for relative units. So an inch is 96 reference pixels, a point is 1.333333px, and a physical pixel may be anywhere from 1px to 2px or more in reference pixels.

The good news


If there’s anything I’ve learned about web standards in the past five years, it’s that people who care about this stuff are in a position to set the standards. At TypeCon in Atlanta back in 2009, there was a panel discussion about what form commercial webfonts would ultimately take. It was less than a year after that that all major browsers were planning to support the new format, woff. Standards bodies followed the lead of the browsers. So ultimately we’re the ones driving this.

That’s it for now. Thanks to Pieter van Rosmalen’s Nitti for setting this week’s nameplate. Using Type continues here Thursday.

Widows, Orphans, Keep Options

This will be very brief and to the point.

Using Type

Orphans are stranded bits of copy, a single word occupying its own line, such as the word type, above, at the top of the right column. Widows are longer bits of copy, a few words perhaps, occupying a single line that’s become estranged from its native text column, pushed to the top of the next column, alone.

Left unresolved, the presence of widows and orphans reveals to the reader the inattention of the typographer. They throw off the margins; they call undue attention.

Keep Options

In InDesign you can specify a minimum number of lines allowed to break between columns. To do it, apply a paragraph style to the paragraphs you’re working with and set the number of lines in the Keep Options page of that paragraph style’s dialog.

Keep Options in InDesignNote that the default Keep with Next value, 0, is probably best to keep at 0 until you know what you’re doing with it.

There. The two lines jumped over to clothe, feed, house, and otherwise eliminate the orphaned status of this previously lost word, type. Solved, right? No. What we’ve covered here is really just one aspect of the craft. The art of typography is creating the kind of flexible compositions that absorb the shocks—both of the problems, as well as those created by their solutions.

Picture 7

Look at the hole left behind in the left column created by our keep options. To fill it, we may have to bump something from a previous page, arbitrarily resize the text frame (and perhaps its surrounding text frames), try something funny with the leading or tracking, italicize some foreign words, or otherwise do some good old-fashioned copyfitting. Or, as I often do when setting longer works, most or all of these. A good layout will permit this kind of editing.

Okay. That’s it. Thanks to MVB Verdigris for setting our examples. Using Type continues here Thursday.

All Caps feature, Braces, Brackets, Parentheses

Just a quick review of a feature today, and by association, some notes on punctuation.
Roman caps have a single descender, generally speaking; That’s Q. And that means that parentheses, or brackets, etc. that are designed for general usage tend to miss their mark when enclosing copy set in all caps. The lack of descenders causes the braces to appear out of touch with what they are attempting to enclose.


Manually adjusting the height of the brackets in InDesign (as I did above) is one option. There is a better way, and one feature in particular (common to Pro font packages) I’ll highlight here. It’s the All Caps feature.

Screen Shot 2013-08-08 at 5.31.15 PM

When setting type in all caps, just highlight it, hit Shift + Command + K (Mac) or Shift + Control + K (Windows) and if the font supports the feature, the punctuation will spring into place. (Of course, not all fonts do support this feature, nor would all necessarily benefit from having it. Some fonts have punctuation that works fine for lowercase, uppercase, small caps, etc.) Speaking of small caps, enabling that feature may reveal an additional set of contextual punctuation, as well as change the figures to a set more appropriate for an all cap or small cap setting.


Above is Jordi Embodas’s Bulo, top, and Tim Ahrens’s JAF Bernino Sans, bottom, with All Caps and Small Caps features applied to different segments of each line. Note how Bernino Sans All Caps feature takes the cap height down slightly, handy for setting things like acronymns.

Screen Shot 2013-08-08 at 4.58.00 PM

How does one check if these features are available in a given font? Check the Case-sensitive Forms box while testing some brackety punctuation. (Select the gear icon first.) The above is set in Frode Helland’s delicate Vinter.

On the grammatical usage of parentheses, brackets and braces—this varies by culture and convention, but the general rule to follow for American English usage is that parentheses are a first resort, and curly braces are the last, (each [nesting {inside} the] other) when necessary. (If a sentence is contained completely by parentheses, the period at the end is as well.) In math, the order of which set to resort to is the same, but its nesting order is just the opposite. {[()]}. British math is prettier: [{()}].

That’s it. Using Type continues here Thursday.

GREP and Styles

Just a few last things to touch on regarding GREP in InDesign. Last week’s piece went over the basics.


Apply a style

One of the great features that InDesign’s Find/Change dialog offers is the ability to apply styles to what you find. Speaking of style, if you’re not all that familiar with the difference in usage between paragraph and character styles, I recommend looking back over the piece on understanding styles, and a related exercise.

So, say I want all numbers to be sized down by a half a point. First, highlight one such number, size it down appropriately, and keeping it highlighted, create a new character style called, say, fig.

Next, open the Find/Change dialog (it’s under the Edit menu), make sure you’re on the GREP tab, and in the first field, type in \d. Then, in the Change Format field, select fig, the character style you just made. Hit change all, and you’ve now applied the character style to all digits everywhere in your document. (You can change the scope of your finding/changing in the box below the Change to field.) You can also (obviously) narrow the scope of your search by specifying a paragraph or character style in the Find Format field.

Grep style

That’s great, but it still means I have to remember to go back through these steps at the very end of modifying my document in order to catch any digits that weren’t previously converted.

Not necessarily. With Grep styles, set up the rule, and it will automatically be applied when anything changes. To illustrate, let’s apply a character style that turns on the OpenType fractions feature whenever there are fractions in a document.

This all happens inside a paragraph style’s settings dialog, so let’s make sure we’ve got a paragraph style already set up. Okay, I made one called Entry and applied it to all the rows in a table of fractions and their decimal equivalents.

Screen Shot 2013-08-01 at 4.52.30 PM

Next, I turn on the fraction feature through the OpenType menu in the Character palette. Then I keep that fraction highlighted while I create a new character style called Fraction.

Last, I go into Entry’s style settings, create a new GREP style, and apply Fraction to the following expression:


Just to decode the regular expression above, that’s \d any digit, * occurring zero or more times (applied to the thing immediately before the asterisk), followed by \/ a slash (typed backslash slash), followed by the same thing at the start, zero or more digits, \d*

Screen Shot 2013-08-01 at 5.15.12 PM

Line it all up with some decimal tabs, and that’s it. The examples today are set in Jordi Embodas’s Pona. Using Type continues here Thursday.


PS. I’ve sat through a number of talks that encourage designers and developers to learn each others’ skills, and inevitably during the question and answer period the designers ask where to start with programming. My answer? Start wherever you’re comfortable, whether that be editing markup, such as HTML, or writing some javascript in an online environment such as CodePen, but somewhere early along the way, learn something about regular expressions. You’ll find that using them trains your mind to think logically, and that when understood, they can save you a lot of effort. By the way: how do you use regular expressions?

Find and Replace, GREP


Alright, this isn’t necessarily about using type—more like using text—but hang with me and you’ll find the information is still applicable if not ultimately very handy. Much of the initial work of typesetting is rearranging and stripping the junk out of your text, including removing that pesky second space after each period.

Find and Replace

The way most go about it is to replace two spaces with a single space, and repeat the action until the search reveals no matches. You’ve likely grown comfortable with Find and Replace by now, but if your text editor allows for finding and replacing with regular expressions, (InDesign does) you may be surprised by the added potential this offers you. Combined with InDesign’s additional ability to apply paragraph and character styles, you’ve got a powerful tool at your disposal.

Slightly more advanced

Grep, a popular term for searching, finding and replacing text that fits a given pattern, allows you to do the above example in a single step, among other, much more complicated and ostensibly more useful things. In InDesign, GREP sits beside the Text Find/Change dialog, looking inconspicuously similar. The below example shows how to find a period followed by one or more spaces, and replace that pattern (or expression) with a period and a single space.

Grep two spaces after a period

Above To find the period, precede it with a backslash (the period has its own special meaning in regular expressions; the backslash says, “nope, just looking for a plain old period,”). This is followed by one space, a second space and an asterisk, (meaning one space, followed by zero or more spaces). When this expression is found, we change it to a period followed by a single space. Similarly this method can be used to replace multiple paragraph breaks:


with a single paragraph break:


or any similar simple substitution.

Replace it with the thing I found

Sometimes when I’m finding and replacing, the thing I want to put in the Change to field is based on what I just found. Say I’ve got a list of names, Last comma First, e.g. Tanner, Jerry; that I want to change to First Last, Jerry Tanner. With Grep, just specify that everything preceding the comma is the (first pattern), everything after is the (second pattern), and then change it to (second thing) (first thing).

Picture 1

Above I use a simple expression: One pattern of zero or more characters, followed by a comma, space, and a second pattern of zero or more characters. The period means any character, with certain exceptions. The two patterns are grouped in parentheses, making it possible to recall them later in the Change to field. Dollar sign 1 is the first pattern in the expression, which we put last. The comma space we ignore. And the last pattern we put first with dollar sign 2. Since the period character doesn’t include paragraph breaks, all we have left to do is hit Change All. Before and after below.


With this level of knowledge, you’re ready to hack into all kinds of text files, turn hyperlinks into labeled URLs, you name it. Next week I’ll go into a bit more styling, including how to apply styles with this and the paragraph style dialog.

In the meantime, how does one learn more about Grep? Adobe has a good resource page on Grep specific to InDesign. There’s also a great general beginners guide  written by Jan Goyvaerts I recommend. Some potentially unfamiliar terminology such as escape and literal will pop up here and there, so keep a programming reference handy, or Wikipedia. Study up on what Grep can do, and you’ll find there’s almost no bottom to the complexity, but—and I’m warning you—don’t get caught up in it. A working understanding of the basics will immensely improve your ability to find and change text. From there, it’s a long descent into the weirdness of cold hard logic and diminishing marginal returns.

Thanks for reading. Using Type continues here Thursday. Thanks to Nicole Dotin’s Elena and Thomas Gabriel’s Premiéra for the illustrations.


Get every new post delivered to your Inbox.

Join 59,465 other followers