Honing a Cross-Browser Font Stack

It is a testament to many things that I didn’t find this problem sooner, but apparently it is more difficult to arrive at an effective cross-browser font stack than I previously thought.

I have the good fortune to be working on a website redesign with a client and a designer who are realistic about web fonts.  They agreed that while Trade Gothic Condensed No. 18 was ideal for the headlines on the site, Arial Narrow would suffice when users didn’t have Trade Gothic installed.  Thus began my epic quest for a font stack that would work in all browsers.

The Problem

At the outset, I couldn’t figure out why font-family: "Trade Gothic Condensed No.18", "Arial Narrow", Arial, sans-serif; was not doing the trick. It wasn’t just not showing up on one or two browsers, it wasn’t showing up on any of them!

So I did a search for “font-family: trade gothic”. After a few modifications to my search, I found a number of situations where developers had “TradeGothic” in their font stack. Having never seen this space-less construction, I dug deeper.

Eventually, my search turned up a very thorough blog post by Rachael Moore fittingly titled Declare a Safe Font Stack With CSS Font-Family?, the upshot of which was that different browsers use different standards for naming fonts and font families.

Some interpret spaces in the name as spaces, some as dashes, and some prefer no spaces at all.  (I am not too proud to admit that this led me to hours of creating the most enormous font stack ever conceived, where I listed every possible permutation of “Trade Gothic Condensed No.18” with spaces, without spaces, with dashes, without dashes, with some words abbreviated and some not, etc).

Furthermore, some browsers (for instance Chrome) will accept styling inside the family name, e.g. “Arial Narrow Bold”; whereas other browsers (Firefox, for instance) take only family names, e.g. “Trade Gothic”, and relegate styling such as bold or italic (or “Condensed No.18”) to CSS properties like font-style and font-weight.

This last revelation led almost instantly to despair for me.  How was I going to be able to achieve bold for one browser but not the other without some very disappointing hacks? What could I possibly do about “condensed”?

The Solution

Luckily in my case I wasn’t trying to bold something.  Going forward, I’ll probably just always do that with font-weight rather than in the font name.  (Let’s not start thinking about what’s going to happen when I want a non-bolded font to cascade down to a bolded one).

But condensed was pretty key to my current mission.   This is where Moore’s post came in handy once again, as it introduced me to the concept of “font-stretch”.

Never heard of it? Neither had I.  Do a Google search for it. Click that w3schools link and then ctrl-F “font-stretch”.  Weird, right? Read on and you’ll discover what I did, which is that apparently this property is a spottily implemented one at best, and at the very least controversial. Luckily for us, Firefox, the main perp as regards strict font-family names, has started supporting font-stretch.  Also, and this is key, the other browsers seem to ignore it.

So, drumroll please, here is my working cross-browser font stack for “Trade Gothic Condensed No.18”:

font-stretch: condensed; font-family: "TradeGothic-CondEighteen", "Trade Gothic Cond Eighteen", "Trade Gothic Condensed Eighteen", "Trade Gothic", "TradeGothic", "Trade-Gothic", "ArialNarrow", "Arial-Narrow", "Arial Narrow", Arial, sans-serif;

Honing a Cross-Browser Font Stack

3 thoughts on “Honing a Cross-Browser Font Stack

  1. Weird, isn’t it? I still don’t have final answers on *exactly* when and why that whole “space vs. hyphen” issue comes up. I suspect a chart would be required.

    I also suspect that from OS to OS a font can have different names, depending on how it was acquired (ie: included with Word, purchased in an Adobe package, purchased from a font foundry, illegally downloaded, etc).

    I wonder how I got by for so many years without ever noticing it! I’m sure I’ll continue to look into it in the future so that I can come up with some bullet-proof principles for constructing these sorts of stacks. That’s what I’d prefer!

  2. jordyn says:

    I too was really shocked to learn something so new about something so basic. When I have some free time (so, probably 2015) I’ll do some testing myself and maybe build a modest table.

    I just can’t believe there isn’t more information out there on this. Your post was a rare gem!

Leave a Reply

Your email address will not be published.