Evaluating Fonts

Here are some visual and technical aspects of a font that can affect how well it serves a particular language or writing system.

Visual

Text is used in various contexts: print (printed out on paper using a high resolution printer); screen (on a screen or monitor); small screen (a phone). A font should be assessed for all the contexts in which it is to be used and at the various print sizes that it is expected to be used at.

How does text in the font look on the page? Consider some of the following questions as appropriate to the font and script being evaluated:

Intercharacter spacing

Is the intercharacter spacing even or do particular letters jump out as seeming to stand alone because they have extra space around them? Or do two letters seem too close together? Do particular sequences of letters stand out as being a dark area of ink?

Is the space character to wide or too narrow?

Diacritics

Are some diacritics too close to their base characters or too far away? Are their combinations where the diacritic is not in the best position? Do diacritics or other marks on different letters interact with each other inappropriately?

Linespacing

High quality typesetting always specifies the line spacing exactly. But is the default linespacing too tight or wide? Is it hard to set text efficiently without characters bumping into each other?

Evenness

Unless a font is designed to be uneven, a printed page of text should have an even colour of grey over the whole page. A fun way to help assess this is to turn the page upside down.

Sizing

When this font is used with another font that it will often be used with (for example a non Latin font used with a Latin font), are the characters in this font of an appropriate size? Are all the characters in this font appropriately sized in relation to each other? For example Latin digits in a non-Latin font?

Technical

The technical aspects of a font are about how well a font handles the text it is given and how easy the font is to use.

Features

Does the font have the necessary alternatives and features needed for the languages being written with it? Are the default settings for these features appropriate?

Uniqueness

No two different sequences should look the same. If they do, the user may use one over the other and their data will contain both sequences for storing the same content. This is a problem for searching because one string will find one but not the other. So we need different strings to look different so that users can see the difference. There are some exceptions to that rule.

• When the strings are canonically equivalent. Since canonically equivalent sequences are required to behave the same, they should look the same.
• When the string is considered erroneous, e.g. a diacritic follows the wrong kind of base. Different strings may look erroneous in the same way. The important thing is that they look erroneous, e.g. by inserting a dotted circle.