Quivira 4.1


Latest Version

Quivira 4.1

Quivira 4.1 contains 11,053 characters.


Using Quivira

You can use Quivira freely, but there are some technical issues that may make it a bad choice (depending on what you intend to do).

Known issue in MacOS

Installing Quivira on a Mac results in an error message. I do not know why, or how to fix it.


The characters in the blocks starting from “Ancient Greek Numbers” are not rendered correctly in every program.

Combining characters

Quivira contains some combining characters, but no instructions how to place them correctly. Thus most programs will not render them in an appropriate way.


Quivira is currently not kerned at all, but may be in future versions.


Quivira is not hinted at all, and will probably never be.

Social Bookmarks

del.icio.us Digg.com Mister Wong StumbleUpon MySpace Technorati Slashdot Newsvine Live Google Bookmarks Yahoo! Bookmarks Twitter Facebook Reddit Ying Folkd

Using Quivira

From 2019 on, Quivira is in the Public Domain. For legislations where this is not possible, I hereby explicitly grant the right to use, redistribute, remix or adapt Quivira for any purpose, including commercial usage, without attribution or any other restrictions unless required by law.

Of course, I do not take responsibility for any derivatives of Quivira made by other people. The only version of Quivira which is under my control, and hence my responsibility, is the one that can be downloaded from this website.

There are some cases when Quivira might not be the best choice:

Your rendering program does not have to support Unicode in order to use Quivira, but in a non-Unicode environment you very probably won’t have access to all characters.

If you write a text using Quivira and send it to someone else, remember that they need Quivira aswell, especially if you use characters from the Private Use Area. You may include Quivira in a PDF, send the .otf file or add a link to the download page.

¹ Unfortunately the last link I knew of doesn’t work any more.

Known issue in MacOS

If you try to install Quivira on a Mac, you will get an error message about the “post table usability”. Since neither I nor any of my friends own a Mac, I cannot reproduce or analyze this problem. The tables in Quivira are generated automatically and have never been edited by hand, so I am not sure what the error message actually means. I am sorry that I cannot provide any help with this issue.

Since Quivira is in the Public Domain now, maybe someone who owns a Mac can take a look, I’d appreciate any kind of help with this.


The Universal Character Set (UCS) is divided into 17 planes, each with 65,536 codepoints. The first and most important one is the Basic Muliti­lingual Plane (BMP, Plane 0) which is supported by every program that supports Unicode at all. However, a lot of programs do not support the other 16 planes.

Quivira contains numerous characters in the Supple­mental Multi­lingual Plane (SMP, Plane 1).¹ Whether they are displayed correctly, as a missing character or even as two missing characters each, depends on both the rendering program and your operating system.

For example, Mozilla Firefox 3, Microsoft Word 2002 and Adobe Reader 8 (tested under Microsoft Windows XP) all render them correctly. (Word does not show them in its symbol insertion dialogue, but if they are pasted in from another program, they behave just like any other characters.)

If you are using Windows 7, make sure you have the Service Pack 1 installed. It does not only fix security holes, but also some issues regarding Unicode text rendering.

As a first test, look at this character: 𝔔

Generally, it is hard to predict where these characters are displayed correctly and where they aren’t. You will have to test it yourself on your system.

¹ These are the following blocks:

Combining characters

Quivira contains some combining characters, especially in the block “Combining Diacritical Marks”. These characters are meant to combine with the characters preceding them, e.g. an A followed by a combining acute (´) should look like Á and not like A´. This means, they have to be aligned both horizontally and vertically, depending on the preceding character. There are a lot of examples, and they occur in many scripts, not only in Latin.

Since version 4.0, Quivira contains instructions for combining marks and a part of the base letters they can be combined with. Please find an overview here: Combining characters (PDF document, {{Combining_filesize_MB}} MB)

As this script is quite some effort to create, and it also seems to cause some weird problems when it gets too large, I try to keep it as small and simple as possible. This especially means that only the following kinds of combinations are supported:

Some rendering programs are smart enough to render combining sequences (or some of them) correctly even without OpenType instructions. E.g. the Hebrew diakritika work just fine in Microsoft Word, Notepad++, Mozilla Firefox and many other programs under Microsoft Windows.


Kerning is the adjustment of the spacing between two characters dependent on their shapes. For example, in the sequence “Te” T and e are commonly moved nearer together than O and e in “Oe”, because due to their shapes the space between them looks bigger. Two characters with such an adjustment are called a kerning pair, and kerning pairs can be defined within both TrueType and OpenType fonts.

However, Quivira is currently not kerned at all. The problem is caused by the very nature of a huge Unicode font: If a Unicode font providing the complete Latin script has a kerning pair for T and e, it must also handle Ţ, Ť, Ŧ, Ƭ, Ʈ, Ț, Ⱦ, Ṫ, Ṭ, Ṯ and Ṱ followed by è, é, ê, ë, ē, ĕ, ė, ę, ě, ȅ, ȇ, ɇ, ɘ, ə, ɚ, ḕ, ḗ, ḙ, ḛ, ḝ, ẹ, ẻ, ẽ, ế, ề, ể, ễ or ệ. So this single example results in 12 × 29 = 348 kerning pairs. Now remember “Ta”, “To”, “Tu”, “Va”, “Ve”, “AV”, … this results in thousands of combinations, and the ratio of effort vs. gain seems questionable. Currently I prefer adding missing characters over kerning.


Hinting is a technology inside the TrueType format used to adjust the glyphs in a font to the pixel grid of a computer screen. It is an assembler-like language, and hinting instructions have to be defined for every single glyph for every font size (for small font sizes). There are tools to do this automatically, but every result I could get so far was harder to read than the unhinted original font. Thus Quivira is not hinted.

Therefore, Quivira might look more blurred on the screen than other fonts. This does not affect printers, because printers have higher resolutions than screens and do not use hinting instructions anyway.

If and how much Quivira looks blurred, depends on your system and the rendering program. However, as Anti-aliasing, subpixel rendering and auto-hinting algorithms get more sophisticated and more widespread, the problem with missing manual hinting becomes less important over time.¹

Quivira already looks fairly readable in most rendering programs at usual font sizes, and I do not have the time to manually hint the thousands of characters. This means, hinting is not planned in the foreseeable future, and Quivira will very probably never be hinted.

¹ If you use Microsoft Windows Vista, you shouldn’t see any difference at all. Vista uses Microsofts “ClearType” algorithm to smooth the glyphs, and the new Vista fonts like e.g. Calibri aren’t hinted either.