You are here: Home / Publications & Online Resources / Athena Ruby Inscription Font / Athena Ruby Users Manual

Athena Ruby Users Manual

Rough draft, 22 October 2012. This manual assumes of the reader basic knowledge of the Unicode standard as it applies to Unicode Greek.


Athena Ruby is a comprehensive OpenType font for the publication of scholarly editions of Byzantine inscriptions. Named after a predecessor font, Athena, and Dumbarton Oaks Publications Manager Glenn Ruby (d. 2004), Athena Ruby is designed to represent Byzantine inscriptions in Latin and Greek. The character set includes letters, letter variants, ligatures, and decoration found commonly in coins, seals, weights, and other media, such as mosaics and frescoes. The glyphs are idealized replicas of letterforms found in the geographical area of the Byzantine empire, from approximately 325 to 1453. Designed to comply with Unicode and to take advantage of advanced OpenType features, Athena Ruby is suited to a broad range of print and digital publications, from simple, single-author projects to complex ones that involve multiple authors and multiple platforms.

Creative Commons License

Dumbarton Oaks, Trustees for Harvard University, Washington, D.C., holds all rights to Athena Ruby, and makes the font available to the public for free under a Creative Commons Attribution-ShareAlike 3.0 Unported License. We request that anyone using the font in a publication, print or digital, formal or informal, include the following acknowledgment (with the version number in parentheses, and if used in a digital publication a link to the Dumbarton Oaks website):

This publication uses the font Athena Ruby (ver. x.x), courtesy Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Font Architecture


The ca. 730 glyphs in Athena Ruby are intended to represent the major letterforms, punctuation, symbols, and decorations found in Byzantine inscriptions. The font incorporates glyphs that appeared in older, non-Unicode fonts developed by Dumbarton Oaks from 1986 through 2008, and introduces other Byzantine-era glyphs well attested in scholarly publications and from items from our collection. The characters were chosen with special attention to the needs of sigillographers and numismatists. No claim is made that the ensemble of characters is comprehensive, or that it reflects any scholar's opinion about the date and appearance of the various glyphs.

Athena Ruby has the Latin and Greek alphabets (normal forms, variants, ligatures); punctuation (modern and Byzantine); and symbols and decoration. To see a complete list of the glyphs, consult the database (HTML | XML), which includes citations to published exmaples, or the pdf presenting the ensemble of characters.

On the principle that the term minuscule is anachronistic, the font has only uncial letterforms, bound to both upper- and lowercase keys. This accords with the Unicode standard, which makes no normative statements about the appearance of characters (in particular because it anticipates special needs in historic fonts such as Athena Ruby), only their encoding.

The only non-Byzantine glyphs allowed in the font are modern punctation marks – | , . [ ] : · and the combining underdot. Athena Ruby has glyphs in the following tables (in Unicode sequence): Basic Latin (61 points supported), Latin-1 Supplement (2 points), Spacing Modifier Letters (4 points), Combining Diacritical Marks (11 points), Greek and Coptic (49 points), Greek Extendend (167 points), General Punctuation (2 points), Number Forms (1 points), Geometric Shapes (1 points), Miscellaneous Symbols (1 points), Supplementary Punctuation (3 points), Ancient Greek Numbers (2 points). All other glyphs reside in the Private Use Area (PUA). Some characters in the PUA may eventually become candidates for standard points in Unicode.

Supported codepoints: U+0020-007C, 00A0, 00B7, 02C1, 02CA, 02CB, 02D9, 0300-03E1, 1F00-1FFE, 200C, 200D, 2180, 25CC, 2627, 2E09, 2E0A, 2E23, E100-E823, 10182, 10184

Athena Ruby does not support Greek characters that are not attested in Byzantine inscriptions. Thus, for example, it is missing the digamma (U+03DC), sho (U+03F7), and san (U+03FA). It also lacks the full set of ASCII characters.

OpenType Features

Athena Ruby is a True Type-flavor OpenType font. It utilizes several complex OpenType features:

  • Character variants. Unicode can be thought of as a very long two-dimensional table. (See our Guide to Unicode Greek for basic background on Unicode.) OpenType font technology allows type designers to assign numerous glyphs to any single Unicode point, thus adding something like a third dimension to the Unicode table. Athena Ruby employs this technique for letters with historically attested variant representations. The font has, for example, twenty-eight variants of the Greek alpha, each one bound to the proper Unicode point. Many software programs and contexts are not designed to acess Unicode-compliant variants. (That is, they can access Unicode only as a two-dimensional space.) For those instances, character variant glyphs are bound to suitable places in the private use area (PUA).
  • Contextual alternates. OpenType font technology allows for a glyph to change its form depending upon the context. Athena Ruby uses this technology, but only for overbars (U+0305), which change depending upon whether placed over a single letter or a ligature. Not all software programs will take advantage of this feature.
  • Discretionary ligatures. OpenType technology allows several ways to turn sequences of letters into ligatures that retain the underlying codepoint values. Athena Ruby uses discretionary ligatures, which are not turned on by default. These are accessed or permitted within application software. Note, too, that several of these ligatures have variants. One of them (U+0305) also has contextual alternates.
  • Stacking decorative elements in the PUA. Decorative elements in the PUA in the range U+E700..E758 are assigned specialized GPOS kerning values to allow the symbols to stack. To prevent nearby symbols from stacking, one may insert the zero width non-joiner (abbr. ZWNJ; U+200C).

To get the most out of Athena Ruby, you may need to choose a software program different from the one you are accustomed to using. Or you may need to modify the way you use a familiar program. To understand the limitations of specific software environments, and for recommendations on best practices, please see below, Editing with Athena Ruby.


Download Athena Ruby. Go to the directory in which you saved the file and extract it. You do this usually by double-clicking on the file, or right-clicking on it and choose "expand" or "unpack". Find the newly unpacked font. The next step depends upon your operating system.

Right click on the font and select Install. Alternatively, find the drive where you have your operating system installed and copy the file to /windows/fonts.

Macintosh fonts are kept in multiple directories, and many Mac users manage their fonts with Fontbook (part of the OS) or a similar program. (Apple's website provides a good expalantion on how to manage fonts on the Mac.) Copy the uncompressed Athena Ruby file to the directory of your choice, and use your preferred font management program to install the font. If you don't know about font management programs, try double-clicking on the file and then click "Install Font."

After Athena Ruby is installed, open software programs will need to be restarted.

Entering text

To enter new text in Athena Ruby, use any of the methods below.

Just Start Typing: Keyboard Drivers

If you already know how to type in polytonic Greek, you are already set to use the font. If you do not know how, you should install a keyboard driver.

The Macintosh OS (10.4 or later) includes a polytonic Greek keyboard driver. Open system preferences, choose language & text, then input sources. Choose Greek Polytonic. While there, you should make sure to turn on the Keyboard & Character Viewer, good ways to find specific characters. Unfortunately, prior to OS 10.4 (Tiger), polytonic Greek was not one of the options. For those on older versions of OS (and for those who run it, but are unhappy with the configuration of its polytonic Greek keyboard), a third-party driver in polytonic Greek should be installed. Try SophoKeys, a free keyboard driver.

For Windows XP and Vista, there is a Polytonic Greek keyboard driver already built into the operating system, but you must activate it. Microsoft has provided very detailed instructions with pictures here. Unfortunately, this driver's assignment of letters and diacriticals is difficult for many to memorize and use. Other Windows keyboards are listed here. Another alternative is MultiKey, which is not a true universal keyboard driver, since it works only for Microsoft Word and Classical Text Editor.

Pick and Choose: Character Finders and Palettes

Athena Ruby straddles several code tables in the Unicode standard: Latin, Greek, Greek Extended, Combining Diacritical Marks, and the Private Use Area. It will be helpful to have on hand a good palette or finder, to help you explore the font and get the characters you need. You will find this tool helpful even if you are accustomed to typing with a keyboard driver. Warning: most character palettes access advanced glyphs exclusively through the PUA. That is, you won't get Unicode-compliant text if you use glyph finders to copy ligatures and variant letterforms.

Windows operating systems have a small but powerful program called Character Map, which allows you to look through the contents of a font and copy the characters you want. Click the start or windows button, click All Programs then xxx then yyy. A similar but more powerful program is Babel Map.

For OS 10.1-10.6 (for 10.7 and above see below), Mac's Character Viewer and Keyboard Viewer are your best resources for navigating fonts. If you have a flag in the upper right corner of your screen, click on it and look for the viewers. If neither program is shown as an option, select Open Language and Text... and turn the viewers on. Or if you don't see a flag, open system preferences, choose language & text, then input sources. The Character Viewer is powerful, and many of its best features are found only after use. Try choosing different items in the View dropdown menu. Click some of the buttons that appear just below the View dropdown. Click the triangles beside Character Info and Font Variation. The Character Viewer for OS 10.7-10.8 is no help in finding glyphs in Athena Ruby. You will need to use a third-party application, such as Linotype Fontexplorer or PopChar X.

Because the glyphs cover such a wide number of tables, quite separate from each other in the Unicode standard, a palette or glyph finder can be awkward to use. Users may find it more useful to use as palettes the character ensemble on the Dumbarton Oaks Athena Ruby website. Dowload the Word, InDesign, PDF, XML, or HTML document, and simply copy the characters you want.

Editing with Athena Ruby

Athena Ruby is a complex font that uses advanced, specialized features in OpenType. The advanced features are too advanced, in fact, for the software most commonly used in 2012, when the font was released. See here for a list of commonly used programs, and their compatibility with Athena Ruby.

There are roughly three ways to write documents using Athena Ruby:

  1. Use a generic word processor along with a markup language. This method preserves the integrity of the underlying data, separates meaning from appearance, and is best equipped to handle robust layout challenges (both in print and on screen). It is also the least familiar to scholars, and will come in use only gradually, as scholars get more familiar with XML, TeX, and similar technologies.
  2. Use a program that uses advanced OpenType features. Because most software of this kind pertains to specialized typesetting, this method is most suited to publishers.
  3. Use a decent word processor. This method will be the most familiar to most users, but advanced features of Athena Ruby will be available only through non-Unicode compliant features.

Each of these techniques are described below, beginning with the method most familiar (but least ideal), and concluding with least familiar (but most ideal).

Method C: A decent word processor

The most common word processors—Microsoft Word, OpenOffice, Google documents—are popular because they are powerful and easy to use. But the software companies that maintain these programs are selective in the typographic features they support. The advanced OpenType features used by Athena Ruby are usually poorly supported, if at all. Although these programs are not ideal for advanced typography, Athena Ruby has been designed to work with them. Variant letterforms and discretionary ligatures are accessed not through their proper Unicode points but through the PUA. If you are using character finders and palettes (see above), you will find variant forms and ligatures toward the end of the font (U+E100 onward). This method is not ideal because it generates non-Unicode compliant text that cannot reliably searched, sorted, or indexed. To make Athena Ruby text created in this environment Unicode compliant, the text must be converted and exported to an environment that can handle advanced typegraphic features.

This technique will be ideal for pdfs, personal word processing documents, and publications that do not require much professionalism or interoperability.

Method B: Professional layout software

Less commonly used but more appropriate for professional publication are a small group of software programs that take advantage of OpenType features. These programs are specialized and require skill to use competently.

InDesign (Adobe), widely used by publishers, can take advantage of many of Athena Ruby's features. See here for an example file.

Unicode text preservation results may vary. Text should remain Unicode compliant from start to end, although in some cases software may embed raw glyph IDs and not the underlying Unicode points. One needs to turn on discretionary ligatures and stylistic variants for the advanced OpenType features to work. InDesign version 5.5 exhibits unusual behavior regarding complex decorations. In traversing a series of complex decorations, the cursor may erratically jump. It is unclear what creates this effect. Experiments with the same string in other programs do not produce the same results.

TEX is a sophisticated typesetting system that uses text markup to take advantage of advanced typography. Special versions of TeX, notably XeTeX, are also equipped to access ligatures and variant letterforms in compliance with Unicode standards. A sample TeX file is under development.

Method B is ideal for generating more professional-looking pdfs than those generated under Method C, since the software programs provide subtle but important typographic finesse. There is also a greater chance that underlying Unicode data will be preserved.

Method A: An XML/HTML editor (plus markup)

The preferred method for getting the most out of Athena Ruby involves a digital publishing workflow. The stages of that workflow, very broadly, include the following four interrelated steps:

  1. Marking up text. The text meant to be in Athena Ruby is typed in ordinary, plain text in a structured document, most commonly HTML, XML, or XHTML. Specific flavors of XML, such as TEI and EpiDoc, are good candidates, although a given schema may need some customization. The editor deals only with the text's Unicode values, not typography. Those characters that should get special treatment in Athena Ruby are marked with unique, special tags.
  2. Defining what the markup means. The markup codes chosen in the first stage are defined, usually in a stylesheet, to facilitate the next steps. The markup language chosen will dictate the kind of stylesheet to be used. For HTML and XHTML some form of CSS is most common. An XML project might rely on a companion file in XSLT, XML, or a similar structured format.
  3. Anticipating reading environments. One usually chooses a markup language with some sense of how the text will be most advantageous to readers. An HTML file is meant for internet browsers. XML files are more flexible, and can be rendered into web pages, ordinary text documents, or even pdfs. One must decide which browsers, or even which devices, to support. Decisions here may affect previous steps, and may change the workflow in other ways.
  4. Transforming the text documents. Once the previous three steps have been synchronized, a mechanism needs to be set up to take the marked up text (step 1), coupled with the stylesheet (step 2), and generate files appropriate for the specified reading environments (step 3). If a web project, this transformation could be done before files are uploaded to the server (e.g., static HTML files), or a server could make the transformations on the fly (i.e., server-side transformations), or files could be delivered with instructions for a reader's browser to make the transformation after the files are delivered (i.e., client-side transformations). If one intends other types of results, such as electronic publications (e.g., pdf, epub, mobi) one may wish to develop XSLT or XSL-FO files to facilitate the transformation.

These four steps are usually negotiated concurrently, as an author or a team discovers the method that works best for them and their audience. Because the four steps require different skills and expertise, this method is especially conducive to collaborative teamwork.

Users of this method will want to pay attention to the draft of CSS Fonts Module Level 3, which is the incipient standard for how to access and use advanced OpenType features on the Internet.

This method will be deployed most often in HTML-based projects, which are browser-dependent. But not all browsers are created equal. At the present, Firefox (PC or Mac) and Chrome (Mac particularly) best support Athena Ruby's advanced features. The best way to compare behavior is to open the database in different browsers.

The use of Athena Ruby on web-based projects requires some customized work with CSS, to accommodate the needs of different browsers. A block of CSS code using Athena Ruby might look like this:

src:url('athena-ruby/fonts/athenaruby_b017-webfont.eot?#iefix') format('embedded-opentype'),
url('athena-ruby/fonts/athenaruby_b017-webfont.woff') format('woff'),
url('athena-ruby/fonts/AthenaRuby_b017.ttf') format('truetype'),
url('athena-ruby/fonts/athenaruby_b017-webfont.svg#athenarubyweb') format('svg');

Study some of the example projects to get a sense of how to develop CSS that works best for your project.

Synopsis of Methods

Method A
XML/HTML editor plus markup

Method B
Advanced Typographic Software

Method C
Common Word Processors

Who should use Anyone publishing inscriptions with XML or a similar markup language. Especially suitable for collaborative Web-based projects (e.g., EpiDoc, SigiDoc) Publishers and those with specialized software that supports direct input of individual glyphs (e.g., InDesign) Anyone who needs to display a particular variant in software that does not provide access to OpenType Layout or direct glyph input
How to use advanced features Mark variants, ligatures, and symbols with markup tags (e.g., XML); set up stylesheets (e.g., CSS) to render the code correctly. Choose variant letters, ligatures, and other glyphs with built-in advanced glyph palette Use word processor or character map software (e.g., CharMap on the PC, Character Palette on the Mac) to select variants in the PUA codeplane
Ideal publishing venues Print and digital (PDF and web-based) Print and digital (PDF only) Print and digital (PDF only)
Advantages Text remains Unicode compliant from start to end. Does not rely upon specialized software for document creation. CSS standards may be used, so readers need not have the font installed One can see precisely which variant or ligature is being used. Text should remain Unicode compliant from start to end, although in some cases software may embed raw glyph IDs. Can be used in many commonly used programs (e.g., Word, OpenOffice, Excel). One can see precisely which variant or ligature is being used.
Disadvantages Cumbersome to see variants as you type. Training in XML can be daunting, and few use it (so far). CSS syntax for advanced layout features not yet finalized and is only just beginning to be supported in browsers. Limited to relatively expensive, professional page layout software. Unicode text preservation results may vary. This method uses the PUA, which is not recommended because it is idiosyncratic and font-dependent. Text is not Unicode compliant, so cannot be reliably searched, sorted or indexed (to make it Unicode compliant the text must be converted).

Principles of glyph design

The glyphs in Athena Ruby were designed with two guiding principles. The first was to preserve our publishing legacy. The custom typography in Dumbarton Oaks books evolved over time, and glyphs were created to suit a particular catalogue or monograph. This habit was epitomized by Nicholas Oikonomides, who pioneered the categorization and dating of letterforms in Byzantine seals. A Collection of Dated Byzantine Lead Seals (Washington, D.C., 1986) introduced the world's first font for Byzantine inscriptions. This led to a number of other fonts developed specially in the 1990s for publications about seals, coins, and silver plate. We wished to consolidate and preserve that legacy, so Athena Ruby's design conserves much of the aesthetic from Athena and older fonts.

The second principle was to foster a scholarly conversation about Byzantine letterforms. We conservatively anticipated variants, ligatures, and designs that had not been included in our older fonts, but seemed especially useful, and noncontroversial, to distinguish as new types. Our focus was on seals, coins, weights. Some media, such as brick stamps, bread stamps, and frescoes, were also consulted, but not as extensively.

The glyphs in Athena Ruby are meant to evoke, but not replicate, types of letters. Authors who need to replicate exact letterforms not found in Athena Ruby should either augment the font on their own, or work with customized drawings and photographs.

Frequently asked questions

How did you assign characters to the Private Use Area (PUA)? Why doesn't your arrangement coincide with those found in other epigraphic fonts?

Other custom fonts with glyphs useful in Byzantine epigraphy, such as Cardo, IFAOGrec Unicode, Junicode, and New Athena Unicode, also use the PUA. Each font was built with a different rationale for the selection and arrangement of the glyphs in the PUA. That is entirely appropriate, if not expected, since the purpose of the PUA is for private use, and therefore arbitrary decisions. We decided early on that we could not align the characters needed in Athena Ruby with the rationale used by other epigraphic fonts for the PUA.
 Most PUA codepoints in Athena Ruby are concessions. In an ideal world, the R-shaped beta, very common in Byzantine seals, would be accessed by keying in U+0392 then accessing variant 1. But because most software programs do not allow a user to choose variant letterforms, we provided a copy of the glyph at U+E11B. This concessive use of the PUA applies to the first three subblocks (U+E100.., U+E300.., U+E400..). Subsequent blocks (U+E500.., U+E600.., U+E700..) have no corresponding Unicode point, and may never have one. Many of these proprietary codepoints are used in different ways in different disciplines, and there is no guarantee that the meanings imputed by users of Athena Ruby would hold true in other proprietary fonts. So assignments to the PUA have been made with attention only to users of Athena Ruby, to avoid giving the impression, that it is commensurate with other fonts.
 Nevertheless, some of the overlap between Athena Ruby and other epigraphic fonts points to agreement on types of letterforms in the Byzantine period. As documentation for Athena Ruby and other proprietary inscription fonts grow, and as consensus emerges in the field, there should be no obstacle either to finding new proposals for Unicode, or in furnishing a cross-walk protocol, to document corresponding characters from the PUA blocks of different fonts. Discussion about this topic is ideal for the Athena Ruby Google group.

How do you use the overbar in conjunction with ligatures?

If you're using the private use area, use U+0305 immediately after the glyph. If you're using Unicode-based ligatures, use U+0305 after each ligature letter that takes an overbar.

Why don't you include other Greek Unicode characters such as the digamma, the sho, and the san?

The font covers the Byzantine period, so characters that were used before the foundation of Constantinople (324) were not included.

I don't see a special character that I need. Could you create a new version of the font with these new forms?

Dumbarton Oaks will add to the official version of the font only as our publishing needs grow, and even then only as scholars develop a consensus about letterform types. But this should not keep you from developing your own special glyphs, either as an extension to Athena Ruby or as a new font altogether. Under the Creative Commons License users are required to share any modified forms of Athena Ruby. We hope the font encourages deeper study, conversation, and discoveries in the typology of Byzantine letterforms. But that task belongs to the scholarly community.

I already have a document that uses the previous Athena font. Is there a conversion tool I can use?

A web-based conversion tool is under development.

Your database of letterforms is incomplete and incorrect. How do I supply corrections?

Join the Athena Ruby discussion list and pose the correction. Should participants react favorably to the suggestions, the errors will be corrected. Proposed new characters will be held in reserve until our publication needs justify the expense of introducing professionally drawn characters.

I'm setting up a web-based project, but the font isn't showing. What is wrong?

Common problems include the following:

  • Different browsers are suited to different font formats. Make sure you download the version for installing on your server, for online publications (zipped).
  • Perhaps your server has not been instructed to deliver certain files, such as woff, eot, and svg types. The way to check this is to go to the address bar of your browser and type exactly the URL of the font (e.g., It should download fine. If it doesn't or if there are error messages, your server's permissions are likely the cause. Check with your server and see if you can edit the file that regulates your mimetypes with text something similar to this:

    .eot application/octet-stream
    .otf application/x-font-otf
    .svg image/svg+xml
    .woff application/font-woff

  • Make sure your page declares itself to be UTF-8 through a tag such as <meta http-equiv="Content-type" content="text/html;charset=UTF-8" />

I'm using Athena Ruby in an online project. I'm trying to get a variation of a ligature. I can get discretionary ligatures to appear, but I can't get the variant to show up. What am I doing wrong?

This problem frequently happens when you try to invoke the discretionary ligature and the variation on two separate lines of CSS code. For example (using character variant 49, alpha and beta):

   font-feature-settings:"liga" on, "dlig" on;
   font-feature-settings: 'cv49' 1;

Because of specificity rules behind CSS you need to invoke all the OpenType features on a single line, e.g.,:

   font-feature-settings:"liga" on, "dlig" on, "cv49" on;

Why doesn't Athena Ruby support the full set of ASCII characters?

Font designers for Windows XP and earlier operating systems generally regarded it as a best practice to include in each font at least one complete Windows 8-bit codepage for backwards compatibility. Since Athena Ruby relies on font and layout technologies that are not found on older systems, backwards compatibility seemed irrelevant. Including a full codepage would have added considerably to the cost of the font's development, without serving any of its main functions.


Designers: Tiro Typeworks

Project manager: Joel Kalvesmaki

Consultants: Eric McGeer, Cécile Morrisson, Werner Seibt,

Document Actions