Frequently Asked Questions regarding the STIX fonts
Content of the STIX fonts
Q. What criteria were used to select the characters in the STIX fonts?
A. The fonts are intended for use in typesetting technical material (mathematics, physics, etc.). Since the members of the consortium creating the fonts are based in Europe and North America, the fonts are suitable for use principally in English and Western European languages. They contain the full Latin-1 alphabet in the four principal styles: roman, italic, upright bold, and bold italic.
The symbol component of the fonts includes material gathered from the following sources:
- ISO Technical Report 9573:13, Public entity sets for mathematics and science;
- symbols required for the publications of the six STI Pub organizations;
- other scientific publishers and knowledgeable individuals, submitted in response to a call for information;
- various other published documents.
The collection of symbols is comprehensive for these sources.
Q. Most glyphs have regular Unicode numbers, but quite a few have numbers in the Unicode Private Use Area (PUA). What is the reason for this?
A. Symbols with codes in the PUA are primarily of several types:
- Additional sizes required for publishing, e.g. large operators, delimiters, or diacritics. Only one Unicode is assigned for a symbol with a particular meaning; use of these different sizes in context is the responsibility of the application.
- Symbols that have the same meaning as designated by a Unicode value but a variant shape; many of these are equivalent to the base Unicode followed by the qualifier U+FE00 (Variant selector 1).
- Symbols that can be composed of a base Unicode plus a combining diacritic, e.g. symbols negated by either a slash (U+0338) or a vertical stroke (U+20D2).
- Partial symbols to be combined to form complete shapes, e.g. extenders for arrows.
- The mathematical alphanumerics nominally in Unicode Plane 1, to allow access for applications that do not support Plane 1. This decision may be reversed for the production release.
- Symbols considered by the Unicode Technical Committee (UTC) to not be math, but relevant to another discipline, e.g. astronomy or chemistry. The UTC has requested that these be submitted for consideration by an appropriate organization in that discipline.
We recognize that use of PUA codes is dangerous, but other mechanisms for accessing glyphs, including by Unicode composites, are not at present recognized by many applications. If another, more appropriate, mechanism for glyph identification is adopted by software providers, it would be desirable to upgrade the glyph identifiers to use such a mechanism.
Q. Some glyphs appear in spaces that are marked "reserved" in the Unicode charts. Why is this?
A. Unicode is a work in progress; new items are continually submitted for consideration, and additional documentation is found for pending items. The STIX Project has coordinated its Unicode character assignments with the latest version of Unicode, now version 5.1, which is to be published in March 2008. Some valid assignments do not appear in earlier versions of the Unicode charts. Please refer to the Unicode 5.1 charts when they are posted, and inform us of any discrepancies that you find so that corrections can be made.
Q. Some characters in various Unicode blocks are missing from the fonts, for example, arrows (U+21EB - U+21F3). Why do these gaps exist?
A. Using the arrows as an example, the ones cited are designated in Unicode as "keyboard symbols". These did not appear in the STIX lists compiled as described in the first question. Except for a few core Unicode blocks, it was decided that extra effort would not be made to include glyphs for all possible characters. Glyphs that did not appear in the STIX lists were not included in the collection.
Q. The fonts miss some essential glyphs to draw mathematical equations, for example those for drawing large brackets and parentheses like U+239B - U+23AD. These are part of the MS Symbol font. The STIX font should make that font obsolete, and should contain at least all glyphs that are included there.
A. The cited glyphs are ends and center pieces from which very large delimiters can be built up. These pieces, as well as complete large glyphs, are present not in the four main fonts (STIX General, STIX General Italic, STIX General Bold, and STIX General Bold Italic), but in one (or more) of the fonts of oversize glyphs. The reason for this is that many software applications assume that the normal baseline for a font is based on the height plus depth of the largest glyph. If oversize glyphs are present in text-size fonts, lines of text will be set much too far apart; proper use of the "oversize" glyphs is the responsibility of the application. To answer the original question about U+239B - U+23AD, these glyphs are found in STIXSiz1Sym.
Other gaps have been reported in the comments on the beta release. Details, including their Unicode status, reasons for possible omission, and location in the STIX fonts (if present) are given in a separate list available to be provided later.
Q. I need some symbols that don't appear in the STIX fonts. How can they be added?
A. The UTC has accepted the principle that mathematics is a language, albeit a specialized one, and that its vocabulary is open-ended. New symbols will continue to be devised, and with proper documentation of their acceptance by the technical publishing community (publication in a book or journal by a recognized publisher), they can be submitted for consideration as additions to Unicode.
Requests have been sent out to organizations specializing in areas other than math (e.g. the American Chemical Society and the American Geophysical Union) to designate representatives to carry forward to the UTC requests for symbols now in the PUA. If other symbols in such areas are identified, they can be included in such requests. The primary contact within the UTC who dealt with the math additions will remain available to help process new requests.
The licensing terms for the STIX fonts will permit additions; however, no formal mechanism has yet been defined for resolving additions made by different users. This matter will be considered by STI Pub, the parent consortium responsible for the STIX project, and reported in this FAQ when a mechanism has been established.
Q. I'm a bit surprised not to find some common combined characters. For example, the letters with a single dot or a bar underneath are missing. Why is this?
A. If these letters are meant to be used as symbols, the MathML model requires that diacritics be added to letters by markup. The letters you cite are not typically needed in text by the organizations under whose direction the STIX fonts were developed. While we understand that mathematical research is carried out by people working in the context of Arabic or other non-Latin language, the time and expense involved in creating an exhaustive complement of alphabets was beyond the capacity of these organizations.
This might be an area for contributors to provide enhancements after the production version is released.
Q. There are some invisible Unicode characters that are widely used in applications like MathML, e.g. U+2061 (function application). These characters do not seem to be in the STIXGeneral font. Are there ways to handle invisible characters in the font, such as included in the font with no visible glyph?
A. The invisible characters specified in Unicode for use in mathematical applications should be present in the font with appropriate values in the various tables; this will be corrected in the production release.
Q. Some symbols have shapes that are different from what I was expecting. How will errors be corrected?
A. Possible reasons for shape differences include:
- Errors of various kinds that will be corrected before the production release:
- Placement of a glyph in an incorrect location.
- Different shapes appear in the same location in the lightface and boldface fonts.
- Misinterpretation of what shape should be used, for example a single-line "equal" component as opposed to a double line (cf. U+2264 vs. U+2266). If supporting documentation is provided, these shapes will be changed, and, if appropriate, a new entry provided with the alternate form.
- Multiple possible shapes may represent a single meaning. In several instances, known variants are provided, although it may not be trivial to find the additional shapes in the fonts. This problem will be handled by documentation, which is not yet complete.
The shapes used for the glyphs were researched carefully, and the entire collection was subjected to a design review before the beta release. If presumed errors can be backed up by documentation, corrections will be made. Documentation should take the form of an example of the questioned glyph, in context, in a published source; a PDF file of the scanned page, with full bibliographic information, is best.
Q. Some letters, in particular some of the Greek lowercase letters (e.g. nu), are shaped differently from what one expects in text. Why is this, and can it be corrected?
A. Many of the lettershapes were designed to do double duty -- in text, and for use as symbols. When used as a symbol, a letter must be able to be immediately and unambiguously recognized in isolation. The nu has always been a problem in this regard, as it is easily confused with an italic "vee". The lettershapes in the STIX fonts have been adapted for optimum recognition in a symbol context, sometimes to the detriment of their appearance in text. There are no plans at present to make any changes on this account.
Q. How were the character/glyph names determined?
A. A strong effort was made to conform to existing Unicode principles and conventions, and, for the symbols adopted into Unicode, the reference names assigned for Unicode characters went through the standard (rigorous) Unicode editing procedures. For example, the Unicode naming convention is "TWO-HEADED" for arrows; if "TWO HEADED" appears in a STIX glyph description, it will be fixed. If you notice such an anomaly, please bring it to our attention.
Glyph names were assigned according to the Adobe guidelines for "Unicode and Glyph Names", using only the core set of names, and referring to other glyphs by their Unicode designation. For glyphs with no Unicode, unique names were defined beginning with "uni" followed by a 4-hex-digit code in the PUA range starting at E000.
Q. Is there a good reason why the font names don't contain spaces? Most modern software recognizes names such as "Times New Roman". There is no reason to adopt difficult to decipher names such as "STIXGeneral".
A. The final names will incorporate spaces for easier human recognition.
Q. Some family names in the OpenType name table are not the same as in the CFF (Compact Font Format) fonts. For example, all the integrals have the family name "STIXIntegrals" in the table, but the fonts themselves have names like "STIXIntegralsUpDisplay", making it difficult to select the desired variant. Can these names be normalized to allow particular variants to be selected.
A. These naming problems will be corrected in the production release.
Q. Once the fonts are released for production, they may undergo updates from time to time. A typical method of keeping track of different releases is to include a version number as part of the distribution file name, such as stixfonts-0.1.zip, stixfonts-0.2.zip, and so on. Will this be done?
A. Version numbers will be included in the names of production releases.
Organization and technical characteristics of the STIX fonts
Q. Many fonts with intentional omissions leave ".notdef" as a place marker in such locations. Without this aid, it is difficult to tell whether an omission is intentional or not. Can this feature be added to the STIX fonts?
A. The intentional "holes" in the STIX fonts will be indicated by ".notdef" in the production release.
Q. Why are the Unicode letter-like symbols and Plane 1 alphanumerics distributed across different fonts rather than being all in one place?
A. It was felt, perhaps unwisely, that an ordinary user would look for an italic letter in an italic font, a bold letter in a bold font, etc. We have rethought this decision, and in the production release will place the full Unicode complement of these two blocks in a single font.
Q. It seems that the combining diacritics do not have GPOS OpenType data, so the system cannot correctly render "base character + combining diacritics". Is there any plan to add this kind of information to make the fonts "smart"?
A. GPOS data will be added for the production release.
Q. For some glyphs in the PUA area, there are multiple code points for different style variations of one glyph. It is more usual to assume that the same glyph, regardless of style, has the same cmap value. Are there plans to merge these to a single code point?
A. Code points will be normalized for the production release.
Q. Are there any kerning pairs in the STIX fonts? None are detected by the Fontforge program.
A. No kerning pairs were present in the beta release. The alphabets intended for text use were reviewed carefully during the creation of the fonts to minimize the need for them. A basic set of kerning pairs will be added to the fonts before the production release.
Q. The terms of the license do not appear to allow modification of parts of the fonts that are not the glyphs, including tables or the README.txt. Will the final license be fully "open"?
A. The STIPub Group plans to release the production version of the fonts under the SIL Open Font License.
We do request that, if any changes are made, these be documented fully, the font names changed, and the identity of the person or organization responsible for the modification be included in the documentation. The copyright holder should be informed of any modifications, and may choose to incorporate contributed modifications into the standard version.
Installation and application support
Q. I looked through your website but didn't find a test page which will show off the STIX fonts and confirm that they are correctly installed. Will such a page be provided?
A. We will create such a page for the production release.
Q. Why are all the size variations needed in a scalable fonts world? Modern *nix systems use <emph>fontconfig</emph} to determine when one font should be substituted for another; your complex font layout obviously demands complex substitution rules.
A. The beta distribution was packaged in a way that we hoped would allow it to be used with the maximum range of software applications. Testing has shown that other packaging is possible. Appropriate changes will be made in the production release.
As for multiple sizes of some glyphs, at least in the case of large operators (sum, product) good quality typography demands that the larger size used in display material (as opposed to text) be a slightly different shape, not simply a magnified version of the text-size form. Similarly, for delimiters (parenthesis, brackets, braces), simply magnifying a small form would result in larger shapes that are too heavy for the enclosed material.
Q. I am trying to get STIX fonts to work with some X Window System applications that use the core X11 fonts system, but have not yet succeeded. How must I install the fonts to use them in this environment?
A. [answer needed; see round 5-1, message from Oussama Rabi]
Q. I used the STIX fonts on two different operating systems (XP Professional 32-bit and Vista Ultimate 64-bit) with Word 2003 and Word 2007 respectively to create a document to convert to PDF. On neither setup would Acrobat 3D, version 8.1.1, convert the Word file to PDF. How can this be made to work?
A. [answer needed]
Q. Will (La)TeX support be made available for the STIX fonts?
A. Several of the STIX organizations use TeX-based systems to produce their publications, so this is a necessary component of the STIX project. However, until the packaging of the fonts is reasonably stable, developing LaTeX support is difficult. A knowledgeable TeX specialist has agreed to develop the LaTeX support; completion of this phase of the project is expected by late Spring of 2008.
Reporting errors or submitting new material
Q. I believe I have found an error. How can I report it?
A. If the report can be conveyed entirely by e-mail, direct it to the Stix Fonts project via the template provided: the STIX Fonts project.
If the report is accompanied by documentation, please send it by fax to .
Q. When should documentation be provided, and in what form?
A. If a shape error is detected, identify both the Unicode and the font information (font name and location in the font). Make a photocopy or scan of a page from a published document that shows the questioned glyph, in context. Provide full bibliographic information for the source from which the copy is taken.
Similar documentation should be submitted for a newly identified symbol. In that case, also provide a suggested name, identify the field in which it is used, and say whether it is related to any existing Unicode character.
Note that bibliographic information alone is not sufficient; the person preparing a submission to the UTC or instructions for modifying a shape in the font may not have access to the cited publication.
[last updated 11 March 2008]