STIX Project Home Page

Updated: 20 October 2006

The pages accessible from here were originally developed for the use of the Team members and other contributors to the STIX Project. (This group was initially led by Nico Poppelier of Elsevier Science Inc.; Nico has left Elsevier, and Wim de Vries is now the Elsevier representative to the STIX project.)

The various documents linked below were developed mainly in support of an attempt to gain acceptance into Unicode of additional symbols used in mathematical and technical publishing. These documents include all the proposals (and their revisions) made to the Unicode Technical Committee (UTC) since work on the project began in 1997 as well as a great deal of background information compiled in support of these proposals. Thus, this page forms a historical outline of the Unicode facet of the STIX project.

Newest files

Oops! Another omission caught and corrected. Here are the new table file and the applicable layout description.

Some minor errors and omissions have been found and corrected in the table file. The layout description remains the same. If anyone notices any remaining errors, please send the details to Barbara Beeton.

The final glyph delivery has been made, and the fonts are undergoing packaging and a final design review. In the meantime, Unicode 5.0 has been finalized, including some additional math characters. The STIX master table has been updated with the new codes from Unicode 5.0 (and some new symbols as well).

TeX control sequences have been assigned to all symbols that are likely to occur in math or other technical contexts (e.g., phonetics); Latin alphabetic characters that are intended for use in text, regardless of language, have not been assigned TeX control sequences since there are other, more preferred, methods of accessing them.

All glyphs have unique Type 1 names. Many of these correspond to simple Unicodes. Some are formed from compound Unicodes (e.g., "uni227620D2") or from qualified Unicodes (e.g., "uni019B.var"). Finally, some glyphs that do not have Unicode representations have been assigned names with the prefix "stix" and a 4-hex-digit code in the range EExx of the Unicode Private Use Area (e.g., "stixEE24").

The updated table file and the corresponding layout description are posted. The column layout remains the same; values of a few flags have changed, and there are several new sources listed for TeX control sequence names. Scott Pakin's Comprehensive LaTeX Symbol List, available from CTAN, has been of enormous help in avoiding name clashes.

Creation of the glyphs for the STIX fonts is almost complete, although a full design review is required. Delivery of most of the remaining glyphs has turned up some items which were not in the master table, and these have been added. Some new characters accepted by Unicode have been added as well. For symbols which have not been accepted by Unicode, "final" STIX IDs have been assigned in the Unicode Private Use Area (PUA), and Type 1 glyph names derived accordingly. The updated table file and the corresponding layout description are posted. The column layout remains the same; only values of a few flags have changed.

The next task is to assign TeX control sequence names to the symbols that do not already have them. In order to make use of names already in general use and to avoid name clashes, we will refer to Scott Pakin's Comprehensive LaTeX Symbol List, which is available from CTAN.

An inventory of the glyphs delivered for the STIX fonts has been completed, and a request has been made for what remains. The STIXFonts web site has been updated with the latest information. Unicode 4.1 charts are posted with quite a few additions, and the pipeline still shows a few math symbols waiting for final approval. The table file has been updated and rechecked; quite a few minor glitches were found and corrected. The layout description has been updated, mostly to reflect changes in type-1 glyph names and a new flag value; the column layout remains the same.

Work on the STIX fonts continues, but is nearing completion. For more information, see the STIPub web site. Unicode 4.1 is under construction, and some additional symbols (including an undotted j!) have been accepted; see the Unicode "pipeline" for details. The new symbols have been added to the table file, along with some more error corrections. Some new flag values have been added to the layout description, to indicate the beta status of some new table entries; the column layout remains the same as the last version.

Concurrently with work on the STIX fonts, the master table file has been updated to correct errors and add occasional items uncovered subsequent to the latest Unicode release. The layout remains the same as the version of 2003/05/02.

Unicode 4.0 was published. This version includes a number of mathematical symbols discovered after the release of version 3.2.

A project undertaken in the W3C in coordination with ISO will update TR 9573-13 to include Unicode references along with entity names. A draft version of the new report is now available.

Unicode Technical Report #25, Unicode Support for Mathematics, has been released in final form, as approved by the Unicode Technical Committee.

In preparation for the release of the STIX font, Type 1 glyph names are being added to the master table file. This requires a change in the layout.

Work continues on the STIX font. For details, see the STIPub web site.

At the 22nd International Unicode Conference, Murray Sargent presented the paper Unicode Support for Mathematics, introducing the mathematics content of Unicode 3.2 and how it can be used.

The STIPub web site is now live:

A news story on the Bulletin site of the Seybold Reports announces the public debut of the project and summarizes its goals.

Another gap. In the interim, some exciting progress has been made:

Work is now underway on creation of the fonts. That effort will be the subject of a separate web site, under the aegis of STIPub. A link will be provided with the site becomes "live".

Once again, this file hasn't been kept current, owing to the press of other business; when time permits, the gap will be filled in.

A new version of the master table file incorporates corrected values for some Unicodes as well as about 70 new symbols identified since the initial submission in 1998. The layout hasn't changed; see the link at 2000/10/19.

Mathematical Markup Language (MathML) version 2.0 was released as an official W3C Recommendation.

The master table file has been updated to incorporate this new and changed information: A brief explication of the layout of this table is available; this file identifies the location and content of each field in the table.

Draft Unicode charts including the math symbols were available for preview from the Unicode site; these preliminary charts have been superseded by the Unicode 3.2 code charts.

This file hasn't been kept current, owing to the press of other business; when time permits, the gap will be filled in.

Briefly, activity has proceeded on two fronts.
Good news!
The UTC at their meeting last week accepted the symbols proposal; a few minor changes were suggested, and these have been posted in a revised version which will be frozen for the record. Code points were provisionally assigned to all the distinct symbols; here is a list of the new Unicodes, keyed to the temporary IDs of the proposal. Some additions have been made to this complement, most notably a ``variation selector'' (which is the formal designation for the MVT used in the proposal) and a collection of brace pieces representing a grandfathered set used with a prominent word processor.

The next step is consideration of the proposal by the ISO WG2 Working Group. This will occur at their next meeting, March 21-23, in Beijing.

An updated version of the symbols proposal is available in draft (PDF) form. Many symbols have been removed as individual entities, and replaced by composites or alternate forms indicated by a Math Variant Tag (MVT). The symbol charts in the PDF file have not yet been changed. HTML versions of the updated charts can be viewed as follows:

A new working draft, Mathematical Markup Language (MathML) version 2.0 was released. Chapter 6, Entities, Characters and Fonts, is still waiting for Unicode action on the STIX proposal, and remains preliminary.

The proposals for mathematical alphanumerics and for mathematical and technical symbols were discussed at the October 26-29 meeting of the UTC. Some additional changes have been requested in the symbols proposal before the next meeting (January 31-February 3); an updated version will be posted when ready.

Another proposal for math/technical symbols, Mathematical brace pieces, was also considered and accepted. This covers a number of pieces used to create various delimiters as composites, and is intended to support existing character sets from Adobe, Hewlett-Packard, and some other sources; it is not intended to cover the existing TeX glyphs from the cmex font -- for input, storage, searching, etc., the use of a single code remains the method of choice, leaving the actual composition of the glyphs to the typesetting software and fonts.

The proposal Request for assignment of codes to mathematical and technical symbols (L2/99-244R) has been revised; this version will be considered at the October 26-29 meeting of the UTC.

Revision 1 of the Proposal to encode mathematical alphanumeric characters (L2/99-195) was presented to the WG2 meeting last week, and was accepted unanimously:

``WG2 accepts the proposal for 991 (subject to verification of this number) new mathematical alphanumeric symbols in document N2086, for inclusion in 10646-2, in the range D400--D7FF in Plane 1. The project editor is to select the appropriate character names and block name in the preparation of the text for inclusion in part 2.''

The 2nd working draft of 10646-2 was accepted and the project editor instructed to prepare the text for the CD (committee draft), including the mathematical alphanumerics, and submit this document for ballot. After the CD is accepted, two more ballots are required: DIS (Draft International Standard), and IS (International Standard). This process, if expedited, can be completed in about a year and a half.

The proposal entitled Request for assignment of codes to mathematical and technical symbols (L2/99-244) has been updated in response to a review by several UTC members, and will be carried into the semiannual Unicode Conference for additional feedback.

An updated version of the table file has also been posted. It contains changes reflecting updates to the symbols proposal as well as some new material from IEEE; the format has not changed from the last version posted.

New versions of two proposals are now in advanced draft stage. After a few final corrections, these may be submitted to the WG2 meeting in September and will be on the agenda of the October UTC meeting.
The table file (prepared by Barbara) has once again been amended, to add references to the new version of the symbols proposal. This is a very large ASCII flat file, with material position-aligned into columns. This is the current table structure.
start  end name description
1 1 flag marks questioned items of various types
2 5 uniq Unicode point in hex
7 10 xref cross-reference code
12 15 AFII AFII identifier in hex
17 20 rq-1 ID of requested symbol in June 1999 proposal to UTC
22 25 rq-2 ID of requested symbol in August 1999 proposal to UTC
27 27 C character class code
29 38 entity entity name suggested
40 48 set ISO entity set
49 52 els Elsevier handle
53 53   source of TeX name
54 79  AMS/TeX name  TeX or AMSTeX name
80 96 APS American Physical Society name
97 106 AIP American Institute of Physics name
107 116 ACS American Chemical Society name
117 136 IEEE Institute of Electrical and Electronic Engineers name
137 167 Wolfram Mathematica name
168 180 Springer Springer name
181 185 KAP Kluwer Academic Publishers handle
186 192 SciDes Design Science MTCode
193 197 M-Pi Adobe MathematicalPi font reference
198 end description descriptive text

A revised version of the symbols proposal has been posted (see above). The revised charts, showing (low resolution versions of) all the symbols can be viewed separately. The collection has been divided into four parts:

At the June UTC meeting, three math proposals were considered:

The post-UTC-meeting version of the table file prepared by Barbara. The table is a very large ASCII flat file. The structure is that of a very wide table. This version incorporates the reference IDs from the proposal for coding of math symbols as well as new material from Design Sciences and the Adobe MathematicalPi fonts.

A revised proposal was prepared for presentation to the UTC, excluding nearly all alphabetics (which had become the subject of a related proposal prepared by Murray Sargent, a UTC member from Microsoft), and consolidating most of the information on symbols from the December 1998 proposal and the March 1999 addendum.

A second proposal concerning math variants was also submitted, introducing tags to extend the symbol complement in regular ways, e.g. by converting a binary to an n-ary operator, adding negation, etc. This material was kept separate since it was expected to raise discussion not relevant to whether the symbols in the first document would be acceptable to UTC.

Previous versions

The last version of the table file prepared by Barbara prior to the June 1999 UTC meeting.

At the request of the UTC, an addendum was prepared, with some changes in the proposal text and the symbols list, omitting items whose status was resolved by recent additions and changes to Unicode.

The submission to the Unicode Technical Committee was reorganized into a single document with seven tables of symbols, grouped by type.

The ISO 10646 WG2 submission consists of some 70 or so files plus PDF equivalents.

The index to the private zone codes shows the sample glyphs with the temporary numbers presently assigned. These are in no way to be considered suggested Unicode numbers. But they really help clarity of discussion.

An update of the table file prepared by Barbara.

The ISO 10646 WG2 submission is done, and consists of some 70 or so files plus PDF equivalents. The index to the private zone codes is private to STIX, though the rest can be accessed as above.

A rough collection of files with glyphs showing, based on an earlier version of BNB's table, is available. Note there are 20 (hex) files. Those from 10 on contain the characters not already in Unicode 2.0.

Mathematical Markup Language (MathML) 1.0 became a Proposed Recommendation of the W3C today [Press Release]. The files containing tables in its Chapter 6 show many glyphs though not the complete set required from STIX. If things are wrong there it would be good to know. Note the Unicode numbers in the private zone used in the MathML document are not necessarily the same as those in use in the STIX tables. It is hoped that the glyphs are the same though.

Barbara's newest table file is now online, as she goes off to a well-deserved Caribbean 'bare-foot' vacation.

Though the new Unicode home page no longer leads to the Unicode glyph image charts, John Jenkins of Apple, who oversees this site, has told us that there is a Unicode chart server. (A later edition of the Unicode home page does contain a link to the charts.)

01 Sep 1997
There are new files merging the character information from all of the sources in a sort of Universal form. As of today not all the links have been verified, so it is best to look at the files from an index, and that's where the previous link gets you.

04 Aug 1997
Chris Hamlin has added a new form of the AIP list including Unicode equivalences, to the others available, such as Arthur Smith's APS list including a comparison with ISO characters, and those to be had below from Elsevier and the AMS.

31 July 1997

There are now compilations of characters in table form with images from Unicode and Elsevier based on the ISO tables (ISOAMS[A-C], ISOPUB, etc.).

A link pointed out by Arthur Smith to the recent (21 July 1997) Working Draft on Fonts of the W3C CSS committee is added below in the reference materials.

28 July 1997
It seems that it may be more useful to have full listings with images of all the characters in the chosen signatures, not just those in the Elsevier and other lists. This allows one to see at a glance which have been left out.

Correction to the links to images in some files posted last week have been made.

The LaTeX 3 Project's early report on math symbols authored by Justin Ziegler, as pointed out by Barbara Beeton, deserves linking here too. It lists the symbols required for math/technical composition, arranged into 256-element fonts for use with LaTeX. There is a local copy for convenience, perusable as TeX source if you wish. The original can be found on any CTAN node in the area info/ltx3pub/l3d007.tex . This reference is relative to a TeX archive root node as follows:
19 July 1997
Building on some of the recent material there is now a pair of sets of comparative listings of characters covering those items of the Elsevier collection already in Unicode. The comparison is with AFII numbers, AMS names and ISO 9573-13 names. There are forms with and without the added convenience for viewing, but disadvantage for loading, of images from both Elsevier and Unicode. The whole has, of course, still the status of a working draft.

The collection is divided into signatures, which are taken to be subsets of characters whose Unicodes are of the form NN** in hex; each signature is divided into tables of up to 16 characters of the form NNM*, which may be thought of as sheets of characters. It is hoped that this way the loading of tables will be quicker. Since the symbols of interest to STIX are taken from a restricted part of the Unicode set the signatures that have been covered in this list are those with NN drawn from the list


Initially only the characters in the Elsevier set are shown in this listing although it is keyed by Unicode numbers. Those not already in Unicode are intended to be prepared for consistency with Unicode numbers from the private zone signatures E5, E6, and E7.

18 July 1997
Although much work has been done in the last week there has been nothing new posted on the STIX site. This is now changing with a bang. There are new tables from Barbara Beeton: and Arthur Smith has made available to the list his ongoing work on the fonts for APS work provided by Beacon. This includes new image files.

Nico Poppelier's new files weren't simply made available by posting, but now can be seen here too:
24 June 1997
Tables from the AMS containing work of Barbara Beeton on checking AFII codes and TR 9573 entity names.

19 June 1997
These working tables have been supplied by Nico Poppelier and the folks at Elsevier:

Reference materials

For reference we have various relevant files:

Last update: 20 October 2006

The nominal administrators of these pages are Please direct any comments to the appropriate person.