Tuesday, 17 October 2017

Chromatic fonts and Proofreading

A useful application of chromatic (or polychromatic) fonts is proofreading of text, an especially valuable technique for complex scripts such as Egyptian hieroglyphic, CJK writing and mathematical notation.

I've not discovered much on the web about chromatic proofreading possibly because chromatic fonts are relatively new and not as well supported in software as might be expected. For instance macOS contained very limited polychromatic font support until the 10.13 "High Sierra" release last month. Windows Notepad still lacks polychromatic font support (as of the Windows 10 "Fall Creators Update" released today) although Window 10 in general has excellent polychromatic support. Chrome on Android 7.1 appears to be still missing polychromatic support and I've not had the opportunity to try Android 8.0.

The basic principles of chromatic proofreading are obvious. In the case of Egyptian Hieroglyphs, some signs are predominantly used as phonograms, others as ideograms. Some signs could be easily confused with others. Some hieroglyph forms exist but are rarely encountered in historical documents. And so forth. Use of colour to highlight certain of these characteristics can aid identification of patterns and possible transcription errors. I've found use of a chromatic font a more practical tool than using markup for chromatic highlighting for proofreading purposes.

I've used up to 18 colours in experiments with polychromatic proofreading of hieroglyphic. I've also experimented with context-specific colours in fonts and alternate colour palettes.  The key issue is usability and I've found some techniques that work well in specialist software can be over-complex when used in plain text editor software or work processors.

The simplest system I've found useful uses a font with four colours as follows:
  1. Hieroglyphs predominantly encountered as phonograms remain black, as normal text.
  2. Numeric hieroglyphs are displayed as red.
  3. Ideograms commonly used as generic determinatives are displayed as green.
  4. Other ideograms/determinatives are displayed as blue.
Even the novice will be aware that some hieroglyphs fulfil more than one role so its important that knowledge is applied when interpreting this simplified scheme.

An example of the technique applied to a Unicode-encoded text is available at Tale of the Eloquent Peasant encoded in UMdC. This example uses a webfont version of my experimental Aaron UMdC Phonetic Alpha font. An installable version of the font is available for download at Hieroglyphs Everywhere Fonts Project for anyone who'd like to try out the method. I've not got around to document how to use polychromatic fonts yet so less technical readers may want to wait until some tutorial information is available.

Bob Richmond

Monday, 16 October 2017

Hieroglyphic fonts for Unicode

I've compiled a list of Unicode hieroglyphic fonts at https://github.com/HieroglyphsEverywhere/Fonts/blob/master/HieroglyphicFontList.md. 'Official' download links are provided where applicable.

All these fonts contain the 1071 hieroglyphs introduced in Unicode 5.2 (2009). Certain fonts such as Aegyptus and Abydos contain hieroglyphs that are not yet in the Unicode standard.

My new Aaron Series fonts are intended to help establish new techniques for working with hieroglyphic in Unicode. They can also help with development of the next generation of Unicode hieroglyphic fonts that feature quadrat shaping.

If you know of other Unicode hieroglyph fonts to add to the list please let me know and I can update the list.

Bob Richmond

Wednesday, 31 May 2017

Unicode Technical Committee (UTC) documents about Egyptian Hieroglyphic (May 2017)

There are several updates and additions to the UTC document register concerning Ancient Egyptian since my previous post on the topic Unicode Technical Committee (UTC) documents about Egyptian Hieroglyphic (March 2017). Hieroglyphic was discussed during the UTC meeting (May 8-12, 2017) and some recommendations made: see L2/17-103 UTC #151 Minutes.

The place for questions, discussion and suggestions is the Egyptian Hieroglyphs in the UCS mailing list (see Informatique et Égyptologie, Cambridge, 2016).

Recent documents

  • L2/17-076 Revised proposal for the encoding of an Egyptological YOD and Ugaritic characters; Michel Suignard; 2017-05-09. This replaces the earlier L2/17-076 submitted in March as Proposal for the Encoding of an Egyptological YOD. The names of the YOD upper and lower case characters to be LATIN CAPITAL LETTER I WITH SPIRITUS LENIS and LATIN SMALL LETTER I WITH SPIRITUS LENIS. This appear to be generally accepted so we can be cautiously optimistic that a future version of Unicode will resolve this long-standing anomaly.
  • L2/17-122R A method for encoding Egyptian quadrats in Unicode (revised); Andrew Glass, et al; 2017-05-16. Minor revision to clarify some technical points following UTC meeting. See my earlier post A method for encoding Egyptian quadrats in Unicode for an overview.
  • L2/17/153 Recommendations to UTC #151 May 2017 on Script Proposals; Deborah Anderson; 2017-05-07. Summary of the Script ad hoc group discussion on hieroglyphic, recommending UTC discuss the L2/17-122 quadrat proposal.
  • L2/17-171 Future Additions to ISO/IEC 10646 (May 2017); UTC/Deborah Anderson; 2-17-05-17. Additions requested include the Yod (as above) and six Hieroglyphic format control characters characters from L2/17-122 (4 corner and 2 bracket-like controls but not the overlay aka stack control at this time) in addition to the two format controls recommended in January 2016.
Current status of Ancient Egyptian for future Unicode

No decisions appear imminent on hieroglyph repertoire expansion in Unicode but Michel Suignard is making progress on his candidate database.

UTC is now recommending an 8 control character set for representing hieroglyph quadrats in Unicode and this will be running through the standardisation approval pipeline. All those interested in digital hieroglyphic will want to study what is proposed and the degree to which it meets their needs. The controls are:


UTC is recommending two new characters for the Egyptological YOD. Namely


There is a considerable amount of work needed to produce working implementations of the control system and guidelines for usage. There is a case for additional controls. I hope to deal with some of the issues in further blog posts an welcome any suggestions or feedback.

Bob Richmond

Thursday, 27 April 2017

A method for encoding Egyptian quadrats in Unicode

A new document 'A method for encoding Egyptian quadrats in Unicode' is now available from the UTC document register as L2/17-122 [pdf]. The system described takes into account discussions last July during the Informatique et Égyptologie Cambridge meeting and afterwards about extensions to Unicode plain text support to handle vertical hieroglyphic and various complex forms of quadrat structure.

The place for questions, discussion and suggestions is the Egyptian Hieroglyphs in the UCS mailing list (see Informatique et Égyptologie, Cambridge, 2016).

L2/17-122 contains a feasibility report based on three prototype OpenType font developments (Glass, Nederhof, and Richmond) which I hope goes a long way to alleviate concerns raised last year by Egyptologists about the viability of flexible hieroglyphic font implementations in Unicode.

L2/17-122 identifies 9 controls as follows:

Basic quadrat structures


These two controls were proposed in L2/16-018 (January 2016). They are similar to the original Manuel de Codage (MdC85) ':' and '*' controls.


These two controls operate in a similar way to MdC85 brackets '(' and ')'.

L2/17-122 does not contain structure extensions such as the group joiners suggested in L2/16-214 [pdf] to simplify encoding of quadrats in vertical text and tall quadrats in horizontal text. Therefore for most applications the basic quadrat structures of L2/17-122 are encoded as exact equivalents to those of MdC85 (itself derived from the Buurman 1976 model).

However, there are subtle differences from MdC85, most importantly (i) L2/17-122 has more clearly defined control behaviour and (ii) quadrat appearance is determined by a font (or equivalent) so there is more flexibility in handling issues such as hieroglyph sizing, kerning, etc. in plain text implementations.

Hieroglyph combinations


This control overlays one hieroglyph on top of another - a direct equivalent of the MdC88 '#' control (encoded as '##' in JSesh).


These four geometrical ligature controls are proposed in place of the L2/16-018 EGYPTIAN HIEROGLYPH LIGATURE JOINER (which was based on an abstract ligature model for non-grid quadrat elements). This set of four ligature controls originates from a consensus formed at the I&E 2016 meeting that four 'corner control' ligatures are sufficient to meet anticipated plain text ligature needs of corpus projects such as Ramses and TLA and that the Egyptologists present preferred geometrical to abstract ligatures. This is a new approach to ligatures although they link fairly well to usage of the original MdC ''&''ligature and MdC extensions familiar to JSesh users.

Bob Richmond

Digital Encoding of Egyptian Hieroglyphic: Origins

Updated 2017-05-17.

I thought it might be useful to summarise some of the background to digital hieroglyphic to help  inform discussion about representations of hieroglyphic writing in Unicode.

This post deals with the early years. Information is thin on the ground so I'd be delighted to learn about any material, unpublished or unpublished, that survives from this formative period.

Apparently, use of computers for hieroglyphic goes back to the 1960s when computers and printing peripherals were hugely expensive and inaccessible to most people except a lucky few. However it was not until the early 1980s that the emergence of personal computer technology started to bring digital techniques and practical tools to Egyptologists and others.

The first Informatique et Égyptology 'round table' meeting (Paris, 26-28 June, 1984) was pivotal in shaping the first generation of digital hieroglyphic that has been used for the last 30 years. Fortunately, the proceedings of the meetings were published in 1985 (although unfortunately and ironically not yet available in digital format) and this post is mostly based on that publication. I'll summarise some papers from I&E 1984 relevant to encoding.

COMPUTER PRINTING OF HIEROGLYPHS AT THE UNIVERSITY PRESS OXFORD (T. G. H. James) describes the replacement of the traditional metal type system at Oxford University Press (used for typesetting the Gardiner font from 1927 to 1983) by a Monotype LaserComp photo type setter adapted for hieroglyphic. The final publication to use the original hot metal font was J.E.A. Vol. 69 (1983).
Typesetting instructions for OUP workflow
This paper gives an insight into older typesetting practices as well as the short-lived LaserComp technology soon to be superceded by desktop publishing on personal computers.

INFORMATIQUE APPLIQUEE A L'EGYPTOLOGIE (Dirk van der Plas) gives a snapshot of his experiences with the Buurman GLYPH program and the practical situation for those printing hieroglyph texts on a budget in 1984 including comparative costs of authography and typesetting.

A PROGRAMM SYSTEM FOR THE EDITION OF TEXT (especially hieroglyphic printing) (Norbert Stief) describes hieroglyph plotting software written in Fortran 77 running on an IBM 370 mainframe driving a CALCOMP plotter. The system was developed at University of Bonn and appears to be what is sometimes later known as the PLOTTEXT system. Mnemonics are used as alternatives to alphanumeric Gardiner codes for encoding purposes in a similar way to Buurman (1976). The 'Bonn Ziechenliste' font catalogue (an extension of the Egyptian Grammar sign list) is given here.

NEW HARDWARE--NEW SOFTWARE (Leonard H. Lesko) gives a short summary of hardware used in his latest setup at Brown University and his earlier hieroglyphic printing workflow at Berkeley during 1973-1982 using the SCRIBE program. The Berkeley system was notable for its use in creating A Dictionary of Late Egyptian 1982-), what I understand to be the first substantial hieroglyphic dictionary to use digital encoding and printing techniques. Not mentioned here is the use of quadrat structure patterns for SCRIBE (rather than a control scheme such as that used by Buurman). Just to prove there's nothing new under the sun, one of the systems I initially considered for Unicode encoding in early 2015 used a similar pattern system although at the time I was unaware of the Lesko work from 40 years ago.

PRINTING OF EGYPTIAN HIEROGLYPHS BY MEANS OF A COMPUTER (Jan Buurman; Astronomer and hobby-Egyptologist). Buurman gives a history of his system which began as a hobby project in 1969. A sketch of the system was first published as The Composing of Hieroglyphic Texts by means of a Computer in Göttinger Miszellen 19 (1976). As far as I'm aware GM19 contains the earliest publication of what would become the basic MdC controls for quadrat structures.
Quadrat structure notation from " The Composing of Hieroglyphic Texts by means of a Computer" (1976)
GM19 reports that the first texts output (in 1971) took an average of 0.2 seconds per hieroglyph to process and 1.3 seconds to plot. The first version of GLYPH was written in Algol 60, running on a CDC Cyber 73 mainframe.

I understand that Buurman showed a video of the plotter in action to the meeting. Far more fun I expect than the laser printers we have grown used to. I'd love to see a video of  hieroglyphs being drawn by a plotter! I am grateful to Hans van den Berg for kindly providing two links showing output from a 1980s HP ColorPro plotter:  'wild bull hunt scarab' of Amenhotep III (https://www.youtube.com/watch?v=8z9aCclxV0U) and part of the Tale of Sinhue (https://www.youtube.com/watch?v=zyp68emXMZM).

RESOLUTION. The meeting resolved to create a standard system for encoding hieroglyphs to be made available before the Fourth International Congress of Egyptology in August 1985. A committee chosen to work on the manual consisted of Jan Buurman, Nicolas Grimal, Michael Hainsworth, Norbert Stief, Robert Vergnieux and Dirk van der Plas.
Proceedings of Informatique et Egyptology 1984 (Paris, 1995). Page 225.
This proposed system was to become known as Manuel de Codage (MdC85) which I aim to summarise in a subsequent post.

Bob Richmond

Monday, 27 March 2017

Unicode Technical Committee (UTC) documents about Egyptian Hieroglyphic (March 2017)

Updated 2017-04-04.

My January post Unicode Technical Committee (UTC) documents on Egyptian Hieroglyphic 2016 summarised relevant submissions to the UTC Document register last year. Additions so far for 2017 are:

January to March 2017
  • L2/16-210r  A system of control characters for Ancient Egyptian hieroglyphic text (updated version); Mark-Jan Nederhof and others; 2017-01-21. Warning of possible confusion. Despite the L2/16-210r document code this January update is substantially changed from the L2/16-210 document and not a minor revision (the earlier document contains alternative ideas and additional material). The original L2/16-210 dated 2016-07-25 is still available as http://www.unicode.org/L2/L2016/16210-egyptian-control.pdf ).
  • L2/17-073 New draft for the encoding of an extended Egyptian Hieroglyphs repertoire (Hieroglyphica based); Michel Suignard; 2017-03-19. An associated spreadsheet http://www.unicode.org/L2/L2017/17073-n4788-hieroglyphs.xlsx is also available for download. A database snapshot is also available at http://www.unicode.org/L2/L2017/17073-n4788-database-add.pdfThis draft contains substantial changes from its predecessor.
  • L2/17-076. Proposal for the encoding of an Egyptological YOD. Michel Suignard; 2017-03-24. A new attempt to sort out the remaining transliteration character.

If I've missed any other published documents relating to hieroglyphic in Unicode or any points came up at the UTC 150 meeting in January please let me know.

The next UTC meetings are in early May.

Bob Richmond

Monday, 27 February 2017

Irregular Hieroglyph clusters

I am compiling a document on irregular Egyptian Hieroglyph clusters (aka quadrats) and would like to hear of any publications relating to this topic. Documenting what is required is an essential precursor to expanding the capability of Unicode plain text orthography beyond regular cluster support.

Examples in this blog post are all taken from the 18th Dynasty Tomb of Rekhmire (TT100).

Here are some examples of the regular clusters that account for the vast majority of writing from ancient sources:

In addition to these simple grid-like arrangements, there are various conventional forms of cluster such as:
The majority of these conventional clusters follow the bird and cobra patterns of non-grid arrangement.

Irregular clusters can be seen when the scribe/artist attempts to squeeze more hieroglyphs into the available space, avoid empty space, or achieve some special visual effect.
In many cases in Unicode text it is satisfactory to transcribe irregular clusters into an equivalent more regular form as has often been done in established practice. For instance the first irregular form in the illustration could be written in traditional MdC as (a:t)*H-A28 with no great significance for transcription purposes. Likewise the last example can use t:t:W10 ignoring the slope.

The goal of future developments of the Unicode writing system is to enable rendering of clusters inherent to the writing system. This principle can only be made precise on the basis of well-attested data about irregular clusters in order to be crystal clear what is to be accomplished, Not all arrangements of hieroglyphs discovered in ancient writings are inherent and count as plain text. Through detailed illustration it should be possible to reach a consensus among Egyptologists that an extended Unicode system meets the needs of the subject.

In the near future I hope to produce draft documentation for comment. Meanwhile I'm interested in hearing of any relevant work that has been done already on hieroglyphic orthography and any examples of irregular quadrats that may be relevant.

Bob Richmond