Introduction

Purpose

This page provides a set of test cases for a Tai Tham renderer. It has been compiled with a view to putting the 'Universal Shaping Engine' through its paces for the Tai Tham script. This set of tests is incomplete in that it does not directly give the correct renderings, although some one in possession of the source documents could visually check them.

I was originally requested to provide words of one syllable for such a test. By syllable, I understand an Indic syllable of the form C+(M*V*M*C*M*)* with a single base consonant. (M = miscellaneous mark). I include cases where the rôle of the base consonant is played by something other than a letter. The post-vocalic consonants occur not only in other SEA Indic scripts such as the Khmer script and Lao script (as in the use of ຽ U+0EBD LAO SEMIVOWEL SIGN NYO in the Lao writing system), but also in Tibetan.

However, many of the interesting cases occur in the second syllable of a word, and certain initial syllables are obligatorily followed by more characters of the word. I have therefore also supplied longer words when a conceivable problem would not appear in a word of one syllable.

The dependent vowel AA (U+1A63 and U+1A64) may form the base of its own little stack of dependent marks. Manual line breaking may also separate it from its base consonant. I have nevertheless counted it as part of the same syllable as a base consonant; the two stacks frequently interact in Northern Thai, with MAI KANG migrating to or towards the base consonant and interacting with its dependents.

The page was originally set up to use either my own stick font, 'Da Lekh', which is based on Deja Vu Sans, or the cut down version, 'Da Lekh Seri'. The Da Lekh font is intended to be suitable for use in preparing (but perhaps not publishing) Tai Tham text. It therefore includes work-arounds for known rendering engine problems. The Da Lekh Seri font deliberately does not include such work-arounds. You may be interested in using or examining my fonts for your own purposes.

I have added two other families of fonts. These fonts are available under the SIL Open Font license. The font 'A Tai Tham KH' relies only on the ccmp feature being enabled; it handles all Indic rearrangement itself.

The Hariphunchai font is an OpenType Layout font that looked promising when used with the South-East Asian shaper of HarfBuzz. Development seems to have drastically slowed when HarfBuzz switched Tai Tham to its implementation of the Universal Shaping Engine (USE). The code for this font is available on SourceForge and there is further documentation elsewhere. I have added work-arounds and a few further touches to enable it to work under the USE; I have dubbed the resulting font 'Lamphun'. I have included two versions in the menus, the 2014 version used for Lamphun, dubbed 'early Hariphunchai', and the latest (2019) version, dubbed 'Hariphunchai4'.

Feel free to adapt this web page to add your own fonts and test cases.

Content and Layout

The test text is given in the table columns headed 'Text', and is the content of the first table cell in table rows with class tst1. Two further columns, headed 'Encoding' and 'Hacked via ASCII', are automatically derived from this text as the page is loaded. The hacked column is intended to show users how the text should look, though it too may suffer from rendering engine limitations. The font used for this column is the member of the Da Lekh font family last selected to display the first column. Ideally, I would include images of the text from credible sources, but that may cause copyright problems, for the Unicode Consortium wishes to be able to use this document for commercial purposes.

The 'Hacked via ASCII' column contains an unambiguous transliteration to ASCII of the Tai Tham text in the column headed 'Text'. Members of the Da Lekh font family contain an open type font feature, Stylistic Set 2, whose enabling may cause it to render the transliteration as the original Tai Tham text as it is intended to be rendered. For more details, see the style sheet in the source of this page.

The 'Meaning and Pronunciation' column is given to identfy the word given as an example. There may be better glosses, and pronunciation can vary extensively within a nominal language. The letter RA is particularly variation between /l/, /h/ and even /r/, and there are regional variations as to whether vowel length distinction exist and, if so, whether they are phonemic. For the Tai languages the pronunciation is given using IPA, while Pali is simply transliterated (as Pali). I have omitted tone, as phonetic tone is also quite variable. Where no indication to the contrary is given, the Tai pronunciation given approximates that of Chiangmai.

The test words may, in principle, be extracted quite simply from this web page. Each test 'word' is the content of the first cell in each row whose class is tst1. For convenience, I have extracted the first two cells in such rows, along with titles, to a CSV file. Rows where there is a plausible case for treating the encoding used as erroneous are marked in pink. (Their CSC class is tst2.) For completeness, I have included alternative encodings which the Universal Script Engine (USE) calls for with an orange background and CSS class tst3 when they are defensible encodings. The USE encoding is not well-supported by fonts and is not robust to alternative classifications of combining marks.

The HTML comments within this web page should not be construed as holding test words.

Font Testing

This page is intended as a rendering engine test, rather than as a font test. However, you may modify this page to try out your own font. The necessary changes will be confined to the style sheet in the source code of this page, unless you use a different ASCIIfication scheme, in which case look at the usage of javascript variable ss02_hack.

My Rendering Performance

When this page was initially composed, in June 2015, the Da Lekh font currently mostly worked for the Tai Tham script in the Firefox and Chrome broswers. It worked in them because they use HarfBuzz to render the Tai Tham script. Since then, the HarfBuzz rendering engine used for Tai Tham has been brought into line with the Universal Script Engine, with a consequent dramatic fall in the rendering performance for the Tai Tham script.

The solution to this problem was to add numerous work-arounds to the font. These work-arounds have mostly restored performance, the main exceptions being subtle positioning errors where mark to base positioning is ignored and the default mark position is used instead.

The quality of the 'Hacked via ASCII' column varies from browser to browser and operating system to system, and also varies over time. For Internet Explorer 11, Microsoft Edge and for the HarfBuzz-based browsers Firefox and Chrome, it is actually the best rendered column. (Script-specific rendering engines have a tendency to make the achievement of advanced script features dificult rather than easy; Tai Tham has many 'advanced' features.)

Font Peculiarities

Traditionally, the consonants used in neither Pali nor Sanskrit did not have subscript forms. However, one significant text book, the 'big blue book', provides a subscript form for LOW FA for use in loans from English. However, this form is cramped and ugly, which goes against the tradition of Lanna script writing. The MFL treats the stroke distinguishing HIGH KXA, LOW KXA, LOW SA, LOW FA and LETTER UU from HIGH KHA, LOW KA, LOW CA, LOW PA and LETTER U as a diacritic. The Da Lekh font follows this interpretation, and leaves this diacritic above the baseline when the letter is subscripted.

If you have difficulty reading the Da Lekh fonts, you may find it useful to consult their glyph gallery.

Control Panel

Font Selection

name

head

Play Area

You may type your own text in the area below. It will, if possible, be displayed in the font selected above.

Table Completion

Tests

Vowel Combinations

These vowel combinations are taken from Revised proposal for encoding the Lanna script in the BMP of the UCS, ISO/IEC JTC1/SC2/WG2/N3207R, L2/07-007R (Everson, Hosken & Constable). Changes have been rung on the initial consonants to check for silly omissions.

A hyphen in the pronunciation indicates a syllable-final consonant that would be specified by a subscript consonant or following orthographic syllable.

Other Explicit Coding Sequences

Other explicit coding sequences are given in Revised proposal for encoding the Lanna script in the BMP of the UCS, ISO/IEC JTC1/SC2/WG2/N3207R, L2/07-007R (Everson, Hosken & Constable), and these are recorded here. Amended and exploratory material is highlighted in yellow; it is not vouched for by the proposal. The remarks are my own.

Other Examples from L2/07-007R

The actual coding sequences to be used here are open to challenge.

Mai Kang Lai

The mai kang lai character can be challenge to a font. The character has a wide range of behaviours. It can behave as a spacing final character (as in modern Tai Khün fonts) to a repha-like character, the old-fashioned behaviour seen in Tai Khün, Thailand and Laos. The MFL dictionary shows an intermediate behaviour, where marks above the following base consonant cause it to be positioned within the previous syllable. This is the style employed by the Da Lekh font.

Ligature NAA

This is mostly a test for readers!

English Loanwords

These examples are taken from the 'big blue book' pp151-6. Some of these renderings are unusual compared with the native tradition, and are included for that reason. The position of RA HAAM is particularly noteworthy.

The pronunciations given are guesswork where Siamese practice and Lanna script orthography conflict.

Tai Lü Blends

These examples are all taken from Graphic Blends at SEAsite. The pronunciations given are Tai Lü.

More Tai Lü

These words are taken from the MA thesis 'Development of Tai Lue Scripts and Orthography' by Apiradee Techasiriwan (อภิรดี เตชะศิริวรรณ). The pronunciations given are Tai Lü. Comparative material from elsewhere is highlighted in yellow.

And Not

The word typically meaning 'and...not' or 'and...then' may be written with a chained syllable, and this may present challenges to renderers. The form of the letter representing /b/ in a chained syllable presented an encoding challenge. N3207R proposed using the sequence <SAKOT, BA> for it, and using <SAKOT, HIGH PA> for the subscript form corresponding to both BA (common) and HIGH PA (extremely rare) in its rôle as a final (Thai sakot) consonant. During the ISO process, a new character was introduced instead for the special form, SIGN BA, and it is widely assumed that <SAKOT, BA> represents the usual subscript form corresponding to BA, both as a sakot consonant and in the Pali /mp/ and /pp/ intervocalic clusters.

When syllables are chained, shared vowel symbols are not repeated. This leads to ambiguity as to which symbol is dropped.

All the spellings in the table below represent the same careful pronunciation in Northern Thai, namely /kɔː bɔː/. The Tai Lü forms are written with different marks and pronounced with different vowels, but use the same two consonant forms in the stack.

Potential Surprises

These words behave slightly oddly.

Test and Tell

These exampless are intended to reveal the behaviour of the rendering system, rather than be clear pass or fail tests.

Rendering Challenges from MFL

These words presented problems, now overcome (Version 0.05), when developing the Da Lekh font to overcome the problems presented by the Universal Shaping Engine of mid 2016. (The solution is not entirely compliant with the Unicode standard - dotted circles in the input are sometimes deleted.) These are offered as an aid to font developers fighting unhelpful layout engines; they are not expected to help developers of the core layout engines.

Notes

References

Document History

This is Version 2.12 of the web page, which has been written by Richard Wordingham.

History

Testing

This web page has been developed with frequent testing on Firefox Version 54 and occasional viewing using Safari on iPhone (iOS 10.3.2), IE 11 (on Windows 7) and Microsoft Edge (on Windows 10).

Switching fonts has been tested in all these browsers.

Font Availability

Da Lekh Font Family

You may freely use my four fonts mentioned here without modification and may freely examine my fonts. See the respective licensing for conditions and modification. I do not own all the intellectual property rights for the Da Lekh and Da Lekh Si fonts. The fonts are available as follows:

If you wish to have WOFF files, you should either generate them yourself from the font files listed above, or simply copy them from this website.

The fonts are generated from the source code by means of a DIY font compiler that still has many rough edges. However, the source code of the font, although spartanly commented, may make it clearer what the font is attempting to do. I have endeavoured to make reverse engineering unnecessary.

The font Da Lekh is partly intended for my practical use in analysing material in the Tai Tham script. It therefore contains a large set of Latin characters to support transcription and transliteration. It also contains work arounds so that it may render properly despite problems with rendering engines.

The other purpose of the fonts is to explore issues in making an OpenType font for the Tai Tham script.

The font Da Lekh Seri is an unencumbered font intended for testing rendering engines. It therefore has, besides the glyphs for Tai Tham writing systems, just a bespoke set of (poor) ASCII glyphs; both the extra characters required by Microsoft Office and the characters recommended for the Universal Script Engine; and the characters needed for transliteration style (feature ss04) and their closure under NFC. Known existing work-arounds have been removed. This removal is implemented by compiler directives.

The font Da Lekh Si (ᨯᩣᩃᩮ᩠ᨡᩈᩦ) differs from Da Lekh in that it aims to reveal the spelling of words. This is useful when using a spell-checker, for example on Firefox. The ideal is that subscript consonants in the coda of an orthographic syllable would be distinguished from those in the onset by colour, whence the word 'Si' in the name of the font. The colour technology used works in the dominant browsers (Chrome, Safari, Firefox, MS Edge and even Internet Explorer 11) and in the word processor of LibreOffice. The colouring is also applied to chained syllables.

It is possible that Da Lekh Si may be reduced to an optional OpenType feature applied to the Da Lekh font.

The font Da Lekh Si Seri is an unencumbered font that colours glyphs in the same way. Like Da Lekh Seri, it deliberately lacks work-arounds for problems with renderers. It is intended as an aid for the development of the Da Lekh Si font.

Lamphun Font

The Lamphun font is available under the SIL open font licence; the applicable customisation declares that "Hariphunchai" and "Lamphun" are reserved font names. The font file is lamphun.otf and what I have used as 'source' code to build the font is an untidy mess assembled in lamphun.zip:

name

GSUB

GPOS

GDEF

cmap

OS/2

head

It is likely that I will create a variant coloured to indicate spelling.

Hariphunchai Font

There are two versions of the font used on this page. The fonts themselves are distinguished by the unique font identifiers in their name tables.

The early version of the Hariphunchai font, whose 'unique identifier' is FontForge : Hariphunchai : 5-5-2014, is available as Hariphunchai.otf both on SourceForge and within the Lamphun source zip file. The reversibly generated WOFF file is available here.

The 2019 version, whose 'unique identifier' in the name table is TragerStudio : Hariphunchai : 19-5-2019, is available as Hariphunchai4.otf on SourceForge. The reversibly generated WOFF file is available here.

The licence is available on Source Forge. The WOFF files, being derivative works, are licensed under the same licence. As the original OTF files can be recovered from them, they preserve the font names.