Links to PhD:

English | Other languages | Typography & fonts


Dictionaries and thesauri


Other languages

Software for computer-aided translation

Translation servers

Shown for each service is its translation of ‘this is the middle ear’ into French. I think that the one marked * is the best.

Dictionaries, grammars, language courses and general information



Other languages

Typography and fonts

Typographic design

Font software


On the Web

Evolution of lower-case ‘g’

Evolution of lower-case 'g'

The modern forms of ‘g’ with either one or two closed loops are the result of a long evolutionary path with multiple branches.

According to Analysis of the Letter-forms of the Vindolanda writing tablets, Fig. 11 no. 14, in the first century AD the cursive letter ‘g’ was written with two or three strokes, with the stroke for the top written last and sometimes connected to the following letter. This is apparently the origin of the small ear at the top right of the modern two-level ‘g’.

Steffens (1910), in his discussion of Merovingian letter forms, said:

La tête du g est souvent composée d’un trait ondulé, mais souvent ce trait forme en avant une boucle tantôt fermé et tantôt demi-ouverte; cette boucle est faite de bien des façons; en souvenir de l’ancienne forme, le g porte en haut, à droite, un petit trait par où il est possible de le relier aux lettres suivantes; ainsi s’explique le petit appendice qu’aujourd’hui encore on donne au g dans les imprimés d’écriture latine. La queue du g est d’ordinaire ouverte.

Hanzi and kanji

‘Hanzi’ refers to Chinese characters. They are also used in Japanese, where they are called ‘kanji’.

Jeffrey's Kanji Lookup provides a very flexible method of looking up kanji.

Wikipedia displays kanji using <span lang="ja" xml:lang="ja">…</span> with the actual characters represented by their Unicode values. I don't know what the lang and xml:lang attributes actually accomplish. Kanji can also be included in a Web page just using &#xnnnn; to represent the hexadecimal Unicode values; for example, &#x8033;&#x529b;&#x5b66; gives 耳力学.

This word-processing document demonstrates the above three kanji in a number of fonts.

The image on the right shows the same three kanji as generated by my Far font-creation software. The glyphs are defined in plain-text .arc files, which can then be compiled into Fortran subroutines by farrd, or interpreted directly by sign. The fin programme provides (crudely) for interactive display and design of Far fonts. The software development was started in the days of the Runoff text-formatting software (ref) and dot-matrix printers (ref), before there were word processors and inkjet or laser printers, as a way of getting math symbols into my manuscripts. It is still of some interest (to me) because the glyphs are defined in terms of strokes rather than bitmaps or outlines. The strokes are made up of straight lines, circular and elliptical arcs, and Bezier splines; they can have different pen shapes, and they can be tapered. Complex glyphs can be made up by combining multiple simpler glyphs, recursively. All of this is particularly appropriate for Chinese characters, where the direction and sequence of the strokes is well established and is useful in looking up characters in dictionaries, and where most characters are actually combinations of simpler characters. Even for Latin and similar alphabets, stroke direction and sequence are relevant for understanding the derivation of handwritten cursive characters from printed characters.

R. Funnell
Last modified: 2018-12-08 14:34:21