Resonate 迴響 wrote about Canto Font

It’s not just a font, but a Cantonese learning tool that does everything

唔好畀「字體」欺騙，根本係全能粵語學習工具

Can you imagine installing a font, and it somehow becomes a Cantonese learning tool? The Cantonese Font can actually do that. Usage is simple: just install the Cantonese Font, choose it for your text, and every character now shows with Jyutping with 99.7% accuracy.

你有冇諗過安裝一個字體，都可以係學習廣東話嘅工具？「粵語字體」真係做得到。用法簡單，只需要安裝「粵語字體」，然後轉換，就可以喺每一個中文字上面顯示出粵語拼音，而且準確度係達到99.７％。

Paucity of Cantonese Teaching Material Motivates the Work

坊間缺少粵語教材，決心做福仝人

Work on the Cantonese Font started in May 2022, and will be officially available in 2024-May-15. Jon didn’t expect this to take that long.

「粵語字體」由２０２２年５月開始開發，會喺２０２４年５月15日正式推出。不過，Ｊｏｎ從來冇諗過製作一個字體原來需要咁長時間。

“It looked simple, maybe two weeks work. Two weeks turned into two months, and then… two years.” Your friendly editor can’t help from chuckling. Good things takes perseverance, and it is never clear where the time went. “A friend, Nathan Hammond, wrote the code the draw the pictures. My own code was very slow: that is, for the 40,000 images needed for the 40,000 glyphs, it takes a month of processing time. For the ten variants planned it would have taken ten months just to run the code once.”

「因為諗住好簡單囉，諗住兩個禮拜咩都得啦，點知兩個禮拜變咗兩個月，然後兩個月……變咗兩年。」聽到呢度，編輯唔爭氣笑咗。好事永遠多磨，外人唔會明白時間用咗喺邊。「之前有一位朋友叫Nathan Hammond，佢幫手寫code畫圖。因為我畫圖嘅code係好慢嘅，意思係而家大概四萬個字，畫四萬張圖，要用成個月。每一個版本有大粵拼小粵拼嘅字如此類推，每樣都係三萬個字起跳，做一個系列已經五萬個……只係做呢啲都已經用咗十個月時間。」

If this is such monumental work, one could just cut the losses half-way. There must be some reason to keep going. Jon feels for the Hong Kong diaspora whose Cantonese seeds cannot find soil.

既然工程咁浩大，大可以中途斷纜唔做，繼續做落去總有原因。Ｊｏｎ有感香港人移民之後，佢哋嘅廣東話似乎無處尋根。

“I think that it’s futile for parents to demand kids to learn what they currently think is a useless language. But in five, ten years, some of the children, now grown up, would want so know ‘where do I come from,’ but that then becomes a tremendously difficult task.”

「我覺得家長要求小朋友學一種佢哋暫時覺得冇用嘅語言，係徒勞無功。但係五年、十年之後，可能啲小朋友會想知道『Where do I come from』，佢哋到時想學嘅話就會變得好困難。」

“People say Cantonese is a very hard language; I don’t think that is true. It is a language that is hard to teach and hard to learn, because we didn’t invest in making it easy.” Japanese can have N ways to read an ideographic character and one must memorize it all; in French, there might be three dozens of ways to conjugate “manger” (to eat). Cantonese, by comparison, doesn’t seem so wildly difficult.

「人哋話廣東話係一種好難嘅語言，我從來都唔覺得係真嘅。認為係好難教、好難學嘅語言，係因為我哋冇資源擺落去。」以日文為例，一個漢字隨時有Ｎ種讀法，只可以逐個死記；法文講個「食」字可能有三十幾種講法，相比起嚟廣東話似乎唔算好難。

“I think the issue with Cantonese pedagogy is the lack of useful material for teaching and learning.”

「我覺得廣東話教學嘅問題，係缺乏教學材料。」

Japanese uses furigana to guide pronunciation of the ideographs, Taiwan uses bopomofo symbols, and Korean Hangul is already phonics. It is only Cantonese that doesn’t show you how a glyph should be read. Many Hong Kongers aren’t even sure what are the six tones (or is it nine?); when you’re learning to write, they’ll tell you how we speak and how we write are different (di-glossic); when you want to learn Jyutping for romanization, some will tell you to use Yale, and others insist that the Guangzhou Cantonese is the orthodox that should be learnt instead. The beginner is confused and discouraged by the endless conflicting options.

日文會標假名發音，台灣有音標，韓文嘅諺文本身就係發音；似乎唯獨係廣東話，唔會喺本書上面教你點讀。好多香港人都搞唔清楚聲調係邊六聲或者係邊九聲；到要學寫嘅時候，香港人會話寫同講嘅係唔同嘅；當你想學粵語拼音，又有人叫你用番耶魯拼音，甚至話廣州廣東話先至正統，太多講法令人好混亂。

“If we want to find parallel bilingual reading material with Jyutping, for a language spoken by 86,000,000 people, there are only 70 of these book of which 40 are more pamphlets than books. The only book with Jyutping is the Little Prince.

The Education Bureau didn’t make a push for phonics, the publishers feel that there is no market, and if there are nothing to read, why would anyone learn Cantonese phonics? Jyutping, or Yale from earlier, is stuck for decades in this vicious cycle. If we can make the tooling, to lower the barrier for learning and creating works, we might just be able to break the cycle.”

「譬如一本字典，又或者想搵一本雙語對照、有拼音嘅課外讀物，粵語人口有八千五百萬人①，呢啲粵語書寫嘅書竟然只有七十本，當中四十本係好薄嗰種。想要一本有粵拼嘅書，就只係得《小王子》。我哋教育局冇去推廣粵語拼音，出版商覺得冇市場就冇咁嘅書，冇呢啲書就冇意思去學……成個粵拼或者再舊啲嘅耶魯拼音，喺呢幾十年都一直都喺同一個位置，如果可以做到工具幫手，令學習門檻降低嘅話，可能打破少少現狀。」

Don’t fool me, it’s not a font! It’s a dictionary!

唔好呃我，唔係字體，係字典！係翻譯機！

“When there are words, like 乾炒牛河, we can know that the ideograph should be read as 乾 gon1 and not 乾 kin4. When we don’t have a context, we need to choose a sensible default. For example, ‘區議員’ (district councillor) should be 區 keoi1; but what if we mean a councillor with family name 區 au1? Built into the font is an override suggestion system, where appending ‘.au1’ to 區 (區.au1) will change the Jyutping while subsuming the suggestion. The system thus allows full expression of the language and gives flexibility to the user.

「詞組比較容易辨認，例如輸入『乾炒牛河』，我哋會決定讀『乾gon1』字而唔係讀『乾kin4』字。有啲係比較難辨認，單獨一個字嘅時候，我哋要搵個default出嚟。譬如『區（keoi1）議員』，但係我哋係想講有個姓『區（au1）』嘅議員，咁點算呢？可以喺『區』字上面加code『.au』，系統會自動吸納另一組粵拼，幫你改做『區（au1）』，當轉番做plain text都可以對應，用起嚟更加有彈性，呢個係其中最主要嘅功能。」

The full family Cantonese Fonts not only includes Hei (Sans Serif) and Ming (Serif; on request of Ichiro) as the Chinese type-face, but one can also choose “show only ideograph”, “ideograph with Jyutping”, “ideograph with large Jyutping”, and “Only Jyutping”. I tried showing text with “Only Jyutping” and passages become just Latin alphanumeric, like some strangely encrypted message. Why was this variant made?

「粵語字體」全套除咗有黑體、明體可以選擇（明體係一郎兄嘅要求），亦都可以選擇「只顯示中文字」、「顯示中文字上有粵拼」，同埋「只顯示粵拼」三種字體。編輯嘗試將文章選用「只顯示粵拼」，結果成篇文章由英文字拼成，好似係睇緊密碼一樣。到底呢個「只顯示粵拼」嘅字體選項有咩用意？

“This in part helps foster some empathy for the teachers. Often when we teach Jyutping, we think it is easy: you just read the Latin characters! This isn’t how it actually works, and you’d know this if you try reading a passage with only Jyutping. When we read Chinese passages, we intuitively segment the run-on string of symbols into words from which we extract the meaning; beginner learners can’t do that. I thought this is interesting not just for learners who isn’t interested in learning to write, but also to help the teachers who may never realize how hard this can be.”

「呢個係同理心嘅問題。我哋好多時教粵拼，會覺得好容易啫，你睇住照讀就得，其實係唔得㗎。當你自己一試過咁樣睇一篇拼音，就會發現係好難。我跟住再做多幾個實驗，發現我哋正常睇中文，會自動喺句子做咗分詞，而學廣東話嘅人係分唔到嘅。我自己覺得，見到呢啲問題之後可以幫到啲老師，知道原來有一個咁樣嘅難處，而自己從來冇諗過。」

Another feature of the font is that when you enter English words (and tag them with braces {}), the word gets swapped to the ideograph and Jyutping. This also works the other direction, and it seems like a dictionary or translator.

另一個功能係，當輸入拼音甚至係英文詞語，就可以自動變成中文字，或者調番轉打中文出到英文，簡直係字典或者翻譯器一樣。

“At the start, a learner may just want the Jyutping. They may then blend in more and more Chinese glyphs. The fonts offer progression based on exactly what an individual needs right now: reading English and selectively choosing Chinese, or later reading Chinese and toggling Jyutping on-demand. There are many activities we couldn’t have done because it was too time-consuming; but now that attaching Jyutping happens instantly, our repertoire of learning activities can expand considerably.”

「喺學習嘅過程，可能開頭只係想要粵拼，然後加少少中文，再加好多中文。純睇英文嘅時候，或者純睇中文嘅時候，可以自己選擇係中文定英文，跟住自動跳返去有粵拼同冇粵拼嘅位置，對教同埋學都好有幫助。因為好多教學活動都做唔到，係可能要慢慢砌兩三個鐘頭先至砌到少少，而家就方便好多。」

Some features are works in progress, and others are Easter eggs for exploration. For example, there are entire banks of idioms and a comprehensive list of measure words, that can be activated with {idiom: n} and {measure: n}. Check out the documentations and try them out for yourself.

有啲功能未完善，有啲就隱藏咗，等用家自己發掘。例如係內藏成語表（輸入{Idioms}）、輸入{measure}就可以顯示廣東話嘅量詞表等等。更多功能可以去官方網站搵吓，然後自己試用。

“I used to be a teacher, and observed that our so-called education often grinds down the student’s curiosity. Without the pressure they actually would explore more. Maybe a student gets interested in the measure (classifier) words: why it cannot be 一個(go3)狗 but must be 一隻(zek3)狗; and ‘hands’ are not animals but it is also 一隻手. So the font provide this {measure: n} syntax, and it additionally let you add a question-mark ? to get explanations, and ex for an example.

「我以前教書，留意到喺教學程序當中，已經磨滅咗學生嘅好奇心，而喺冇壓力嘅時候，佢哋自己反而想知道一啲廣東話。有個學生覺得量詞好有趣，點解『一個狗』唔得，一定要『一隻狗』，手明明唔係動物，又要『一隻手』。你可以輸入個問號，字體就會話你知點先至啱，加個{ex}，就可以見到一個example。」

Jon taught Cantonese (as an extracurricular) in an international school, and his wife is Argentine. When he developed the font, he made sure that it can adapt to different (software) applications and (life) interests.

Ｊｏｎ教過國際學校嘅學生中文，太太又係外國人，更加明白學習中文面對咩困難，所以開發字體嘅時候，希望喺唔同嘅應用程式或者生活層面都用得着。

“We are demonstrating the font in Keynote, a presentation software. But maybe the user just wants to read, so the font can also be used in browsers, making the whole universe of Chinese — like Yue Wikipedia — accessible. Some people come to Cantonese because they liked Cantopop or specifically the songs of Cass Phang. By extending the functionalities of websites, they can now display all the Spotify lyrics with Jyutping. Many people gets attracted to Cantonese from cultural exports like film and songs from the 80s and 90s; with only glyphs and not pronunciations, how would they be able to know how it should sound and sing-along?”

「譬如我哋用緊Keynote示範，係Presentation程式；再簡單啲，譬如你上網睇嘢，整個粵語維基就可以睇晒；或者有啲人鍾意廣東歌、鍾意彭羚，顯示Spotify所有歌詞都可以幫你加上粵拼，然後跟住唱。有好多人學廣東話係因為睇電影、聽歌，仲係好渴求呢啲八、九十年代嘅文化輸出，但係如果只得文字，唔知道個字點讀，點樣跟住唱呢？」

Obsessions revealed

在命名中心呼喚愛

Jon is unabashedly King of Awful Puns, and we can see the depth of care even from the choice of version names. (Which, alas, no one sees.)

Ｊｏｎ自問係個爛gag王，佢將唔同版本都改咗名，從中睇得出佢對成套字體幾咁有愛。

“Version 1 (released last year 2023-05) was code-named Boundary; it let us see if there is a boundary, and whether there are New Territories beyond. Version 2 (upcoming) is called Pokfield (蒲飛) — y’know, they are all named after Hong Kong street names — and 蒲飛 also reads as ‘buffet’, where everything is on the table and the user can choose what they need. Version 3 [if it ever exist] is named Sycamore 詩歌舞 (the street name here is transliterated as ‘street of poems (si1), songs (go1), and dance (mou5)’), and the font would integrate with a reading pen and democratize these tap-and-read material for all teachers and parents.”

「第一版係舊年（２０２３年）５月嘅V1版本，個名叫boundary，因為我哋要睇吓有冇一個界限，然後界限以外仲有冇New Territory。V2版本叫蒲飛（Pokfield），大家可以想像到係關於香港街道，讀音都似Buffet，100％畀晒你，你按自己嘅需求嚟使用。V3版本係Sycamore（詩歌舞街），諗緊可能會引入電讀筆，老師同家長都可以用嚟讀到。」

The character and words content in the font is astonishing, and relies heavily on several open data-sources. These include Words.hk, character pronunciations from Prof Qin Lu’s team, Jyut.net by K T Shek. We move forward not alone, but by working with and learning from one another.

字體收藏嘅詞彙量驚人，全靠幾個開源嘅粵語資源。「粵典」、Prof. Qin Lu、石見田嘅「粵.net（粵音資料集叢）」，證實咗推動粵語並唔係一個人單打獨鬥就做得到，更加需要互相學習同幫助，先至可以令到語言環境更加充實。

Jon perceives the “Cantonese Font” not as a font, but a set of tools whose usage is determined by the needs and imaginations of the user. With this in hand, for example, it is far easier to prepare books with Chinese and Jyutping.

Ｊｏｎ認為「粵語字體」與其話係一套字體，不如話係一套工具，至於點樣用，全憑用家自由決定。試想像，而家有咗「粵語字體」，想出版一本中文字同埋粵語拼音對照嘅書，當然容易得多。

Lastly, our editors highly recommends the font. Just like Resonate, the Cantonese Font is self-funded and user-supported. If you are interested in the project, you can consider purchasing the paid programmes, which comes with extras as incentives.

最後編輯幫推，「粵語字體」完全係自資項目，同《迴響》一樣冇任何基金支援，有興趣使用嘅朋友可以購買收費版本，享用更多方便功能。