Unicode Conference 36: Day 1

These are live blog notes from the 2012 Unicode Conference in Santa Clara, California, 22nd October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

Developing an OpenType font for complex scripts using Fontforge

Pravin Dinkar Satpute, Senior Software Engineer, Red Hat

Pravin Satpure: I am working for RHAT 7 eyars in i18n, working on Indic scripts. I developed 6 unicode chars for devanagari. Pashmiri language. I’m project lead for Lohit fonts, supporting the 9 major scripts and are the default fonts in Fedora, Debian, Ubuntu, and some are used in ChromeOS and Android. Wikipedia loads some

I am also leading the Liberation Fonts project, the default fonts in LibreOffice.

I want the world scripts in Unicode, so everyone on earth should have their language on their terminal.

I was to show production of OpenType fonts, why they are important, FontForge – a 5 minute demo for Latin – then whats involved in OpenType fonts for 20 mins – then demonstrate making a Devanagari font with FontForge.

Imagine a developer has made an excellent applicatoin, but the first screen in that app has broken fonts. What will you feel if you see the square with 4 numbers, or a dotted circle?

A friend was working in Embedded systems, and faced a problem, he was getting U+0916 instead of U+0915. It took 2 days to realise it was the font that was buggy! Understand fonts is important 🙂

Why this session?

1 If you see the Unicode from early version today, many scripts were added. 5.0 to 6.0 added Khadoshthi, Lepcha. If normal people want to use these langugaes, they can’t, because fonts either don’t exist or are VERY expensive. So we should have libre fonts for these languages available to all developers, so they can support these users in their applicatoins.

2 People in those communities WANT to do font development. I tell them Unicode is a good thing. We don’t have users from this session here today, but I hope they will learn the things we are covering today.

The number of script in unicode is increasing.

There are several knowledge domains involved OpenType font. Linguistics, knowing the language and writing system. Art, the visual drawing of letters. Technology, since the letters are shaped by an operating system, so you need technical knowledge to write the layout table.

In the RHAT office in Pune, there are people from all the language communities I can consult with.

Why FontForge?

Its the only libre font editor tool available today.

I started using FontForge in 2005, it was very complex to write OpenType then. Today its very easy!

Continue improvement: There are active users and developers around this tool, and you’ll typically get a reply to your query within 1 or 2 days. I found a problem with the grid fitting tables and posted about it on the mailing list and got a fix later that day!

It runs on GNU+Linux, Windows and MacOS X.

What is a complex script?

[audience ideas]

In Devanagari we have reordering of characters, what we type and what we see are different; what we type looks totally different to what we see when its ligated.

Can we call CJK complex? I don’t think so, at the rendering level no. There are huge character sets, 6k in Japanese, Chinese is more, but the complexity is on the input level. I feel personally that Indic scripts can go like that; every syllable in an Indic language will have a key code and the rest will be automatic. I hope one day 🙂

The complexity can be in the OS level or Font level and its complicated. Win, Mac and GNU+Linux have different shaping engines for OpenType standard.

Indic and Arabic have re-ordering all the time.

I will do a small demo of OpenType in FontForge.

[GNOME3 accessiblity panel has a zoom tool]

I open FontForge git latest. [Default theme]

I open a LiberationSerif-Regular.sfd and copy the 4 abcd glyphs. The em size isn’t matches so I scale them from Glyph origin 50%.

Then I go to MS Typography site,and find the OpenType specification, Features page, and see the documentation for the LIGA feature.

Then I go Element, Font Info, Lookups, Add Lookup, Ligature Substition, add a liga feature for Latin (default) script.

Add subtable, add the Ligature Glyph Name colom first, tahts the name of the final shape, and then in the Source Glyph Names col, I type

a b c

And a mouse hover shows a preview.

Q: Its glyphs or characters?

A: You can see the glyph with glyph name a is associated with the unicode value U+0061 which FontForge does automatically.

Now I generate a TTF and install the Test1 font, then in gEdit I can pick the Test1 font, and if I type abc then I see d!

Its the font doing the magic here 🙂

Q: How does Japanese ruby typesetting work? Where you have kana written above kanji?

Jungshik: Its done by browsers with CSS, its Harakana glyphs placed by the layout engine entirely.

Q: How is the order of subtition rules defined?

A: Its defined by the rules, you can see in the microsoft.com/typography/…/otfntdev/features.html its ccmp, liga, clig,

OpenType is a cross platform standard by ADobe and Microsoft, extends TrueType format by Apple. Its cross paltform, has i18n characer seupport, large glyph sets, and supports many advanced typographic features.

How do they work?

Unicode characters are input into a OPENTYPE LAYOUT SHAPER (OTLS) which also takes as input a OpenType font, and outputs GLYPH IDs in that font.

The OTLS reviews the sequewnce of unicode chars and asks what kind of char is it, and based on the OT spec it applies a number of faetures to the input chars. the shaper searches for these features in the fonts and processes the features and finally outputs the glyph ids.

In FontForge we can see these Glyph IDs in the Font View as the first number on the toolbar.

The whole magic behind OpenType is the OTLS. We can develop an OpenType font but we can’t b

Jungshik: Glyph ID alone isn’t sufficient, the OTLS also emits x y positions

Q: And with the GlyphIDs and positions any dumb renderer can render text?

A: Yes, this is a part of the overall rendering stack

Here i write in Gedit ‘pravin’ with 3 syllables in Devanagari.

The OTLS does this:

1. Analyse the text

2. Reorder chars as per script requirements

3. share glyph sequences with GSUB then GPOS

eg, Uniscribe on win, ICU in LibreOffice, Harfbuzz-NG on many free systems. A few years ago pango, qt and icu were all different OTLS and a few years back, harfbuzz was started to unify them. harfbuzz is meant to be fully compatible with uniscribe, so a single font will work the same everywhere. This makes font development easier.

Jungshik: What is Apple using on Mac OS?

A: They use AAT

Jungshik: But they support OpenType too, in addition to AAT they support OpenType to CoreText. I wonder what is behind that.

A: I wonder if they use harfbuzz?

Jungshik: Perhaps ICU, but I don’t know.

A: I have used AAT and I like that approach

Jungshik: I like it too

Q: Split vowels?

A: No, not in Devanagari, but in say Malayalam


[Brief talk about Unicode encodings]

designers draw shapes on paper and scan that. drawing on screen is different, they are very used to drawing on paper. they draw and scan and place on background. once its in the backgroud, they say Element, Autotrace.

FontForge has vectorised the points. Now remove the background image. The problem with this is that there are SO many points. We can remove these with ‘merge’ points, select them and CTRL-M. We can also draw around the background image instead of using autotrace. I am no artist and it took me a whole day to draw the indian rupee symbol.

Q: Do you have different Rupee shape for each script?

A: We do a little bit

So, lets copy some glyphs from Lohit Devanagari to a new font

You can see that font development is time consuming process and I think thats why there is not much community contribution to libre fonts.

The ligatures dont have Unicode points so the glyphs belong in the Private Use Area.

Q: are there rules to assign PUA points?

Roozbeh: We are meant to not point any CMAP table at those glyphs

We can set the glyph name to ‘khsa’ and assign the unicode point to -1, and pste the base glyphs into their correct slots.

Lets sets up the GDEF table. Go Element, Compact, we only see the glyphs that are IN the font. Now we go to the base glyphs in Font View and right click, Glyph Info, and set their OT Glyph Class as Base. THen the mark glyph we set as Mark.


now we set up the GSUB table. Font INfo, lookups, gsub, new lookup, ligature substitaion, akhn, for deva {dflt} script, and we can see this is set up correctly in the tooltip preview

Now in the metrics window we can see this working live.

GPOS, FontInfo, Lookups, GPOS tab, Add lookup, we want a Mark to Base position feature, abvs, for deva {dflt} script.

For testing purposes, we can see if we add the wrong feature.

Now we add an Above Base Mark GPOS lookup, and a subtable, name the anchor ANCHOR. then go to the base glyph, Point, Add Anchor, place it.

We add this to the mark glyph and then Point, Add Anchor, place it, and we can see in Metrics View that it works.

You can see moving the anchor point in the glyph view updated in real time in the metrics view.

In GPOS, the anchor point is a key idea.

How to debug for problems? there can be many issues, especailly with ‘cyclic’ features, where a b c becomes d and b c d becomes e.

I’m lucky! Are you aware of Nastaleeq? Its one of the most complex scripts. I’ve never seen a more complex script. Needs 1,000s of lookups. Debugging it is very hard!

Be patient. THat’s the only option we have.

Q: Do we have automated tools for subsitution that can become infinite and lokc up?

A: Yes, its possible. In FontForge if we do a sequnece, we pass that to harfbuzz. If we can test directly frmo FontForge it would be easier. Same problem with MS.

Q: Who defines the tags?

A: MS and Adobe, its an ISO standard, OpenType and Open Font Format are the same.

Roozbeh: Another good libre tool is TTX. It dumps the font as XML and you can edit tables without any side effects. Its nice to make precision edits. Play with existing fonts, dump them, see how they are arranged.

Jungshik: There is no automated sanity checking.

A: I’d like to see harfbuzz integrated into fontforge

Q: Is opentype sufficient for everything?

jungshik: nastaleeq is hard. it pushes the limits of opentype. you can not fully implement it, as needed by urdu speakers, you must do some compromises.

roozbeh: you want geometric calculations, move dots based on actual text, so fonts will always been a compromise even with a more advanced engine than OT. www.flickr.com/photos/pimu/4671362490/ You can see 2 baselines, every word has its own base line and there is the line baseline. tools arent ready for this.

A: in OT fonts, the font developer is totally dependent on the OT spec. AAT gievs total control to the font developer, so i think it is the best. OT shaper has so many implementatinos that differ. we should have a single shaper to make life easier. long term, a single base can solve all problems and move forward.

Q: when creating gsub rules, how can you verify they are correct?

A: thats where a linguist role comes in. you test by installing the font and typing text using those rules and see it is correct with native reader knowledge.

Finally, here is a testing tool:


Its libre licensed, under MIT license, in 2010.

Posted in Knowledge | Leave a comment

Moral rights and the SIL Open Font License

Originally published in Libre Graphics Magazine Issue 1.4.

Designers often express caution about Libre culture. Letting other people make modifications to their designs is unusual for them. This is especially true for type designers, since the harmony of a typeface design can easily unravel if changes are made carelessly. Copyright is the major legal restriction that stops others from making modifications. In some places, though, there is another restriction: moral rights.

Moral rights are explained on the English Wikipedia as “the right of attribution, the right to have a work published anonymously or pseudonymously, and the right to the integrity of the work.” In many places such as France, authors are not allowed to transfer, sell or give up these rights to other people or parties.

Recently I was discussing the SIL Open Font License with France’s highest profile type designer, Jean Francois Porchez. He asked me if these moral rights conflict with Libre licensing. The SIL Open Font License is the most popular license for Libre fonts and—I believe—it does respect moral rights.

Let me walk you through the OFL’s compliance measures.

Section 2 of the OFL requires that “each copy contains the above copyright notice.” As a copyright holder, your copyright notice (which can include your url and contact email) will remain with all copies and derivatives. This ensures that your attribution is made available.

Additionally, there are FONTLOG text files commonly distributed with OFL fonts. These have a detailed description of the fonts, a timeline of releases, and details about all copyright owners and authors including those of any “parent” fonts. This is not required, but it is recommended by the OFL as a best practice and it really helps with good attribution.

Another side to attribution is a case in which the author wishes to be misattributed, in a way: anonymously or with a pen name. In this case, designers can simply use their handle, or “Anonymous,” in their copyright notice and other fonts’ metadata.

A third side of attribution, what we might call bad attribution, involves the name of the designer of a parent font being used to promote derivatives. Section 4 of the OFL provides a guard against such misuse to “promote, endorse or advertise” a derivative.

Such attribution issues are quite straightforward. The concern is really the moral right to the integrity of a work. English Wikipedia explains that “preserving of the integrity of the work bars the work from alteration, distortion, or mutilation.” Allowing others to make alterations is fundamental to Libre culture.

Personally, I believe that saying an alteration of a typeface is a “distortion, or mutilation” is an entirely personal opinion. If I chop off the serifs from a typeface and re-space it, I think it would be strange for anyone to say that I mutilated it. Others can say I did it poorly, that making a related sans is really a new design project and many other details of the typeface must be changed to re-harmonize it. But for my personal identity to be so wrapped up in my work that I could have hurt feelings in this way seems bizarre to me. Yet many artists and designers do feel strongly that the work they have made should not be altered by other people at all, even in the countries where moral rights are not instituted in law, like the United States.

The identity of a designer is associated with his work. In type, there is a common association of the name of a typeface with the name of its primary designer. Making a change naturally invites altering the name to reflect differentiation. Using the original name in part could cause the association with the original designer to carry over too far. The OFL has a solution to this. It allows designers to set Reserved Font Names, which act in a similar way to trademarks.

If you declare a Reserved Font Name, the OFL ensures that when your font is modified then the name you chose for the font will not be used — unless you give additional permission. This naming restriction encourages people who make ‘corrections,’ who adjust minor details in the font—tweaking a kerning pair here, adding some metadata there—to contribute their changes upstream, back to the original designer who maintains the font under its most well known name.

2014-07-02 Update (not in Libre Graphics Magazine): However, the Reserved Font Name restriction means that such fonts may not be conveniently used as web fonts, since format conversion and subsetting are very common modifications for people hosting web fonts (The FontSquirrel web font converter is very popular.) Therefore, I recommend never using RFNs. Encouraging upstream contributions is important, but happens much less often than web font hosting, so not using RFNs is a good tradeoff. Plus, times are changing: today many OFL font projects maintain their source files in version control systems like Github. With version control, many copies are made by many people, each intended to eventually be part of an approved, tested, release; but it doesn’t make sense to keep switching the names around, and contributors won’t find your project if it uses a ‘codename.’ So, instead of RFNs, I recommend telling people that you welcome contributions. That will do more for this goal. Larger organisations with valuable names are likely to already have unregistered or registered trademarks, which can also be used to achieve the same goals as the RFN with more flexibility; as a trademark holder you can allow various uses of your mark and have legal recourse if you need to pressure someone to not use your mark.

In all of these ways, among others, I believe the OFL respects moral rights, while maintaining the principles of Libre culture.

The author of this column is not a lawyer. This column is meant as a set of opinions, not as legal advice.

Dave Crossland believes anyone can learn to design great fonts. He is a type designer fascinated by the potential of software freedom for graphic design, and runs workshops on type design around the world.

Posted in Knowledge | Leave a comment

ATypI 2012 Hong Kong: Day 5 Morning Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 14 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

A study of punctuation marks’ history in Hanguel – Jaehong Park

Here you can see the ancient comma and period marks, comma are small circles connected to the foot of a stem and periods are the circle under the glyph.

Jungshik Q: Western punctuation marks are well established in Korean print. Will your suggestion be adopted?

A: No, its not a suggestion for adoption. Its an experimental typeface design. In 1933 we borrowed English, and I couldn’t find so many in the history. Korean type designers made type designs without thought to this aspect. This talk is just to show and rethink the punctuation marks.

Q: Quotation marks; the good reference for a quote mark, the French/Italian/German style << >> looks better than the English ” ” and would perhaps fit the forms of the Korean script better.

Jungshik: Korean do use French style quotation marks, such as in book titles.

Q: Thanks for this interesting presentation. 2 questions, 1, your design is experimental but doesn’t reflect the characteristics of the script, the forms seem unusual to the current korean forms.

I’m a type designer. I work at MS. I wrote a brief, Can a true italic be designed for Korean? The next question is SHOULD an italic be designed?

I don’t believe western type concepts should be added to non laitn scripts, but this case, its a genuine fit. Roman italics is a faster handwriting style, then it became part of an emphasis.

We saw earier today the emphasis mark as traditionally done. Why we need an italic? Korean italics, you have mechanically slanted forms be used. Its a mark of coolness, differentiation. Users see slanted text in latin and want to harmonise the korean so it looks better. the emphatic use of slant is useful.

But auto slanted doesnt look good. The handwriting type in korea is from beautiful hadnwriting; these are sandoll’s foundry specimen book pages. Here is the upright form, and in an example of use in their book, its artificially slanted to the right. Maybe harmonised with latin, maybe it feels cool. Not sure.

But why not use the handwriting forms with the type forms? In a headline it can work (example) but it wont work in text.

At U o Reading I looked at signage. You have many sytles, fat rounded, uneven alignment lines, brush styles. But an Italian restaurant “Santa Rosa” had this signage. A right slant, a brush style forms, calligraphic. But it has a horizontal axis, not vertical.

Traditional korean claligraphy has very vertical lines. Its part of the key elements of beuatiful korean tyep. So this with its horizontal axis/stress was special.

The cursive form or veernacular is not the masters of calligraphy, its just fast writing. it jeeps the vertical stress and axis but it adds ligated forms for speed so you get cool interesting forms.

I wanted horizontal forms, high contrast – brush inspired – but it was too brushy. ‘This reminds me of Edwardian Script’ – too decorative, too far from upright and scribal hands.

We just saw a lot exmaples of this handwriting. The writer said slanting it was unnatral. Why? I tried to write Korean quickly and it has a right slant. I realised my elbow position was differnet. Close to my body, there is more slant, than away.

Here is a piece of lettering on a CD cover that has horiznal axis.

So I made Saja, a different feel to normal uprights. Its ligated forms, 5-6 strokes, the top has 1-2. This merges latin slant, brush forms. Here it is like emphasis in an upright body of text, like latin use of tialics. how can we do an italic that stadns out? I think this will do it, a slight slant and a cursive form.

This works well when mixed in to roman tialics. if you have design elements that work well together, you cna put them in use together and they work well. You can do separate-but-equal typesetting as well as interleaved.

This is a new concept fo rkorean. Noramlly thye just have slanted type, its not beautiful, so i think a new appeoach is imprtant, but its not unprecedented. its already done but poorly. A new dierction will help diversity while maintaining teh value of the text.


Q: The tilaic you did, will that be a style?

A: Yes, just hit the itlaic button 🙂 thats the magic

John Berry: The last 3 talks all focused on modulation of text, especially punctuation. They all seem to be able ‘how to read the text’ like brinhurst says all punctuation including spaces are like musical notation.

A: Yes, i am excited about korean type because there is a focus on function and form, people think about how to display content.

Q: How does yours compare? in every script there is their own italics. can korean italics

Posted in Knowledge | Leave a comment

ATypI 2012 Hong Kong: Day 4 Morning Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 13 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

11:25 Typography between Chinese complex characters and Latin letters – Mariko Takagi

I wondered what I can do to help bridge the 2 cultures of east and west. My EN <> CN type primer, has 6 princpiles of chinese chars and some key ideas.

To talk about typography, we must have the same vocabulary. I saw a slide in yesterday’s Thai talk, talking about ‘looped terminals’ and I’d been to a train station show in the slides and said, “I’ve been to the looped terminal” 🙂

I am a native German reader, so I had to learn the English typo terms. How can they be translated to CN? In many books there are the ‘8 principles of yong’ which are often shown with made up characters, that show different strokes but aren’t real words. The have quite a square form.

WE started collecting and comparing sources from CN, JP, Taiwan, and other places like Germany. Last year Susanna Licko (?) wrote a book about CJK typography.

CN chars can have a single stroke, but the most complicated has 64 strokes. You can see the 2 CN and EN texts, there is a row of headings, DOT DOT DOT, and the CN headings are visibly different terms for various dots. In EN we dont have discrete terms.

Horizontal bar, vertical stroke, these are easy to identify. downstrokes can vary a lot in angle, but all are downstrokes to the left; downstrokes to the right also vary in angle. Upstrokes left and right. Angles or corners, ‘returning stroke’ or folds,

11:50 Hanzi of the West, letters of the East – Christoph Stahl

‘Translation Server Error’ as headline on a billboard 🙂

The CN text input will add a space after a comma, so in EN you see a comma place ,after the space, and a space added after an apostrophe.

Some peopel can see ‘made in china’ from lousy latin fonts. Slowly CN foundries learn they should have better quality latin.

Now lets see how europeans use hanzi. Frederic William of Brandenberg had 2nd biggest CN library after French king. You can see a book made totally second hand, looks nothing like the real chars.

Here is a 1641 book, ‘relations of china’ and it uses the christian crossl + symbol, lol

1688, ‘new relations of china’

1655 ‘history of china pt 1’ – there is this idea that they are pictographic, so a euro man saw fishes in the ‘river’ char when there is no such thing.

Here’s a typeface with a thick border around each hanzi glyph.

Soon the euro founders invented split sorts, which allowed euro typographers to make up bizarre combinations of radicals.

1815 first bilingual EN-CN dictionary by Robert Morrison. He had one size of a wooden hanzi font. his workers desrted, all kinds of drama, but the book came out. Like roman said, euros got tired of this, they saw too much effort for too little effort. So sinology books have CN terms with romanisation, but there are many different romanisation systems. This means the same person may appear as 2 differnet names.

proportionality and monospace is tricky. latin text in chinese body means the latin text offsets the grid. so you align the hanzi on the grid and place latin words in the gap?

You can also use half wide and full wide latin spacing. not great.

Hanzi usually today uses justified setting, which means the same latin word has very different tracking even in the same sentence.

You can put latin in CN with a baseline alignment, but that means a step down for each latin word; so i think its better to have a middle alignment. even capitals and lowercase cna have a differnt baseline which looks odd to euro eyes but makes smoother reading for CN native readers.

A high x height helps fill the tall space of CN.

Line spacing can be baseline aligned with 3 lines of latin next to 2 lines of CN.

You must have a feeling for a language, you can’t just go ahead and apply the conventions of your writing system’s typography to another one, as they did in the past. This awareness should be done in all typography education arund the world.

100 years ago euro typographers gave up on hanzi and now with digital type we have the change to try again, if we work together, we can make it work, and the effort to bring out cultures together can never outweigh the benefit.

Posted in Knowledge | 1 Comment

ATypI 2012 Hong Kong: Day 3 Afternoon Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 12 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

16:25 Black and white in Indian typography – D. Udaya Kumar

Tamil Nadu is a southern Indian state, its also an fficial language of Singapore and Sri Lanka. Its literary history starts in 300BC.

It has 247 letters with 12 vowels (life), 18 consonants (body) there are manyn vowel consonant conjuncts, loan letter sfrom Grantha script, some ligatures, and our own numerals.

All indic scripts come from Brahmi script; the Cankam script was 300BC-500AD. There were 3 great kings in india, and a kind promoted tamil as the official script in the 9th century.

There were stone carvings, copper plate etching, and palm leave manuscripts.

Stone, metal, palm.

A palm leaf survives for 300-400 years. There may be text from say 300BC but that has been destroyed by time. But every 300 years they copied the text to a new palm. The copying meant the script evolved. We can see the original forms because of the stone forms.

I tried to research these changes over time. Palm is used for many things in India, building houses even, and the palm leaves are boiled and dried. You can’t just take a leaf off at tree 😉 The processing various across india, north and south, and in S E Asia too, each have their own process. This is a processed and prepared leaf; its smoothed with pebbles.

Tamil has a unique way for using the medium. Here is a northern manuscript, and their tools of writing and script forms are totally different; a reed pen with an angular cut, that writes. THis makes the forms very calligraphic. Tamil is written with a pointed metal stylus that scratches the palm leaf, and then we apply the ink to the etching.

The manuscript is not written on a desk, its held in two hands. You must sit, you can’t write when standing. There is no ink in the writing stage. I believe there is thus no concern by the scribe about the black and white of the forms. Malayalam, you can see lots of curved forms, as a horizontal stroke is easy and a vertical stroke is hard to go across the palm lines. If you go horizonal, it will easily cut the lead. So the south indian scripts are rounded so that a scribe can QUICKLY write that overcomes the physical limitations of the medium.

Malayama is used in the neighbouring Kerala state.

So Indian typgoraphers like to use multiple bright colors, and lts of text effects. All the magazines have these effects. There isn’t a grid strucutre. We love to play around, use as much space as possible.

These are popular magazines in the south. At least 5 typefaces on the cover. This is INSIDE the magazine – each sentence has a different fore and background color!

This is a wedding invitation, I did it before all nice with a grid and helvetica and they said WHATS WRONG WITH YOU and did it the normal way.

A metaphor?

If I show this to my Mum, she says the traditional lady on the left looks good and the western lady on the right should stand properly, put more clothes on and eat better.

We want more ornamentation, more decoration, more colors, and we like it.


Modernism ideas are rejected. We should embrace what is there and understand the culture and do typography with that point of view. In Modernist type we look at the grey value, balance things and seek harmony. For the Tamil typographers these qualities may not be desirable.

I would like to thank my team members, Prof U A Athavankar, G V …

Thank you!


16:50 Simplified Arabic – Titus Nemeth

[nice quote]

Arabic type was slow to arrive. It was a slow and tedious process to compose arabic with movable type. 500 sorts, heavy kerning prone to breakage due to the pressure of the press. Until the mid 20th C the region wasn’t as supportive of typography beacuse colonial empires were in control. Ottoman empire.

This changed in the 1910s, a migrant arabic newspaper in brooklyn used a Linotype to compose its text. It was faster, a simpler workflow, but it imposed its own characteristics; it could not kern, it had a physical limit to character sets. 2 magaziens meant 180 chars. Reducing 500 sorts to 200 is a drastic reduction in type form variation.

Printed type remained marginal until around 1950. The biggest arabic newspaper at 1900 had 10,000 copies. Daily Mail in UK at that time was 1,000,000 copies a day. The region was fragmented with new and somewhat artificial national boundaries, and they sought identity. This is where arabic print really develops. This is what arabic typography as we know it today really developed.

The Al-Hayat newspaper in Beirut, Lebanon exemplifies this. Kamel Mrowa was a Shia muslim in Beirut, and it was a commercial success. In 1954, he approached LInotype for new machiens; he wanted faster production, to increase print volume.

He said he knew the arabic typewrite, around 50 years already, and they wrote arbaic with a simple tool. Unlike chinese or japanese typewriter, it had 45 keys and a shift for 90 total. From 500 to 90, it was a crude represenation.

Here’s how it worked.

Most arabic letters are connected, and can be reduced to 4 forms: initial, medial, final and singular. …

Walter Tracey was the type director at the time, he thought it was a good idea. Middle east was a new market for type manufactorers to enter.

Remember the wealth of letter shapes in arabic. Here are the 24 shapes a letter can take at the start of a word, changing based on how what is followed.

A typewriter reduced this to just 3 forms, singular and connected. (?)

LInotype was slow. They took an existing type and threw out some chars and modified existing ones. Kamel Mrowa was frustrated it was so slow going, and went to England to help. The University of Reading has in its collection some samples of this work.

It took years, but in 1959, simplified arabic was annoucned. It was big news for newspaper publishers, a 30-50% increase in output in a market where time is money is good. The author of the press release stressed that Kamel Mrowa was the author; perhaps Linotype wasn’t too sure about the invention of simplified arabic and put him with top billing. But in the following years, it was heralded as a great breakthrough by LInotype.

the type was expanded, an isntant sucess for the company. it found appreciation in the copmetitoin with imitation. Intertype had copied LInotype since the patents expired. Intertype was a total clone of linotype, to the degree it could exchange spare parts.

Intertype annoucned their simplified arabic a year later as ‘Abridged Arabic’

While Linotype had a patent on the system they didn’t go to court. The idae of shape reduction was copied by Intertype, and they improved the system; they added some characters to give a slightly more authentic, less simplified look.

Linotype took another route. They hired a calligrapher, to design a typeface specifically for this technique, and the Mrowa-Linotype Number 8 with 9 combiend the improvements of Intertype with a new design.

Intertype customers soon switched to Linotype since it was so much better, and by 1967 almost all typesetters were Linotype users.

In 1967 another type was released for the simplified arabic system, Yakout. A calligrapher was assasinated in his office and disappeared from Linotype marketing materials. It was the most successful and most copied.

In 1980 they had 80% of the general market and 90% of the newspaper market. You were guaranteed to read these types. This remains true today, with clones very popular.

But in moving to digital, the clones continued. Letters that are not the same have the same base forms, when there are no limits on digital character sets.

MS Arial and Times New ROman’s arabic used the simplified arabic system too. A typeface made for a narrow, specific cirumstance of newspaprs int he 1950s became a de facto stardard of typography.

A handheld device, it may be ambiguous and curious that this shares ANYTHING with the design for a linotype machine in the 1950s.

The quote from the start, we can subtitute script for type and writer for designer. There is no need to repeat the compromises of the past in the present.

Please learn the history, learn why things were done in the past, and consider how to do things newly today.

Jungshik Shin: You said the final and isolate forms were 1 form and the initial and medial forms were 1 form.

Titus: They CAN be treated separately.

Arabic glyphs in arial still use simplified arabic, just 2 forms?

Titus; Yes. The new font for a mobile has the 2 forms. There is no constraint any more for using more than 4 forms. Returning to my quote, its more work for the designer of course. If something contributes to the convenience of a reader, the designer or writer should do that.

Jungshik: OpenType has a restriciton on Nastaleeq?

Titus: yes, there are issues with opentype, it can’t handle everything – that is WAY further than simplified arabic.

Jungshik: Simplified Arabic ever used for urdu or pashtu region?

Titus: Urdu is a specific case. They use the arabic script in a different way, and until VERY recently, their newspapers were handwritten, and their websites are all single images linked, there is no real text. There is little or no use of ARial and TNR arabic by Urdu writers.

Jungshik: How do they do printing?

Tirus: Lithography. They handwrite on a lithographic stone and print from that.

Gerry; You can see a lithograph nastaleeq newspaper in the exhibtiion tonight.

17:15 Notes of Helvetica Thai: loop-less terminal in Thai typefaces – Anuthin Wongsunkakon

I look kinda chinese but I’m not and I got help from a friend with the Chinese text in my slides.

I’m from Thailand, and we speak Thai there. Its a 513,000km^2 sized country, about the same size as france or california, 65M people, 95% buddhist. Bangkok has 6.5M people, a single state; capital, port, political center.

Think of Thailand? An elephant ride? People ask me if I ride an elephant to work 😉 but you do see them on the street for tourist purposes. There are the buddhist ruins, the floating market (that is totally far from Bangkok). Its like any big city, lets see the street typography; Air Asia ads, KFC ads, a loopless terminal type in both. CP, a big food company, blackberry, epson, citibank – all loopless in their ad headlines, and citi use it in text too.

Averege income, $490/month for someone with a BA degree. True? Yes. MA, $685/month. The foundry business, you think how much a person like that spends on fonts. 26 minutes a day reading, 40% prefer TV to books. So what kind of type we read in thailand in books? it doesnt match the street signage and billboard ads.

This is a knockoff of Interstate. This is street pedestrian signage in Helvetica stretched narrower, next to a looped terminal thai type. This bank poster and this Ikea signage is nicely harmonized.

This is a magazine spread, the way we teach in design school, you MUST use looped terminal type for text. Smallest use is 14pt, you do 1 full page col or 2 cols 50%. But they use loopless terminals for captions.

So if loopless terminals are so bad, why would we use them for captions? When you reduce point size you can say more, the area can contain more text. The loop terminal at 9 or 8pt, you see the little dots everywhere.

Thai Letter vs Thai Typeface

Old Thai stone carving lettering. This kind of text set the tone for what our type looks like. 700 years ago it was upright, then it became slanted like in this stone. But with movable type EVERYTHING became upright again.

It effected our hand writing. We all started writing upright, lost our little slant.

More bizarre, this is the face everyone in Thailand thinks is THE default type, if you want to type soemthing in an office, you use it. We call it French Face, designed by French missionaries, it is latinised a bit, euro style stroke contrast.

If it was introduced now it would be awkward but after 150 years its highly readable.

WE have looped terminal, this is Tonbori, the default Mac OS X thai type, and this is more ‘latinised’, Sukhumvit, with a non-looped terminal.

Wallpaper, Maxim, they use no loop. This ia school textboook with looped terminal.

This is a VTS station, its helvetica with somtehing that looks odd to me, not the looped terminal, but the way the classifications dont match.

This is more matching to me.

Legibility for the old generation? A latin type has a family of styles, weights. In thai, before digital type, phototype had little variety in weight. This estabilshed workarounds, so its acceptable to use a different type that is bolder to get emphasis.

This one, you see the outsider typeface; an old Ford newspaper ad. They wanted to look advanced, imported, reliable, modern. They used a type that looks like Futura Bold to me.

This concept crept in, people felt good to use loopless terminal in 2nd and 3rd level headlines.

Today, you can see morea nd more loopless terminal designs in text.

Many people mix it up, loopless is display+headline and text at 14pt with a looped terminal only type.

When you work in thailand, you may be as a co worker, or as a family business. If you dont have a true bold you use another type that is bolder, another example.

Helvetica in the 1970s, dry transfer was very popular here, an alternate way to get fancy type in your layout. That’s where this begins. Anuparp was designed to be Helvetica’s best friend, that I made. its looped terminal, monoline, with the same proportion as helvetica. Similar color.

But when helvetica wants to speak thai, what would he do?

Its born out of being a display font, so it went over the classification line into a text font. Differnet weights help with this. This is nutrition facts on a orange juice box. Its loopless narrow type. This is a loopless terminal in an Ikea catalog.

Capturing Helvetica physically, its not possible 70 years ago. But today its more and more used for serious text usage. You can copy and paste the forms from latin that are similar to thai forms. But it doesnt work very well. You can see thai text appears much too narrow in this way.

When I designed Helvetica Thai, i wanted to move on and design new forms. If you want to design something, you must make it new, not just copy. We have to keep the essentials, but the x height is different so I redrew everything. the text color and pitch is harmonised.

I did light, regular, bold, and the 44 thai letters. I also did helvetica neue version too.

Thank you!

Posted in Knowledge | Leave a comment

ATypI 2012 Hong Kong: Day 3 Morning Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 12 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

10:45 Hanzi: the past, present and future – George Gu

Intro: Masters in Tokyo, he works with Chinese MM type design.

I studied in 1988 in Tokyo, and I finished in 1993, and moved to Canada. I designed a set of fonts, 0x-08, a MM font, and finished it in 1998. I showed my work to John Collins, D Knuth, in ATypI in Boston. After that I stopped my work but not my thinking.

Last year I started my work again. I agreed with my business partners to show my work to you today.

First, the type history of asia. Second, the thoughts about producing a typeface. Third, the practical results. In the USA, a company invents a lot of technology, like MM. You need a US computer company to make a big CJK font, but I can show you my work is fair, trustable and honest because it lets you export CJK but can work well for Latin too. I will show you my own software for designing. If a firm has this already, thats great. I am just an amateur type designer, so this may not be new. We have 4 guys, 3 with a PhD degree and me. 2 guys are Computer Science PhDs. We did lots of work together.

In 1836 Marcellin Legrand started compositing CJK with composite radicals. 20 years later this was adopted in Berlin and in 1912, a Japanese type designer biought a Lynn Boyd Benton machine which allowed them to move to large scale automated production. It was sold in China in 1950. Later Morisawa Noboo were inspired by Monotype to do phototypesetting of CJK but the scale was hard. This is lead type in 1861, 1875. A western type designer went to China and Japan and help with this work. A number of flaws can be found in the skeletons; loss of aesthetics in scaling a single origin skeleton to multiople weights (counters get closed); uneven alignment (dense characters are too tall, uneven baselines; vertical alignment from widths uneven too), disruption of normal spacing (monospace not done well), uneven strokes ().

Digital tyepface are extended by handwriting; a person doesnt change their brunch when writing but digital type often has variable stroke wdith in an unnatural way. Here is a serif and sans serif design

Here is an example of my teacher, a master of chinese art, Chu Lai Chen, 1982, its 2cm^2 and done very well by hand.

Current font tech:

1. Stroke based types by Dyna Font, in Taiwan, and they make Hanzi in Changhai. 1.4Mb fonts, can change weight.

2. Radical based type, like Paris Type and Berlin Type, is the most common technique in China, like WenQuanYi in Boston. Chienese military use this.

3. Fontographer EM function is excellent. We started this in 1995 and I asked the programmer to stop this.

Finally I chose to use Multiple Master for my font production. Range of weight by interpolation, allows range of wdith axis too, and I have a program to calculate the grey balance. We use this data to design with.


1. Smaller storage memory for JIS X 0208 WOFF (around 1.1mb)

2. Use of condensed fonts can increate number of chars displayed on a mobile screen by 130%

3. EM512 is 9 bit, 1024 is 10 bit, 2056 is 11 bit. Quality can be the same. My partner said, if a USA company chose a CJK font, we can’t match their quality.

Here is Basilisk II, hmm, doesnt work. Win XP, SoftMac. Here is my font in 3 weights. Here is a my font condensed. This is in Ikarus, and this is the grey balance output, returning a value like 0.123 for each Glyph ID. You can see this glyph has a low value, its a single stroke. We had another software on the ‘gravity point’ of a glyph. The 3 PhDs said they can see a central point but they can’t see the gravity point I can see. So we developed this tool. But I don’t have it here.

This is my iPhone mockup showing 1,155 characters on a screen. This is mine, 1,589 – a 137% increase. This images compares legibility, you can see mine is an improvement.


11:10 Designing CJK typefaces under a unified concept – Shinya Yagami

Japanese comics are popular and loved around the world. The quality and workmanship is famous, as is animation.

Why do we need many typefaces? CJK can express something with a single character. The phrase ‘Made in Japan’ reminds of things produced so far.

Issues in typography

11:35 Digging into the ATypI archive – Gerry Leonidas

This talk is dedicated to Cynthia Batty.

ATypI was started in 1957, 55 years old, and in 1984 things changed rapidly with Mac and 2 years later Fontographer. Unicode, internet in 1994 with browsers 2.0, but documents didnt change until later. No memory of online until 1997.

We see a reversal of permanence, what was on paper vanishes as its not searchable, its hidden in an office in a box. digital things can be more permanent. but what happens to reference works, the last years brittancia shifting totally to digital. its a trend.

From 1957 to 1984 there were also big changes. hotel metal to phototype. There were big players; In 1968 when i was born, there was half a pge of new type being published. You could INDEX ALL TYPE THAT EXISTED IN LATIN.

The association handled everything byt he secretariat, employed by Monotype Linotype Etc, they kept copies of all the association’s stuff, evthing was on paper. Business transitions, conference records, meeting minutes (!). It was North Western centric; evrything was in English German and French and translater, copied to 3 colored papers. Like a world wihtout mobile phones; people type something, 2 weeks it arrives, 4 weeks they get a confirmation. its all filed in BOXES AND BOXES of stuff.

AtyPI had some transitions and the archive followed through each one. People felt they should maintain it somehow, it grew and grew. I have a presinte 2001 AtypI bag. Most things are similar to today, like Sam Berlow, and there are lots of pristine empherma. SoTA partnerships. LEgal history. And the founding principles, arbitration between members.

We got the archive in 2006, it was a store house in New Jersey and sent to Reading. We can maintain things that tell us stories. We kept it in storage for 5 years until I got funding and hired a student to sort thorugh it.

If you forget your credit card numbers from 15 years ago, we have them. It was all on paper so remains. We scrapped all personal financial informaiton and tried to keep all the best things.

‘Mostly gentlemanly disussions between mostly north western men, represtingn different strages of oligolpolies in transition.’ it was a series of dinners among friends. The gala dinners is a tradition.

The interesting thing was the committess. The type deisgners, manufactorers, and educators. Education was key from the start. Typefaces were no longer tied to heavy machinery. Portability. And portability of designers. So the challange was, who maintains this, who owns it.

Copyright for typefaces?

ATypI was active in trying to design copyright for typefaces. They put together a document in 1973, to clarify what type is and how to protect it. Lots of other discussions about how own them.

Who owns what?

THere is a tension between manufactorers and type designers. ANd they wanted to educate users.

How do you educate users?

get them young, its easy to copy a floppy, and bad habits set in fast.

The minutes are often very boring. There are a lot of whitepapers explaining to outsiders what typefaces are. They tealk to the general grpahic communication community. The main companies are behind this, they wantto get their act together – they are people who talked to each other instrution to sintiution and all faced the same problem.

Most of them will not exist in their current form in a few years, there is no awareness. 5 years into DTP, they talk about funding a bank service to fund buying metal type. There is a lot of stuff like this. Initiatives to get people together. THe industry is not like what we know today. 5-6 figure funds for ‘piracy protection’ activities. £10-20k per year across a dozen companies adds up.

The ease of ephemeral publication digitally, we dont keep things that are worht keeping. Producing things on paper takes time and attneion. so people had experence and its useful to circulate. No one I know has made any copies of thise online. Its lost to the modern world. This is in 73, and in 94, its people at the top of their game, education and industry.

Walter Jungkind, largely resposnibile for seriosu type in Candada.

David Kindersley in Rrading; Gerrit Noorszji and Simon Daniels in Reading, in a room thats still used today.

Legibility is fashionable today, and theres old research in this archive thats quite different to what is done today.

What is a typeface?

That was explained to non type people a lot.

What is a PROTECTABLE typeface?

‘This is a little bit too much like mine’ without a legal environment; explaining to users was done thorughfully. 2 definitions, one was german first and translated:




Novelty was ‘the stlistic elemnts or overall appearance’ – this is not yet known to eh general public,
reasonably expected t be known by professionals.

WE cant talk about type separate from the users of type. all justments of type are in the conditions of use. familiarity by a general audience is a relevant factor.


if the distinction features of a typeface exhibit […] individual creativity exceeding […]

the technical skills and craftsmanship of a lettering draftsman.

Today, if a junior extends a character set, the art director has set the design parametrers and the junior doesn’t have much influence.

But there are hazy boundaries. What is ‘just productoin’? We have production professionals.

There is an unashamed recognition of expertise. There are people who are qualified to make judgements that are worth of attnetion. Today we avoid recognition of expertise.

WEb Fonts FORCED type desiners to open up. Browsers said they’d do it on their own and there are records of that all teh way back.

We used Omeka, a libre archival tool. Its online,


We started uploading text and images. When its more complete we’ll link it to the ATypI site. Its taggable and we want to make it available as a reference to the world, and then as content.

Jungshik: People take many photos of presentations. Why don’t we distribute slides of every presentation to attendees?

Gerry Leonidas: A list of participatnts is easy to do. We used to do that on paper. We can do that online. My view is that slides are not that useful. Theres little text in mind. You dont know what I said with the slides. I can put things on speakerdeck but you don’t know what the gist of what I said is. A paper presentation is different to a slide based presentaiton. I’m happy to put mine up. the images on my slide are not semantically tagged either, so you won’t find them if you search for phototypesetting. Taking a photo of a photo on a project is silly. Lets not forget that, as a teacher, I put my slides online, is that a substitute for teaching? The words that go with things is important.

Brian Stell: Its very valuable for us in teh audience to refer back to, though

Gerry: Yes. I will put my slides online.

Q: Why did you appear in Adams slides yesterday?

Gerry Leonidas: Oh, that reminds me! [Glimpses of Adam in various poses]

Posted in Knowledge | 1 Comment

ATypI 2012 Hong Kong: Day 2 Afternoon Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 11 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

14:00 Making CJK web fonts faster – Brian Stell

You can subset based on common character usage stats

You can subset for a given page, 1,400 chars for 99% of the pages.

A typical CCJK page only uses 5% of the characters in the Big5 set.

the GLYF table is 80-90%+ of a font’s data.

Google Web Fonts subsetting, for Latin this per-page subsetting is a big win. The text= parameter doesn’t have range support so its limited to 60 chars (?)

Page subsetting COULD be done all on the server but its not yet done anywhere.

HTML5 file system is local disk speed, its a web standard although its only implemented on Chrome. I’d like Mozilla to join in on this, I don’t know about MSIE. Can be used normally or used as a background process and then appears on 2nd time. a flash of loading at 0.5s is super annoying.

Load and use next time is great, if you know users will visit multiple times.

Lazy loading is a programmer term, you load the parts as you need them. This loads just what a page needs now, but each chunk is re-usable by all other pages, so over time the whole font will be loaded.

We do this by creating a ’empty’ version of the font, with 000s in the glyph table space, and a gzip compression will turn that from a 18mb of glyph data to just 5 bytes. I’ve started this as a libre software project:


I hope to ship working code for this in the coming summer, and it only works on Chrome now but I know Mozilla plans to implement the required parts.

I’m now working on HTML5 ‘Load and use later’ for Google services.


Thomas Phinney: Having the client merge in the fonts?

Brian Stell: Yes

Thomas Phinney: Where does the merging happen?

Brian Stell: The client is a small library, but the server needs some magic sauce.

Font production

14:25 Managing font families just got a lot easier – Adam Twardoch / Yuri Yarmola

FontLab just makes type design software.

What have we done recently? We released Fontographer for Mac OS X. It simple, robust, It Just Works.

A free update for Mac and Windows is available today on FontLab.com

We released FontLab Studio 5.3 for Mac, as a preview, 2.5x faster, 500+ bugs fixed, new ADobe FDK 2.5, many other things.

FontLab Studio 5.2 for Windows has a preview today. You can finally preview OpenType features in the metrics window.

We have a new codebase, Victoria.

Design process isn’t a big problem for designers. They want technical stuff to just work. Like font families. There’s FLS, Fontographer, DTL, Glyphs, RoboFont, FontForge. All these tools have trouble making families that just work.

There are 2 naming systems. Styling groups must be in sets of 4, then theres the extra stuff. There is complex technical information to memorise. Designers get frustrated with this, unless you are a technical expert.

Our new technology makes this easier. We looked at our Font Editors and realised a typical editor isn’t designed for families; like Photoshop isn’t made for managing photos, like Lightroom.

TransType takes a different view to fonts, it was a Mac/PC conversion tool but that use is over. So we took our Victoria tech to TransType for organizing font families in a ‘just works’ way.


Select all styles, ‘Optimal Family’ and all the width/weight information is analysed and used. Regular and Bold have 4 members, and oethers just 2, regular and italic. Thats the recommended method.

You can change this, make the black a bold of the medium. You just drag and drop them into these 2×2 squares. No more check boxes, no more tables. When you’re done you pick the target font format – we still didn’t pin them all down – and select an output folder and how to structure that output and click the ‘play’ icon to convert, and you get the fonts re-organized.

Linotype used to release 2 sets of fonts, one with style linking and one with all fonts in their own name.

There is a side panel with al the metadata inside the families.

It opens VFBs and UFOs. These UFOs have 10,000 glyphs, 50Mb of data, opens very fast.

If you open 2 fonts like an Extra Light and a Black, there is a ‘filter’ menu, and this allows us to very simply interpolate these fonts.

There is a rounding filter, so I can quickly make rounded editions.

There is a distortion effect.

I have a fake Bold and Italic tool.

It supports Overlay fonts, Chromatic Fonts, to create using the ‘overlay’ operation, and export a PDF from Chromatic ‘multi layer’ fonts.

It is not yet available, it will come out December 1st.

Yuri will now show something else.

This is a DTP page showing the font nicely. If you zoom in, you can select a glyph and make in place editing right in the page window. If you look at this, its our Dynamic Shapes tech. The serif is converted to a Dynamic Shape. The junction will stick to a stem. They are intelligent symbols.

We have better annotations.

We have a nice sticky guidelines with section counting. Our measure tool is very smart.

We really wanted to move away from many windows for different views of type. We want a single view where you can have a paragraph of text, and you can zoom in to a single glyph. Edit everything in context, in place, all the time.

14:50 GlyphWiki: a wiki-based glyph design and font production system – Taichi Kawabata / Kamichi Koichi

How to encode and display unencoded glyphs?


Anyone can create a glyph and name it, embed it, as font or image data.

270,000 glyphs made here.

You can use existing components from prior glyphs, and place them.

15:15 Glyphs and non-Latin typefaces – Georg Seifert

I’m a type designer in Berlin, here is some of my work, I did a new type for the Berlin airport, and I made my own type design tool, Glyphs App.

I had some basic ideas from the start, as Adam said, I implemented the idea that you draw in place. I heard it was 2 days of work to do VOLT arabic mark placement. I take some arabic text from a test doucment, switch to RTL mode, you can see the FEA code is generated automatically. I add anchors from the menu with default positions.

Posted in Knowledge | 1 Comment

ATypI 2012 Hong Kong: Day 2 Morning Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 11 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

11:50 Web fonts for non-Latin scripts – Toshi Omagari, Monotype

Non latin web fonts. Nassim for BBC was a breakthrough.

Web fonts are good [for the usual reasons]

Fonts.com has the most non latin web fonts

GWF Early Access is an experimental page with the widest range of scripts available.

In fonts.com web fonts or MArch 2012, 6 of top 100 were non latin

July had 10 – 4 chinese, 3 arabic – so we can see the steady growth of non latin web fonts

Here’s a list of example sites using non latin web fonts, including Wikipedia which uses Bengali and Burmese

There are simple and complex scripts. Complex scripts shape text by substiting glyphs depending on the context of other glyphs.

OpenType, wide support. AAt for Mac Safari, Mac Opera, iOS Safari.

Graphite, in Firefox as an option to enable and LibreOffice.

Last 2 are more powerful than OpenType but less widely supported.

DIN Next Devanagari, is OT and AAT. AAT works better than OT in this font.

There are unsupported scripts – Burmese only works on Mac. Syriac, Mongolian don’t work on Mac, Ethiopic on iOS, and Nastaleeq and Javanese don’t work anywhere.

Huge character sets can be an issue. CJK fonts are usually only a fe Mb, but can be up to 50mb.

Alan Tam explained yesterday that dynamic subsetting can help. Web Fonts services can help a lot with this.

I set up testing pages that are freely available to me, Fonts.com and GWF Early Access. Here you can see the results. Black text is a web font, blue text is a system font.

Some Loads and Works fine (LW), some fallback to system fonts (FB) and some load but don’t work (LD)

Some scripts can be really enhanced by web fonts because the system default may be a lot of boxes. But loading and not working may be worse than the system font.

GWF EA looks like this, its an experimental stage but here are lots of scripts available. North Korea loves Google Web Fonts! The government websites use it a lot.

Here are 2 tables showing these LW/FB/LD results and web fonts have less red cells, but Burmese may be LD. The table looks much better than I expected.

When I proposed the talk, I expected the talk to look more like this (all red with a 🙁 smiley made of yellow cells)

I found that mobile browsers support is worst, and all the fonts are improving quickly. Google web fonts are very reliable, and more reliable than those from fonts.com

Raph Levien, said at Google IO 2011, that they are working to improve the reliablity of fonts.

I really thank Google for pushing this forward and setting the bar. Its up to us to make awesome web fonts and make the web richer, not just for Latin for the whole world!

Q: Do these work in mainland china when google doesn’t work in china?

A: If a website is blocked, not sure what we can do.

12:15 Solving the challenges of Asian web fonts – Bill Davis, Monotype

Technology Adoption Curve, the top 1,000 websites measured by traffic, around 10-15% are using web fonts. For latin. Non latin and Asian web fonts have some way to go.

Web fonts are easy to use. Web fonts services really easy. Web fonts services

Technology Adoption Curve, for non latin, we are really right at the very start of this. There are 5 asian web font services, typesquare (morisawa), Fontplus.jp, Dekomoji.jp

These are in Japanese, there are some in Korea, and China has justfont.com

Why not seeing asian web font take off? Sites are very text heavy, lots of text at small sizes, where system fonts are tuned; in latin web fonts are used to replace flash with html5 or images of text for real text.

System fonts have 1,000s of hours of hinting. Web fonts tend to be autohinted or

Filesize is a challenge. But its solvable. There are now more smartphones than PCs, so it will remain an issue for a while. A 2mb file takes 61 seconds on a 256k GPRS, versus a 4G mobile with 4Mbs takes 5s.

Subsetting is crucial, and dynamic subsetting is effective and available from fonts.com

So I think THIS is the year that non latin web fonts will take off!

Posted in Knowledge | 1 Comment

ATypI 2012 Hong Kong: Day 1 Afternoon Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 10 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

CID Keyed fonts – Ken Lunde

Everything used in the workshop will be published on the Adobe CJK forum.

CID Keyed fonts are notoriously hard to work with.

They enable things like different hinting zones for difference scripts in the same font, with more than one FDArray element. Its easy to control FDarray elemenets.

The end game is making an OpenType CFF font. Almost all AFDKO tools support CID keyed fonts

Everything I show here scales to 10,000+ glyph fonts. This demo uses a Japanese font but applies to Chinese fonts too.

tx is a very useful tool. `tx -pdf fontfile` produces a glyph synopsis. Its 73 pages for this large font. Often you want to look at a subset. `tx -pdf -g /1200-/1299 fontfile` – this runs faster as its a subset 🙂

`extract-cids.pl -r -s fontfile.otf` will extract CIDs from the font in the text format needed as input to tx. `extract-gids.pl` will also work; GIDs must be contigious and it will extract eg `0-17499`.

`fsarray-check.pl` Checks the structure of the FDArray element. The output can be repurposed by copy and paste; we can put this into `tx -pdf` so you can see all the glyphs associated with that FDArray element.

[I bailed to see lectures]

Shades of grey: a look at how the brain processes typographic information – Myra Thiessen

[10 mins late]

All were right handed. So we showed letters one at a time wired up with EEG gear, awhen wht y saw 2 of the same letter in a row (regardless of typeface) they hit a button. The letters and numbers were randomised.

Fluent type were Times New Roman and ARial, and Disfluent were Edwardian SCript and Lucida Blackletter. These are in the Windows bundle. We had 2 groups of 4 letters, e c a o are rounded and at x height, and i l t f which are thin and tall. The x heights matched at 90px and reading distance was 30cm. The stimuli was on screen obvoiusly.

Results! The x on the line graph is time in miliseconds, its PRE ATTENTIONAL. Before the individual is aware they’re looking at something. The consciousness hasn’t processed it. it takes 300ms for that to happen. There is a time that the brain is aware that there is a sensation before cognition. This is shown in pink area. We can see in the left hemisphere there is a big difference between fluent and disfluent type.

There is a significant differnet in the left hemisphere between fluent and disfluent typefaces. The peak/trough in the graph means the brain is engaging more and its doing more for the disfluent type. The difference being right handed means language processing happens more in the left hemisphere.

So these results match what Diemand-Yauman’s study said, the brain is working harder to read disfluent type. But does that mean less cognitive capacity is available for higher order functions, like the gorilla in the room.


* EEG can tell us things with data about legibility and how the brain processes typographic information that behavioral tasks cant

* another point

* some other point

Typography on medicine labels: user studies on typeface, type size and spatial cueing for senior citizens in Hong Kong – Brian Kwok / Keith Tam

Brian: … 96% felt type size should increase from 12.5pt to 16pt or more, and KaiTi font was a favourite. Is 16pt really appropriate? Is KaiTi really the best? And does point size really improve legibility?

Is it possible to use 16pt on labels? Not really because the label size is fixed, the invest a lot of money in the whole labeling system. The content is legislated and must contain all that is there.

SungTi has high stroke contrast, HeiTi has no contrast, and KaiTi has a brush stroke modulation. This means the height for the same point size is quite different.

So we scaled the type to be uniform height. So we had 6 labels to test, each font at the same point size and each at the same height.

We looked at the text that the public hospitals use; there are 315 possible instructions. We looked at the longest and counted the number of strokes. We found 4 classes, 108 chars with 1-6 strokes, 200 with 7-12 strokes, 122 with 13-18, and 23 with 19-24, for 443 total characters.

We interviweed 79 educated seniors and ran 5 tests. …


HeiTi is the more preferable font, with better contrast. But it causes many errors in the reading test. The Regulation should be about font height too not just point size.

16pt can’t work with the public hospital.


Re-trans-formation of Chinese typography – Jackson Choi / Monica Chiu / Sylvia To

Jackson: In our school we learn latin typography. for learning chinese typography, our output is often not as good. we struggle to apply what we learn about latin typgoraphy. We asked students to manipulate a chinese typeface to make custom forms, since they typically fall back on system fonts.

We see chinese characters as living coenpts, not ancient relics. Type is all around us and is evolving. We want to have an experimental learning environment and personal engagement from students. How to get students engaged is tough, so we asked them to look at their chinese names and make a logotype for their name. We don’t read stroke by stroke, we read the overall character form. This make students curious about chinese type because the forms are familiar yet alien.

Monica: We came up with a modular framework for these workshops. We had 2 parts, a literature review for rational knowledge and a practical part. We asked students to break down the structure of chinese characters and draw these structures on a grid. Another workshop used art material and artifacts to explore the type forms with texture and emotion.

Students start by knowing and imitating forms and move to appreciate and express them in their own way. This leads to transforming and interpreting the forms.

We saw that our graduates win more awards after we introduced this course, which seems to show it helps. To have fun is a key aspect of learning and we achieved this.

Sylvia: A transformation workshop looked at poster design. …

Jackson: Our focus is not on creating full typefaces but to enable better chinese display typography. We asked,

What did they learn from the modules? How did this interest them in chinese typography? New areas of interest? How did this effect their design abilities and professional work? 75% said this raised their interested in chinese culture and heritage. 80% said it helped them in their jobs and 60% used the skills directly. 63% of clients or bosses said craetive display type is unique and eye catching. But 23% thought it was a useless waste of time 😉

In 2006 30% of students made lettering and in 2012 80%.

16:00 And we forgot about the time: flow, type and graphic design – Chris Ro

[5 mins late]

I was into architecture and then went into graphic design. In architecture I couldnt be invovled in the whole process. In graphic design I could be. I love the instant form making. I love thinking about having type move, playing with the forms. I think of Angus Young from AC/DC and how he is when playing, and how graphic designers are when they are working. Glum, hunched over a computer.

Flow state is a subjective thing. I interviewed a lot of designers, about flow states in design. It was more about graphic design.

Many designers said computers themselves reduced flow. They felt mastery of a computer is impossible. The immateriality of a screen, its not tactile. Designers love working with their hands. The screen is flat, its a blocking point of flow. CTRL-Z makes designers careless. I read Matthew Carter say in the old days that before CTRL-Z he would choose what is doing more carefully.

Keybaord and mouse was made for word processing, not design. Interfaces, the CS UI is a model that is 22 years old. Tool bar on the left, palettes, mouse buttons.

A funny graphic for career evolution of Motion Graphics designers. and as you progress you do less design work and more email.

I found that positive associations with flow were about typography; Keywords with type were felt to have lots of flow. Designers set up their grid, their sizes and proptions, and then laying out the content with a typeface on their design system was great and had lots of flow.

So future study, we can look at physical interactions with typgraphy and …

16:25 Creation of two original type families intended for reading contents on electronic media – Andre Baldinger / Philippe Millot / Thomas L’Excellent / Virginie Poilièvre / Christina Poth / Haruko Sumi

We were 126 students, and we made ELT Foundry. Our studio is like a science lab, we have research topics chosen by directors, 4 students involved and 6 people working 2 years on it.

ELT Gaston

We used interpolation and discussed in the group which option was best.

ELT Incision

This was designed as a pixel font first. We looked at Corbel, Verdana and Droid Sans. We did some nice pixel traps here. We chose cursive a and g for open counters. We looked at a interpolated set of weights all together and in context with regular weight, and we chose the type we all thought worked best. Similarly for the italic angle, looking at it all together, in context with regular, and also as rasterised type. Here are our tests on various devices.

We are considering releasing them at the end of the year, with a reference book, a bibliography, and a concice chronology.

17:00 From the ashes of war and oppression: Korean font development at the turn of the 20th century – Fritz Park

We are a research magazine, very new, started in April 2012, research is our main focus, we are looking at Korean typography. The history is vague. 1900-1950. Opression> Japanese occupation was 1910 to 1945, and infleunce started in 1880s. The Korean war was 1950-53, the start of the cold war.

Korean langauge was banned, hangul repressed, a ‘bantu’ education, and japanes names were enforced. The N/S war isnt over just a ceasefire.

Metal type was often melted to be reused as metal.

There was a 40+ year cultural black out. It wasn’t until 63 that rebuilding started.

A little about Hangul: there are three main elements, a dot for sky, a line for earth, and a bar for man. . _ | This is a zen appraoch to writing. Its a system for vowels. There was an inner reasoning; similar pronunciations have similar forms. Consonants have shapes similar to the shape of a mouth or in nature. Confucious philosphy was prevalent and influenced this.

Complex pronunciations have an extra stoke or are otherwise derivative from the base form. Non aspirated sounds, there is doubling of symbols.

The baseline can be hidden in Hangul. Its read left to right to bottom. It was originally meant to be read top to bottom on a page.

All the vowel and consonant structures are meant to fit into a 5×5 grid.

Korean type designers juggle the vowel and consonant shapes in each character. They are all unique. Shape and size vary. As more syllables are added, the color becomes hard to keep even.

What happened in typography in the cultural black out time?

1880-1954 is the ‘new letter era’ or ‘new printing era’ – Electrotype, we found no punchcutters from this time, so electrotype was very prevalent. Counterpunches don’t work well with the hangul shapes. All the technology was brought to Asia by missionaries and remade in Japan and reimported into Korea.

This image is from Ji Hoon Parl, Mushashino Univeristy, who helped provide may images for our research. We see that in 1880 the euro left to right reading order was introduced. Here is an 1881 with the traditional vertical reading order.

Hangul was bought and sold as type in 1896 as we can see from this type specimen. Its hard to research WHO type designers from this time were.

1954-1990 was the ‘original drawing era’ which ahs 2 parts, matrix cutting machine era in 1954 (official trade documents show their import here) but didn’t really start to be used until 1956. And photo type setting era 1970-1990.

Here are copper plate matrices used with a pantograph to etch punches. The original drawings! These are by Jung ho Choi, a famous designer from that time.

Phototype. In korea we tend to just dispose old technology, not preserve it, but we do have a phototype machine that is preserved. We can see the craft and detail that was put into them.


A lot more work needs to be done. The history of Hangul needs to be reassed, arhcived and documented.

I want to thank many professors who helped, including Ji hoon Park, Eun You Noh of Ahn Graphics, also the Text Book Museum, and many others.

Why are we doing this? Many people do restoration of old fonts. Neu Haas Grotesque is typical. In Korean, we dont have the well established history, so we dont have these revivals. This is an effort to lay down a foudation for that kind of work.

Q: What about contemporary type design?

A: We are looking at the first generation and the 2nd generation of designers. The period from 1990 is a 3rd.

– – –

Keith Tam: We covered a lot of topics today! I hope you enjoyed them all! Tomorrow we look more at the technology. Thanks to our sponsors!

Posted in Knowledge | 1 Comment

ATypI 2012 Hong Kong: Day 1 Morning Sessions

These are live blog notes from the 2012 ATypI in Hong Kong, 10 October 2012.

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the event, and may or may not be similar to what was said by the people who spoke on these topics. Probably if something here is incorrect it is because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately (dave@understandingfonts.com) – or post a comment.

09:30 In with the new: today’s font licensing landscape – Christopher Slye / Bryan Mason


09:55 Future challenges of font licensing – Ivo Gabrowitsch

[10 mins late]

Customers now offered mobile app font licenses, and they ask,

“Are mobile fonts printable?”

It may be technically possible but not allowed.

“In addition to iOS, Android, and WinPhone phones, can I use these fonts on Set Top Boxes”?

No. An additional license is needed for that use.

Desktop fonts are licensed a long time. Are there any open questions? Its 1-5 usrs, pay once and use in any print project. But then people ask, ‘Can I get a single user license, I don’t need 5’

Ebooks, people ask ‘can I embed fonts’. ePub 3.0 has font obfuscations, so its okay. For other things, users require an ‘editing/embedding’ license.

“We would liekt o know the costs to embed your OT fonts in our Kindle Fire app using HTML5″

We want to say ‘tahts what web font licenses are for’ – but while its HTML its not web, its an app. Web Fonts licenses are not for HTML its for websites, and we want to avoid web fonts being used in non-website environemnts without another license.

As we just learned, web fonts cna be made specifically for websites.

We want to move away from keeping track of each license. Customers don’t either. Its not a question of expense, its a question of too much admin work. So a big publishing house wanted many licenses, a perpetual license; they are willing to pay but not voer and over. THey want unlimited users. Most companies dont, they know how many users they have, and they just dont want to discuss licensing again. They want unlimited non-affiliate users, like print houses and consultants and so on. They want full cross media use, web, mobile apps, embedded apps. They want editable embedding. THey want all kinds of optimised formats, and permission to convert to any new format themselves.

There are people who will give them such a license for a cheap price. I don’t know what this customer ended up doing [but it wasn’t with FontShop.]

Things are getting more **complicated.** Users dont want to talk to you on a regular basis, even if you’re really friendly.

So licensing needs to become more simple. Its not price related but complexity related. If its too complex, they will walk. Eg,

‘AFter some internal discussion, we decided NOT to extend our use of FF MEta, but use Segoe UI which is a similar font and available as standard on Windows 7’

We want to give a good service and ensure our FontFont designers are rewarded with royalities. I believe this happens best with SIMPLE licenses.

So please, make your licenses more simple. I think this is win-win-win, win for users, type designers and publishers.

I look forward to talking to you all about this.

ADam Twardoch: The top floor of licensing is customer comissioning a custom font. They go away froma library to a custom font as then they can do anything they like. The pricing is to me simple. If someone wants to do all the things, they pricing and licensing terms should be competitive to commissioning a custom font. Your license can not be close to that price and simpler.

Ivo: I agree.

10:20 Crowsourced font funding: Kickstarter’s impact on type design – Thomas Phinney

Intro d

Predictor of success:

Size of your social network – nice Seth Godin quote

Low price goal

Short fund raising period with preparation work

Googel funding – Googel has funded open source kickstarters, often by 50%

Ethan Mollick on AppsBlogger, great research! Failed project really fail, only 10% reach 30%, only 3% reach 50%, and if they get over 50% its likely they will go all the way.

Delivery problems plague Kickstarter. 25% of projects are late and of those, 25% deliver after 9 months of their target. There is no guarantee the money will lead to the goal.

Success by Location graph of the USA. If you’re in NYC or Chicago, Portland, Chicago, its more likley (because of social network)

When I asked desigenrs, they said its hard to do large families because its hard to raise large amounts of cash. Natanael Gama said the STORY matters very much. The south american desigenrs all said the same phrase, ‘creating empathy for the project’.

‘Kickstarter is not a magic money machine’ – thes video took a long time to prepare.

KS gives you an admin panel with some analytics, here’s some graphs of the funding total over time.

Chatype, you can see it trickly up.

Folk, you see Google’s big 75% contribution near the start.

Montserrat hit 200% – it had a great story

Cristoforo – a victorian art neauvou. Why succeeeded on the 2nd time? My pledge levels were adjusted and the new $24 was very popular.

Stretch Goals are important to have, because it can help people contribute after you reach your goal. I put mine perhaps too close so that it created a lot more work than was justified, but having 1 or 2 helps a lot.

Lessons Learned

Scalability must be watched, if you have to package and send 200 items its a lot of work.

If you’re thinking of doing a Kickstarter for fonts, PLEASE contact me, I’d be very happy to review your project and help you optimize it.

Effects of KS? I think this will effect what kinds of typefaces get designed, because it allows for funding things that wouldn’t fly on the typical retail market.

Type and brands

11:00 Multi-script brand identity, harmonisation versus standardisation: the case for Air Inuit – Jean-Baptiste Levée

Air Inuit had old planes, an old visual ID, and the website, reservations, didn’t inspire reliability; they had newsletters and christmas sales and stuff with home DTP style design.

I started with a new logotype. I made the Inuit script harmoised with the Latin; its a plain sans serif and this was the base starting point.

Christian Schwartz told me to ask, do they really need a new typeface; it will cost a lot and take a long time. What other options did they have? There were some already available Inuit type, Nutaaq – Nunavik, the latin doens’t match the brand’s tone. Huronia, by Ross Mills at Tiro Typeworks, they do great work for Canadian Syllabics, and this is a great design. But its not for branding, its for literate and scholarship works.

Euphemia is another Tiro type, a sans with a similar tone; good character support but lacking in weights.

So it was easier to start from teh logotype and design from that. I was worried in the start that it would be Yet Another Helvetica. I wanted to steer the design away to something else. I started backwards with the bold weight first.

My method for non latin is to design both at the same time, make a few variants. The Inuit script isn’t bicameral, it has only one case, so there are weight adjustments you can see during development.

This was my first proejct with this script, so I could only trust my eye as a designer and read the documentation I could find, and thanks to Tiro for publishing a lot of that.

There is an old fashioned, romatntic thinking that you have to respect tradition in the way you engage the design; Arbaic type designers say you should mirror calligraphy. But I think Inuit is an invented script by priests 150 years ago during colonisation. It was to translate the bible and for everyday things. So for me this is a wide open field and I can do experimental things.

The shapes of the Latin and Non Latin script should influence. I avoid copy and pasting shapes. Each has its own form. You can see the latin C and a shape similar to a C in the Intuit script, there are slight variations.

I wanted the feeling to be serious overall but with little informal touches. I shopped prototypes quickly so the designers working on the graphic design could start to work with this type.

There are some shapes that produce odd shades of grey. A Latin wordspace in this script is too small. I could kern them against the latin wordspace to add space.

Before the Bold was finished I started the Light. Deadline, they needed to paint a new plane with the new branding, shown to the press and public, so numerals needed to be ready for the license plate of the plane! I learned a lot about airline companies, and if you Google these registration code you can find a lot of fun trivia.

So I can never underline the importance of a good graphic designer as a partner so type gets used very well. The designers feedback was essential; they said the light was not light enough. I had to draw a new extra light weight by hand, I couldn’t bring myself to quickly interpolate it.

Matching line lenghts is a key goal for designing multiscript type. I wouldn’t have bet that Latin was the longer script but it was!

These were my initial design glyphs – CEHIOV – and this was tricky for the Inuit script. There is a ___) and (””’ shape and one looks longer than the other when its flipped. So needs hand adjustment.

In all my type designs, I always add the little symbols like hearts, its not in the spec but its fun for me.

You know when the design works or not when the design is used. I’m a kid at heart and I love the idea of my type at 2,000pt on the side of an airplane. When this mock up arrived of the plane decal, I was very happy with the 16 hour workdays that I put into the project.

This project shipped 12 months ago. There are few options if you are a designer typesetting documents with the Inuit script. At the end of the exclusivity period, there are 80,000 readers of this script, and Google Web Fonts and Kickstarter would be good options for them.

11:25 Nokia Pure, or: how I learned to stop worrying and love the scripts
Bruno Maag / Amelie Bonet

Ted Harrison: This is a presentation about 16 scripts! A worldwide typeface before long!

Bruno: 15-16 scripts delivered so far, including Chinese, and in 20 mins we cant talk about all of them. Cheinese, Latin and Bengali.

China. Many if you know this script. Its a huge market. For Nokia as a mobile phone company, its 100M+ users, $3Bn market. THey want to be in there. The sheer size of the market, everyone wants to be there.

But China has restricitons [Great image of Mao saying “GB18030-2005”] – there is a state mandated character set that the government says your font MUST have if it ships in a computer or mobile. It must be CERTIFIED at a cost of $20,000 per font. And the process of certification takes 3 months, but the fonts must ship in 8 months.

So unusual problems there 🙂

We delivered the Nokia Pure fonts, 3 weights, 82,500 characters. Tencentype took 7 months to do these 3 styles, that is a record time for such a high quality product.

How did we do it?

Before we contracted Tencent through URW, we drew 50 glyphs in house at Dalton Maag to define the design to harmonize with our Latin. Everyone seems to respond to a design style.

Chinese friends, please dont be offended, but we felt the drawing quality was often not that great. We understand the character set scale demands compromises, but our ambition was to draw a high quality UI and print typeface. On the right, you see 3 variants with different terminals, strokes. We went from traditional to avante garde.

We then sketched 3 weights and tested it on print and screen. There can be dense symbols that become blotchy in a text block. We tested a lot of stroke weights in Regular and Bold to get a design that well harmonizes with Latin.

We also worked with various CJK type consultants to advise us. The font is for mobile devices with data limitations. We needed 27,500 character set that had a filesize no more than 3Mb.

Its a big achievement. We designed the font file with elements that should be repeated. We used TTF components to do this. This creates design compromises, that we will address over the next few months.


How did we design the Latin? 2 years ago, we worked with Nokia who needed a new brand. Their fortunes were down, they remembered what design was, and they wanted a BLAND type design, they felt Nokia Sans was designed 14 years ago by Erik Speikerman with so much personality that brand designers couldnt do anythig with it. Even the in house type designers. They wanted something that would form the foundation for brand work based on other graphic elements.

We did a lot of variation in ef the crossbar of the B or the Q tail, to add a little bit of personality back in.

We did 8 styles, 3 for UI and 5 for display, ExtraLight to ExtraBold, and the UI had increased letter spacing and some design tweaks to be better in a UI context.

Type makes you money!

Clients too, not just type designers. In March 2011, Nokia threw a big party for the typeface lauch, and an indepdenedet marketing study said the hype created about the nokia type craeted €1,000,000 of value.

That means we type designers should charge more than $1,000 – don’t sell your font to Google for $1,000!

Amelie; Bengali is a phonetic script, this has Bengali Devanagari and Latin on a street sign.

To create a sound, you have a few component glyphs for the component founds, and they conjoin into the final glyph.

Bruno: Look at it! They are on drugs!

Amelie: Fiona Ross and others helped consult for us to work out the linguistic requirements. There are 500 glyphs in the font. Its not as much as CJK, and a lot is built on the fly, but everything is decomposed outline; the position is mark positioning.

Here is Nokia Pure compared to Benda (?) a system font. Benga is a good systtem font, but we wanted to have a slightly more calligrpahic stroke. We built a lot of vowel arms of various lengths. We avoided collissions of strokes above the body of the type which we can see in Benda.

It was hinted with VTT.

We are Devanagari, Bengali, Tamil, Sinhalese, Kannada, and now we are working on Gujarati, Telegu and Oriya. A lot of work is coming up!

Bruno: its not just Amile and I doing this. Its a huge team, here is the team. Supporting such a team means you can’t sell them for $1000 to GOogle and make them free fonts. Type desigenrs have to charge reasonable prices. Tom Phineey said $5,700 could be raised. That won’t pay rent for 2 motnhs! I am very upset with this stuff, because it undermines the value of the work we do. A doctor won’t treat you, doing a social favor. Type designers study for years, we invest in design adn technology, we train people. That has to be paid for. This huge team needs a fair wage, and clients needs to pay me a fair price for the quality we delivery. Thank you very much!

11:50 How web-based fonts and dynamic subsetting improve brand performance in East Asian writing systems – Alan Tam

Web fonts!

CJK is much more complex than Latin especially for web fonts. Andy Payne, Interbrand Chief Global Creative Officer “Brand starts with the name, and the name is written.”

The digital medium can be a barrier for brands. Education, understanding how things work. It shouldn’t be feared, brands should extend and meet there customers. You can engage and enhance your customer relations, reaching many more people online than in the print world.

Fonts are at the core of a brand. So having the same type in print and on the web is important.

When you extend a brand globally, you recognise the brand by the type alone. Cocoa Cola, GOodyear Tyre, Cathay Pacific. Their message, their logo, brings trust.

Web fonts help reduce steps needed to bring your brand online. This makes globalisation local. Cosmopolitan cities like HK, Tokyo, Seoul, people wish to consume content in their language. Web fonts allow brands to do that.

East Asian writing systems have large fielsize. 6-50Mb! 10-30k characters. The average Big5 font is 13,000 fonts with 8Mb, and Unicode 6 has 109,449 characters can be be up to 50Mb.

MYuenHKSemiBold has 11Mb WOFF, 33Mb SVG, 11Mb TTF, 11Mb EOT. That adds up!

Mobile devices really can’t deal with that for bandwdith, memory or disk space.

Subsetting can help. That delivers only the characters used for the displayed content.

Raph Levien has blogged for Google WEb Fonts their text= parameter, and the example ‘MyText’. This is ‘Pre Subsetting’

Dynamic Subsetting looks at the content and extracts the glyphs needed. This is available on fonts.com and uses a JS file that is part of the HTML/CSS/JS page.

What are the tradeoffs? Both give a smaller font file size.

Web fonts helps a lot with reinfocing brand indentity and this can even have security implications for helping with phishing.

Before web fonts, all this was done with static images. Its a huge load on designers time, bandwith, and its also a poor visual experience for users.

On the web, speed is everything!

Fonts can be so much faster than images anyway, but subsetting takes it to a new level.

Different screens and devices means having a consistent customer experience is important and web fonts makes it happen effortlessly but images are a lot of work. Its future proof about new devices and screens. Images will have to be redone in the future.

Content is searchable. Theres a lot of workarounds you can do for static images or even Flash workarounds that don’t work with this. They use a lot of static images of text and this ruined their search ranking. SEO would have cost a lot with that technology. But switching to web fonts, they got all 3,000 pages properly indexed and searchable very fast.

Accessibility is important, and in the EU and USA there are accessible compliance laws, and there have been lawsuits against JetBlue and Disney, over these laws. Web fonts is great for accessiblity.

**You get up to 90% savings on developer time. IU-HQ CEO Frank Lampen said “we estimate that ethere ia time saving (and therefore a cost saving) in excess of 90% by using web fonts over static images.”**

I’d like you all to try out seubsetting. Monotype made a demo site:


This is a great way to experiment and find out how subsetting works. I can select a number of fonts and a text and see the file size reduction possible with subsetting. The full font on a 3G network took 3 minutes to load, but the subset was under 1 second.

Who is using Font Subsetting today? Many top brands are:


Typography and reading

12:15 Reading expertise: what does it have to do with typefaces – Mary Dyson / Keith Tam / Brian Kwok

Mary Dyson: I’ve been interested for a while in how we read and how we process different typefaces. I’m interested in perceptual abilities of type designers and typographers differ from normal users. I’ve also added some CJK type.

Before I start, I should say this is about reading. Its not about the expertise at the University of Reading.

Clare Leake is the 4th person on this paper.

What is font utning. How we can look at so many different forms to understand the abstract Platonic Form. This has been researched by Sanocki. Psychologists think of typeface differences as noise, being things to get rid off to identify letters. But they might be the opposite?

Regularity is the similarity of letters within a word. Same font for each letter in a word, you are more accurate in teling the difference between 2 strings of letters. Tested with Verdana and Bodoni, Garamond and Bodoni.

Congruency is also a relevant concept. If a type style is consistent with the meaning of a word, we are faster to recognise the word.

Reading expertise, just means being able to read the language.

My research qyestiosn: Can type designers demonstrate font tuning in a script they can not read? We found people who couldn’t read Chinese and people who could (bilingual), and readers of chinese who were/not typographers.

Font Tuning may occur with meaningless shapes, when people have typograhic expertise. Typographers can separate the stylistic aspects from teh character’s essential shape. If we dont have that expertise, those 2 thigns are bound together.

THomas Phinney Q: How could this influence how we design typefaces?

Mary: Not really, I think designing scripts where we cant read, I threw this to an ex student, and they said its helpful if you can’t read the language. I’m not a type designer, you can answer that 🙂 It seems helpful to me to know how you percieve things, that you DO perceive things differently to those who do have that expertise.

Posted in Knowledge | 1 Comment