LiveBlog of ATypI 2011 Day Two

Usual disclaimer for live blogging: These are informal notes taken by Dave Crossland at the ATypI 2011, and may or may not be similar to what was said by the people who spoke at ATypI on these topics. Probably if something here is incorrect its because Dave mistyped it or misunderstood, and if anyone wants corrections, they should email him immediately – or post a comment!

Also, Dave apologises for scribbling these notes in ascii, when many of the talks today are critiquing the reduction of subaltern characters to angloamerican technoimperialist ascii πŸ˜‰

Frank E Blokland: The Mythical Eye

Fournier said that the eye is key.

What is type design? Allen Hutt in his book on Fournier 1972 said being a type designer is ‘an indefinable quality’ which is silly

Lecturers, you know when start teaching that new students dont see what to see, and you teach them to see the way you see; conditioning.

Typography is formally representing hand written letters on the surface, but calligraphers can do a lot, avoiding collisions, as everything is custom. Typography, everything is a bound by a rectangle, its totally different. A calligrapher doesnt care about the space, he doesn’t have sidebearings. Where is the sidebearing of the C?

So its possible to have th elette

Could it be that these guys were even more clever? How were the rectangles decided? the x height is a key measure, adhtn the body. A system for small and large point sizes, and for

The ‘m’ has an ‘Em square’ that follows the golden relation. there’s also a golden relation between the x height and the jenson descender.

jenson worked in mainz for a perioid and made texturas. this is from gutenbergs 42 line bible, and the same system is used, with a ‘root 2 rectangle’

Font formats today are for all scripts, and these were very smart guys, who invented movable type, so i would guess they would have a sophisticated system.

Garamond’s Paragon Romaine (used by Slimbach for Adobe Garamond) and this relates the m to the ascenders and desceners of h, p, and cap height of H too!

Here’s an old display type with very long ascenders and descenders and huge caps. the paragon romaine is conditioned to be more familiar for text type.

Here’s a display type by van der keere, and the body size is related to the ‘E’ but it seems he made a mistake in the system πŸ˜‰

You can vary the propotions and see how the system proportionally adjusts everything else.

I looked for how the caps could be related to the lowercase, and the ‘m’ fencepost can be used to see the H and N are matched to the m outer stems, and this is too wide for us, and Griffo condensed these letters.

I looked at the original prints, not Slimbach’s revival, and original metal types, and you can see the height of the O and the width of the ‘m’ are equal.

In gutenburgs textura, i could see a spacing pattern that is very simple, the downstroke. iif jenson knew this system, would it apply to roman type? i think so. the h n u m d p q all fit the same width. this makes casting easy! on a 30″ screen, then ‘ni’ and ‘nl’ means you can adjust the ‘nl’, but the old guys were doing it very small, and in metal, they couldnt be so fine.

Jenson, and Griffo, these letters have the same rhythm which is no suprise. Van Krimpen’s Haarlema from 1938, using an existing unit arangement system. i think he drew the letters to fit an ‘m’ rhtym.

I divided the ‘m’ into 8 units, and each unit is the stem width. the sidebearings fit into that grid, and are almost the same as the adobe jenson!

If there is a ‘natural’ rhym to a type, can you blindly do the spacing? I did this with Adobe Jenson, and there’s a set of 4 images of cadence units.

if the letters are morphologically related, it seems this works!

I looked at Garamond’s de grow (?) romaine type original matrices, and the proportions for the display type is the same when scaled digitally from the paragone romaine.

You can see this 16th type and te b d g o n u o are all the same width, and this later type has many differnet widths.

these are granjon matrices, its the cursive of ITC Galliard but matthew carter. i tested this at plantin, casting their original matrices and the casting was so easy and simple! as soon as we had one wdith, we could cast many letters. we used an ‘o’ to set the overall width.

I think this system became forgotten by the later 17th century, and we found in the plantin museum ‘letter de la justification’ from te 18th c – but before that, you know if the ‘n’ sets the spacing on a system and he knoew the system, you didnt need such documentation.

I will post this informatin on typohile.

You know moxon’s mechanik exercises. this is thought of as the first time a grid was applied to type. this grid is similar to the one i found in jensons type. i think moxon did the same, he didn’t draw the grid though, just placed numbers. so i dont think the romaine du roi is the first gridded type; moxon was not an expert, he got his knowledge from all over.

Monotype Bembo is spaced this way too.

Swift, the ‘e’ in my system has too little space on the rigth, but its very close.

Documenta: my own spacing, my cadence spacing from stems, and the ‘kernus’ autospacing that looks at the white space around the letters

I think if you define the systems in a writing system, that is what type design __is__

‘Conditioning is based on conventions and subsequently …’

Q: Are there measuing tools from the 16th C?

A: No, there’s only the pacioli measuring of roman captials before the romaine du roi. my research seeks to get proof from the actual material, there is no written support. a historian asked me, what was the need for the punchcutters to make a system? I think its about quality, to control the typographic quality. its so simple, everyone knew it. its a tricky point of the research.


Here’s an email from a friend, “im abroad….” – Thorn becomes th, -D becomes D, AE becomes AE, U E lose their accents and O looses Umlout.

The IcelandAir inflight magazine does a similar thing, they got the accessnts but los the TH D and AE. This is their policy! They have a foreign audience, and perhaps its done this way to not frighten them πŸ˜‰

In the last few months, the ‘reduction algorithm’ in their database and search strings of Amazon and Biblio has got a lot better. so if you type ‘AE’ you will get hits for both ‘AE’ and ‘A E’ – and this is really good for Icelandic booksellers!

In 1988, license plates from one area stopped being started with a thorn, but you can still get these chars on your vanity plates. Will they be able to use number plate recognition across the EU with these chars? πŸ˜‰

Drupal doesn’t see thorn and -D as lowercase alphabet chars, and rejcets them. Its a Uncide support issue πŸ™

Many icelandic people habitually reduce icelandic chars in file names. uncide support isnt robust across all contexts.

SMS is similarly usually reduced.

There are 2 visions of the future, one where ascii style reduction is dominant and repalcing -D with D and thorn with TH, and one where unicode is made more robustly supported and peple are educated to use their ethic writing more often.

Fred Smeijers: Life after Counterpunch

Research since 1996

We printed 6,000 copies in 1996, and I thouht we’d never get rid of them. 4 years ago we did, and needed to print another edition.

I wondered to expand the book with new research, or keep it as it is. Its slightly revised but essentially the same book.

Counterpunch 2 is coming, as a complement to the original. I don’t know why the book is alive and still interesting for type designers. Perhaps because it explains basics without extra fat. It has plain language, it can be read in one evening, and its a book with silly ideas πŸ™‚ I don’t take them back yet πŸ˜‰

What happened since? I conitnued, I looked at matrices too. There are big ones, odd ones with air channels, neat ones or strange ones. With matrices we can find punches that were lost.

If you look at matrices, you need to clean them and the photograph them. Is it in relief?

This is a counter. Huge. deper than is technically needed.

you can also use ‘digging’ rather than a counterpunch.

There are ‘shoulder’ punches that allow metal ‘composite’ chars πŸ™‚

roman is easy, italic is more difficult, i think greek is tough. thank god i wasnt born in 1520. 24 matrices for the alpha’s accents. the alphas are the same? no, we can see their forms are different. they were willing to use shoulders one at a time but not two.

all accented matrices do not differ in width compared to teir parent within a margin of 0.15mm. this is important as it sets a clear reference.

Justin Howes thought I should take up casting. Type Spaces by Burnhill is an interesting book. He looked at manutius type, divided em square into 12 units. Justin said, If you have an em square of 4mm, and divided by 12, you can’t cast something so thin by hand!

I’ve been casting in 2006-2008, hundreds of times πŸ™‚

I can go into a print shop in antwerp from 50 years later than manutius, and its a mess πŸ™‚ The Petite Nomparelie … from 15XX was the first 5-6pt type. we can look at the print, the type and the matrices.

The proportions of a lettes wdith, height and counter are defined by the width of the main stroke.

Michael Everson: A ΓΎorn by any other name

Thorn is from a rune. Its in the text of Beowulf, line 3,059.

Those of you old enough to remember and appreciate todays tech, the font on a typewriter was the only ane available. I would write a p, backspace and write a b, for a thorn, and draw in a -d with an x over it or by hand πŸ™‚

You can could attachments for Smith-Corona typewriters.

Last year I asked, what makes Gealic type Gaelic?

Eventually I moved to a Mac Roman computer, and then ISO Latin-1 computers ,and there was a Mac Icelandic codepage, and then in 1991 Unicode made ASCII and Latin-1 as it first 2 things. Turkey put pressure on for Latin-5 to come before Latin-1, arguing that there were many turks and fewer icelandics. But Latin-1 prevailed as the first Uncide table.

European standard was drawn up for a sorting order for multilingual data, EN13710, and in iceland thorn is after Z, but in danish, it comes between T and U, and also for english.

The EU standardistas met in Iceland to decide. There were Danish and Icelandic actors in the conflict, and me πŸ˜‰ My paper was presented in 1994, and thorn was sorted in the icelandic way.

Its now supported in OSs.

‘A Concise Dictionary of Middle English’ from 19th C used the danish sorting. I remade the dictionary fo the web and used moern conventions πŸ™‚

On an iPhone, the Danish locale sorts it in the Danish way.

I filed a bug with Apple to replace he Β£ with € on the numeric keyboard layout. The IRISH keybaord on the iphone includes thorn as an option of T (but not the UK or US keyboards)

I added 4 more thorns, with strokes in the acedner or decender and cap or lowercase, from medieval writing.

Could there be more? A german linguist made a thorn with a back loop from the downstroke into the bowl, and anodhter style of thorn.

Historically the thorn was reduced to ‘y’ as in ‘ye’ like in Lewis Caroll’s poetn of ‘ye carpette knyghte’ – but there was never a ‘ye’! it was always ‘the’ but written with a thorn that got reduced πŸ™‚

The Wycliffe bible from the 1300s got reduced, replacing the thorns with TH, and this occurs so frequently that thorns can save trees πŸ™‚ There’s 104 pages of ‘th’ in the bible, and its only 72 pages of thorns. a major saving! πŸ™‚

www.ΓΎ πŸ™‚

Ralf Herrmann, Nadine Roßa: The German letter ß

@nadrosia and @opentype

Nadine: German has its special character too, ß

Type Designers often have their favourites, the ‘g’ or ‘a’ or “Q” and I like the ß (set) (or sharp S from unicode name)

Its in my surname, which makes a lot of trouble πŸ™‚ its often reduced to B.

Its rarely used, 0.31% combination (e is 17%) in german dictionary. reduces to ss. internationally its unknown, and within Germany it has an image problem.

it was used in parallel with ‘s’ for a long time, and used for ‘sz’ although the s was often a ‘long s’ that looks more like an ‘f’

there was a competition in 1879 to design the standard form, and the salszberg design won

there isn’t really a capital ß form. people find their own way to deal with this, like adding a descender to a ‘B’

there are lots of names for the char, ‘eszett’ is common, ‘sharp s’ or ‘rucksack s’ and even ‘german B’ πŸ˜‰

Ralf: so for over 100 years, a capital ß has been proposed. its still controversial. Latin has a principle that the number of lowercase = uppercase, but in Germnany and Austria there are 26, umlouts with a o u, and then the ß that lacks a capital.

many people are habituated to this and dont think of it. there are some funny tweets about this from @letterror, @kioskfonts and @kupfers and @espeikermann said “how can a lowercase ligature become a capital letter?”

hte idea is that there are characters and ligatures. ‘fish’ and ‘FISH’, the ligature isnt used in caps as it makes no sense, so ‘Straße’ becomes STRASSE. thats a fine argument.

but this is only true if the ß is really a ligature like ‘fi’ and ‘fl’ – but its not used like a ligature.

German orthography isn’t like this. Here’s a Palatino specimen page showing it as its own char.

when blackletter was used commonly for writing German, there were may common ligatures, ff ch and sz (long s, z) but then when German was written in latin, the long s was lost, but the ß was kept.

Tje orthographi reform 1996-2006 made new rules for writing german.

there are now ligatures that became characters, like w, oe and ae; those are pronounced differently than ‘o e’ – so if you think ‘w’ is a true letter in the alphabet, so is ß

it doesnt matter about the history, its about what contemporary orthography assigns to the visual symbol

Eg, there are 3 wrods, Busen, Bussen, and Bußen (meaning bosom, buses and monetary) and they are pronouced differently.

But setting these words in all caps, they are BUSEN, BUSSEN and BUSSEN – so the context determines which word is being used.

The official replacement rules to conveft ß to SS aren’t round tripping. If I get a name in uppercase, I have no idea how to spell the name correctly.

A german dictionary in 1919 said ‘the use of two letters for one phoneme is just an interim solution and should be stopped as soon as an uppercase ß is made’ πŸ˜‰

Many type designers have made such designs, 120+ fonts in MyFonts have it.

120 years ago, German lacked capital umlouts, and it was the same story; theres a google book n gram search that shows the ‘Ue’ replaced by the :U

@twardoch said that ‘were in a transition period, people will adopt ß more and more because ITS THE ONLY LOGICAL THING TO DO’

So I encourage you to put this in your fonts and let the German speaking people decide if it has a real use πŸ™‚

Everson: to ID people in databases, you need to spell their name right or you will arrest the wrong people πŸ˜‰

Gudmundur Oddur Magnusson: Driftwood

What is Icelandic?

Ideas are not original, solutions are.

I teach at Icelandic Academy of the Arts.

Fiona Ross: Newspaper Non Latin Exhibition

This is Linotype donated material.

Relevance, significance and reliability.

This is a unique collection

The drawings of these text types, from the last 2 decades of 20th C, its 90% of the type in these cultures.

Reliability? Its biased toward LInotype, and it shows they kept informatin from problematic projects, the easy ones had less documentation, so it seems harder than it is.

its useful to have hydribs and interpretations. this is an experiment with phtographic negatives …

Alice SAvoye: I’m doing a PhD and I’m looking to document how design processes changed in the move from metal type to film setting. i spent time doing archive research, the collections in Reading of non latin type, you can see the effect of technology changes. you can hardly see this over time anywhere else. theres a diversity of material there. we can understand how designers worked day to day when the adapted type from hot metal to photo to digital.

Theres 2 bengali newspapers, one with metal with breaks between chars and the matrices couldnt always be aligned, and this made little jumps. whereas digital setting is smoothly aligned. this can be observed on the printed page, and the darwings and correspondance … what you can observe from teh printed page is reflected in the correposndence about hte desicn decisons. we dont show the correposdance here as its not as visually fun as the drwaings, but its important to credit the right people who did the work, and see the collaborative aspect of the work.

Fiona: There are drawings of devanagri chars by matthew carter and you see a revision. someone could misinterpret the meaning of the revision.

Sarah Owens: 12pt Times New Roman

Sarah asked me to remove the detailed notes about this presentation. You should contact her if you want to know the details; the following is a review with my own views of her presentation.

Sarah has been observing lay designers working practices and doing interviews about how they see themselves and what they are doing.

Popular computing can redistribute power, and I believe the commoditisation of design by word processors is as fine an example of this as the commoditisation of finance by spreadsheets. Sarah suggested that commoditisation of design has led experts to shift to generating templates for non-experts and assisting in their use, although her examples often featured people not using templates but manifesting their own visions.

Sarah noted how their design artifacts are often demonised or idolised, and felt this can be too simple. I agree with that! Sarah did interviews to understand how they are designed and why they look as they do in detail. She found people using apps they already had available (Paint, Word) and I was sad to hear how people were not curious about using more sophisticated tools. I hope that as libre licensed apps like Inkscape become more widespread, and video tutorials are YouTubed, this will change.

Sarah noted that active design choices are being made (type choice, layout, and an iterative process to overcome the imprecision of word processors’ page previews) and so her conclusion was that lay graphic design isn’t “The Other” of professional design and is worth examining to understand design as a fundamental activity.

Overall it was a clear presentation, and the photos used to illustrate the talk were great – my favorite was a A4 sign with very large all caps posted in a door window, saying ‘push button to open’, and placed above the generic, professional sign that was too small to command attention. I hope Sarah’s research will be publicly available some day πŸ™‚

This entry was posted in Knowledge. Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.

Post a Comment

Your email is never published nor shared. Required fields are marked *

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    Markdown is turned off in code blocks:
     [This is not a link](

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>


  • What we do

    Understanding Fonts is a type design training business. If you'd like an event in your college or city, let Dave know: