Book on how to design type

[AngelFlonne]

This topic was imported from the Typophile platform

Hi there, first posting here, can anyone suggest a book that will teach me the anatomy of letterforms and type design. I'm a graphic designer of about 10 years standing but have never designed my own face. I did Fine Art at college rather than Graphics and a friend was telling me that on his graphics course they redrew Gill Sans and learnt the intricasies of each letter shape, that's the level of detail I'd like. I'm looking for something very hands on, telling me why different letters are particular widths, relationships between shapes etc. not anything theoretical, historical or overview.

Any suggestions? Many thanks.

[hrant]

There's nothing really very useful, but I'd recommend Doyald Young's "Fonts & Logos" for some great detail, and Walter Tracy's "Letters of Credit" for a broader view.

hhp

[eomine]

There are some great tips at Briem's website.

[beejay]

Leslie Cabarga's book was scheduled for Fall 2003.

http://www.flashfonts.com/flashsite/entries.html

[aquatoad]

This is not a book, nor authoritative wisdom, but...

I find myself saying the same thing very often over in the critique section. So here is my list of some of the top mistakes beginners make. (re: Beginners. I designed my first font less than two years ago, so I'm no expert, but I have spent a rediculous amount of time learning about and drawing letters since then. You will find the learning curve is steep.)

Structural Problems:
1. m != n != r != u (does not not equal)
2. Stoke modulation problems. I recommend looking at Bodoni. Trace over it with a chisel tipped pen. Notice where it is thick and where it is thin.
3. The Trapazoid Effect (scroll down to diagrams)
4. Width problems. The best advice I know is to look at a lot of type.

Optical Problems:
5. The mathematical center != the optical center (it's too low!) Trust your eyes not your ruler.
6. Two lines of equal thickness one vertical, one horizontal. For them to look the same, the horizontal one will need to be thinned a little bit.
7. The X Ordeal (scroll down to diagrams)

Rather than a book take a class from somebody really good. Surely there must be something happening in london? (Help people, please!). The only resource I know of is Reading. Do they offer weekend workshops?

Finally, hang around typophile. Post your work in the critique section! Dont be afraid, thats why its there. Email people privately too. Most everyone is eager to lend a hand. The people at typophile have been so helpful to me as I learn my way around the alphabet.

The water is warm. Enjoy!
Randy

[beejay]

I missed the trapezoid effect and x ordeal
when you originally posted that ... gotta go
back and look at the Xs in some stuff i'm working on.
great tips.

[AngelFlonne]

Muchos gracias everyone, there's a wealth of information here. The books look great and in fact the info at Briem's site and in Randy's post is a great start so I figure the best thing to do is get on with it then come back for feedback from you once I've something to show.

Be a couple of weeks or so but "I'll be back". Thanks again and I've already taken your advice and joined the mailing list.

[aquatoad]

Edwin before you go, here's another online resource: Underware's Workshop Basics

Randy

[AngelFlonne]

Oh thanks Randy, it's exactly the kind of info I want. I'm going to disseminate it over the coming days and try to take as much as possible on board. I hope this thread is useful for others in my position, sure it will be. Catch you all later this month hopefully.

[hrant]

Edwin, just be aware of the difference between making "calligraphic" (more properly called "chirographic") type (like Zapf or Majoor*) versus making "synthetic" (Bilak's term) type (like Carter or Licko). You might end up favoring the former, but I'd recommend the latter. Anyway, it's never to early to realize the distinction - probably the first thing a type designer should learn, in fact.

hhp

[AngelFlonne]

Thanks Hrant, comments noted, and sorry all for the multiple postings above, my browser hung when I was trying to submit and I must have replied repeatedly

[setmajer]

quoth Hrant:
Edwin, just be aware of the difference between making "calligraphic" (more properly called "chirographic") type (like Zapf or Majoor*) versus making "synthetic" (Bilak's term) type (like Carter or Licko).

Could you elaborate on this for the benefit of the lurking ignorami? Or do you have a link that would do so? I expect the difference is that chirographic types explicitly build upon 'written' letterforms and therefore show evidence of the pen, whereas synthetic types do not. What I am not sure of, however, is why you feel the distinction is so significant. I do understand that if one wants to work in one or the other style one needs to be aware of the characteristics thereof, and likewise understanding how the techniques/technology of writing impacted the forms of letters is valuable. However, I'm not sure I understand why you draw a bright line between them.

quoth Hrant:
You might end up favoring the former, but I'd recommend the latter.

Are synthetic types easier to draw? Is this a matter of taste? Or something else I've missed completely?

[hrant]

Chirographic type is type that is somehow influenced (usually by structure, but sometimes also of finish - think Ex Ponto) by the arm/hand/broadnib_pen system.

I don't know of any resource (online or otherwise) that explains the opposition between chirographic and synthetic type. Peter Bilak's Fedra specimen has some stuff. There are resources that elaborate on chirography, but they generally avoid an analytical treatment of the functionality of typography (with good reason). When you pin down a chirographer, he ends up invoking WRBWWRM as an escape clause for functionality. Which is hilarious because chirographers disdain the Emigre/PoMo school.

Is the distinction significant? Is type significant? To most of the world, not a bit, so certainly chirography shmirography isn't going to affect the world very much. But we do have to focus on the context of this way in which we've chosen to spend our time, and if we mean to help type progress (instead of vegetate, or even regress), then the distinction is central.

Which way of type design is "easier"? That's a very rich question. It depends on what kind of person you are. If you feel that control over you own creative process is the primary concern, then chirography is very attractive: it lays downs easy rules that you can follow cheerfully. But if you a) are worried more about the readability at the user's end, and b) are willing to realize that total control is illusional anyway, then chirography is nothing but an artificial limitation.

In synthetic type, the rules you need to build (in order to make "good design" easier) start form elsewhere. So I would say that chirographic type is easier if you master its rules, while synthetic type is easier if you just don't bother so much with rules (PoMo). In that pairing the former is certainly better. But chirographic type is never very hard, while synthetic type can be surreally hard (by accounting for the true nature of reading), and that's the "real stuff" as far as I'm concerned.

--

Let me try to give a concrete example:
In chirography, the bowls of the "b/d/p/q" generally have to be the same way (not identical, I mean structured the same way). In synthetic type, they don't - and this creates a problem: what do they need to be? How to avoid arbitrary results? The answer is:
1) Don't have such a phobia of arbitrariness - the control you have is proportional to the control you relinquish. This can be psychologically very hard.
2) Work backwards from the process of reading. This is technically very hard.

You can see why chirography is so popular: it strokes your ego, and it reduces your analytical load. You can be a free, happy artist, in your own little world of rules. Selfish.

hhp

[setmajer]

Wow. Thanks Hrant.

If I understand you correctly, your preference for synthetic types is based on a rejection of what you perceive to be arbitrary rules derived from the characteristics of now-outmoded (in the context of lengthy, mass-produced texts) technologies in favor of new rules derived solely from the function of the type.

Is that a reasonable summary? If so, what about limitations imposed by current technology? Are they to be dealt with on a type-by-type basis depending on the intended use of a given design? Or, insofar as most all designs nowadays are executed with beziers drawn on a computer, are there more general rules for legibility/readability that are a necessary consequence of that process? As well, are other old technologies still exerting undue influence on type design (punchcutting, hot metal, phototypesetting, etc.)?

[William Berkson]

How much should the pen influence digital letter forms in text type?

Letterletter by Gerrit Noordzij is on the side of strong pen influence. Counterpunch by Fred Smeijers argues that though the starting point was calligraphy, even old style serifs are to a significant extent constructed, not pen drawn. Hrant, as you see, is militantly opposed to the influence of the pen.

To me, and I am much less knowledgeable than any of the above three, Smeijers' view rings true. On one hand, it is very hard to draw a good looking two story a or g that doesn't show the influence of the pen. And when you do an upper case M, if you don't let your weights be influenced by the broad-nibbed pen, the result tends to look unbalanced. On the other hand, Smeijers makes the case effectively that the look of Garamond and even Jenson are decisively influenced by construction of the punch, not the pen.

Also an important point is what Eric Spiekerman said: that text faces can only vary about 5% and still be readable and good. To me this means that you can be too pen influenced or get too far from the pen.

So to me the dichotomy that Hrant is setting up is misleading. It is rather that there is a sliding scale between pen-written and mechanical, and most successful faces, and probably all successful text faces, are somewhere in the middle range.

In display faces you can get away with very pen drawn scripts or very mechanical sans.

[William Berkson]

Hmmm. I probably should have written 'constructed' rather than 'mechanical' above - this is more accurate.

[hrant]

> a rejection of what you perceive to be arbitrary rules derived from
> the characteristics of now-outmoded (in the context of lengthy,
> mass-produced texts) technologies in favor of new rules
> derived solely from the function of the type.

Couldn't have said it better myself!
And really, it was "outmoded" the second Gutenberg did his thing.

As for technology, in any form it always exerts "undue influence" - it forces deviations from the [perceived] ideal. But that's life, so we have to work with it (like by using TT hinting to make outline fonts look nice onscreen). However, one thing that we have to mind is potential cultural damage where it does more harm than good, such as the desecration of Arabic during its application to Linotype.

--

Smeijers and GN are ideologically very close (in the context of the whole)*. In a broader sense even Crouwel and GN are pretty close: they both apply superficial visual construction rules to something that must be derived from vision/perception/cognition instead.

* GN doesn't say type has always been totally influelnced by the pen in every way, he just says it should be, in most ways. To me Smeijers essentially says the same thing. And neither of them make room for conscious improvement of communication.

And I'm not "militantly" opposed to chirography (have you ever met a militant?). I'm willing to let people choose their own way in life (even if it's a waste of potential), and I'm even on good personal terms with some strongly chirographic type designers (as long they're not so insecure as to shun me for my typographic views). Not many of them though, because artistes can't handle criticism very well.

> the result tends to look unbalanced

To whom? Type designers (and users) who have been conditioned to see things in a certain way, sure. What do we really know about what readers need? This actually reflects on the difference between display versus text type, between deliberation versus immersion - that whole network of parallels.

> text faces can only vary about 5% and still be readable and good.

What exactly is this benchmark, from which we need to measure?
This doesn't work.

> probably all successful text faces, are somewhere in the middle range.

It is tempting to see sliding scale (because it gives you control), but it's an illusion.
And what exactly is "successful"?

BTW, "synthetic" is a much better term than "constructed", because the latter implies ruler & compass stuff and/or making type by putting pieces together, neither of which is required (or even encouraged) in synthetic type.

hhp

[William Berkson]

Yes, I agree, 'synthetic' is better.

My basic difference here with you, I think, is that I don't think it is likely that we will ever greatly improve on the readability of a well designed and set and printed page of Garamond or Baskerville. While I think the readibility stuff is very interesting and an important direction for improvement, I don't think it in the end is going to make a huge difference.

But I could be wrong, and would be happy to be proven so.

[hrant]

> I don't think it is likely that we will ever greatly improve on the
> readability of a well designed and set and printed page of
> Garamond or Baskerville.

1) What exactly is the basis for this opinion?
2) What is "greatly"?
3) How great does the improvement have to be for it to be worth the while of type designers to push their beziers in a certain way, which costs nothing but awareness?

> I don't think it in the end is going to make a huge difference.

Well, it certainly can't make a huge difference for most of the world's population.
Nothing related to type can.

hhp

[hrant]

Also: Shouldn't we use new technologies to improve functionality?
Typography no longer has to be static and linear.

hhp

[William Berkson]

>1) What exactly is the basis for this opinion?

Well, I am considering first that people had a very long time with the alphabet and calligraphy to try to make letter forms easily readable. And after all the end test is the person reading, which hasn't changed. And some of the scribes were, very, very good.

The second thing is that I don't see much change in readibility over thousands of years. If you look at a page of the Torah scroll, which has looked the same for probably 2,500 years or so - since they switch to the pen drawn Aramaic script - it is pretty much just as readable as anything today.

The third thing is that changes from the early model of Garamond haven't improved readability, but if anything reduced it. One change with the 'modern' styles of Bodoni etc. which are less readable in text. Another change has been the sans serifs that proliferated in the 20th century; again not a gain but a loss in text.

A fourth thing is that you have different alphabets, and Chinese characters and I haven't ever heard that is makes a radical difference in the rate you can read with comprehension.

The basic thing with all of these written language systems is that they are systems of lines in closed or open forms with branches. The brain seems to get used to dealing with one, and is very, very good at decoding.

All this makes me think we are near some limit on what the brain can do, and the problem isn't shaping the alphabet a little differently.

What I think we can do is understand better what makes for comfortable reading, and for reading with less strain. Also at the limits perhaps some innovations can really help. These are such things as very small type, signs at night, low res. screens, etc.

I am interested in improving readibility, but all these things make me think we are talking about small improvments.

>2) What is "greatly"?

Well, I would be very surprised if you could increase reading speed with the same comprehension by more than 5% by any improvments in type design.

Any improvement is great, and there are also aesthetic considerations, which are very important as well.

There is something beautiful and magical in letters. I just don't think there are any revolutions to come in letters, as there surely will be in science, for example.

All this is explaining a hunch. I don't claim to have any iron-clad proofs.

[aquatoad]

Just weighing an a quick note about the

[hrant]

> people had a very long time with the alphabet and
> calligraphy to try to make letter forms easily readable.

1) They were essentially concerned with writing, not reading. Remember that the lc letters evolved from the UC through subconscious writing facilitation.*
2) They didn't know how we read (Javal made the first breakthrough, very recently), so how could they properly address it?

In fact the only pro-readability (more accurately pro-legibility) feature of the alphabet is the force of conscious "heterogenization" to alleviate blatant problems resulting from writing facilitation. For example, in cultures where the numeral "1" is written with a prominent "beak", the numeral "7" is given a horizontal bar to differentiate it.

* See "Canons of Alphabetic Change" in The Alphabet and the Brain, p. 133.

> the end test is the person reading, which hasn't changed.

The reader hasn't physically changed, but that's moot here. And the act of reading is in constant change, sometimes cataclysmic change, like when the Irish scribes introduced the word space in the middle ages, thereby enabling immersive reading for the first time. But more often it's subtle, like when people started reading less books on pedestals and more newspapers on trains.

> the Torah scroll ... is pretty much just as readable as anything today

The Hebrew script is actually known to be extremely slow to evolve. But Latin has changed a lot, not least since Gutenberg.

> changes from the early model of Garamond haven't improved readability, but if anything reduced it.

Different time periods have different amounts of respect for typographic quality. Sure, the 19th century mostly sucked, but what about Unger's fonts for example? It doesn't sound right for you to say there's no more room for improvement. That's never been true about anything in life.

> Chinese characters and I haven't ever heard that is makes a
> radical difference in the rate you can read with comprehension.

Where have you looked?

> 5%

I think we can do 15% pretty easily.
But even 5% is worth it, when you consider the aggregate savings.

And that's not counting improvement via dynamic and non-linear techniques.
My hunch is that in an ideal setup 50% is not unrealistic.

> There is something beautiful and magical in letters

As long we don't think they're Holy - thus stunting progress.

hhp

[hrant]

> I am a far sight better with a pencil and bezier program

The good news is that -contrary to what the chirographers will insist- the important thing is to control the boundary between Black and White - which is best done with outlines. When you focus on the relationship between the two edges of the Black instead (GN's "moving front"), good notan becomes demoted. You can't have both.

> constructed type

The modularity you describe is actually not a requisite of synthetic type. Not only that, I would even say that such modularity is a sister to a central tenet of chirography, where the modules arise from the motions the hand favors. This is what I meant by the [unwitting] brotherhood between Noordzij and Crouwel, and how they're both hopelessly superficial. BTW, the punchcutters used modularity too (through the use of counterpunches).

Modularity is its own thing (that needs to be controlled).

> you must learn how to fix the problems

And it's the same for chirography! Think of the "z" for example. Or optical compensation (something the GN school never addresses).

> people have forgotten how to write

They've also forgotten how to forge their own horseshoes. So what?
As long as they're leaning new things*, it's all good.

* Which I actually agree they're not doing enough, but that's tangential.

> Virtually every letter shows the effect of the pen.

I wouldn't argue that type has generally reflected the pen (and it continues to do so). Some people even point out that the famous Romain du Roi -which was supposed to be the first synthetic type- was in fact chirographic (at least to some extent). History is not the point, the future is. What does type need to be?

> don't forget your pencil.

Certainly, but only because the pencil is very different than the broadnib pen: the former creates outlines, the latter merely Black. I use a pencil myself, but only because computer software simpy isn't good enough yet - or you could say I'm not good enough to use it to good effect, since some of the most accomplished designers (like Carter) work directly on the computer. That's totally different than chirography.

hhp

[aquatoad]

I'm not going to try and argue the readablility issue of chirographic vs. synthetic. Just trying to suggest which might be easier for a beginner since they likely have more practice with the tools. (my main point, in response to the original question)

As I alluded to, chirographic type is a dying art, mainly limited to display at this point. So you've already won the battle Hrant. Perhaps the proof is in the pudding. Regardless, I can think of dozens of designers who are doing synthetic type for text. I'm having a hard time at this early hour coming up with one chirographic designer.

It is not the trade of blacksmithing that I'm notalgic for, it is crafting something unique and the message sent by doing so. It is the limited edition factor of letter writing. (totally different than making chirographic type for text