H&J is a broad topic, with plenty of room for debate.

Nevertheless, here is a set of basic settings that I give to the designers I work with as a starting point.



Make sure that in your Document Preferences > Character tab, the box for Standard Em Space is checked. (If you make these setting changes while no document is open, then these will become your default settings.)

These make for a decent starting point. I'm sure others will have different defaults from which they start. Whatever you do, don't use Quark's defaults.

The J settings will vary depending upon the line measure, the leading, the typeface, and personal tolerances (or preferences) for tightness vs. looseness.

For instance, with a longer measure and more leading, I may allow word spacing to be looser Max. And depending upon the fit of the typeface, I may allow character spacing of +1%. Conversely, with a shorter line, I may prefer a tighter Min. space. Also, with fonts that have an excessive word space built in, I may have to change all of the percentages accordingly.

The H settings may vary depending upon editorial standards and line measure. For instance, in narrower columns, I prefer to set Min Before to 2 and allow Caps to break and unlimited hyphens in a row. These looser parameters will allow Quark more options to try to achieve more even spacing. I then manually override any unacceptable breaks. However, some editors will not allow a 2-letter break before under any circumstances.

Regarding squooshing/stretching, I say Just Don't Do It. Same with tracking letterspacing.

-- K.

Kent's values are close to what I use.
However, I won't hesitate to horizontally scale a line a few percentage points, and/or tracking units, to get a good-looking paragraph/page.

Yves, I used to be a purist and set Character Spacing to 0/0/0. But I found through some experimenting that I actually couldn't always notice the difference in -1% (after putting it down and picking it up later, that is). So I started allowing it, just to give Quark a little more leeway to find a better line. But it does depend upon the typeface.

Regarding Cap breaks, it's just a difference in approach. I let the machine try it and I reject it if I don't like it, instead of trying to find places where breaking caps will help. The thing is, I'm usually going over everything anyway, so I let the program take a crack at it first. It depends a lot upon the content. In an Acknowledgement or something with a lot of names, I turn it off because it's a hassle to track them all down. And in medium to long measures, it's usually unnecessary. But in narrow columns, breaking Caps is one of those sacrifices I often consider.

Nick, I won't ever scale a line. But I will use additional manual tracking on occasion to fix problems -- depends upon the typeface and measure. But I tell designers working under me not to do it, as a rule, until I can teach them by example, and over time, what's acceptable and what's not.

Before resorting to tracking, though, I will usually try tracking just the word spacing. It's a useful trick, and more like traditional, metal typesetting. In Quark, if you have the TypeTricks XTension loaded, then highlight the line and using the following keyboard commands:

Shift-command-option-control-[ will reduce word space by 1 unit;
Shift-command-control-[ will reduce word space by 10 units.

A line will rarely tolerate more than -10 if my H&Js are set well. The same combinations with the right bracket will increase word spacing accordingly. Note that these key combinations are similar to the ones for applying tracking, with the addition of the Control key.

On my current projects, though, I have the good fortune to have a really good relationship with the editors and can usually get a rewrite (or an okay on *my* rewrite).

-- K.

Gentlemen, thank you. I'm touched by the attention you've given this thread. I brought some work home this weekend and will experiment with these settings as soon as I sign off.

I'll let you know how it works out...

"Avoid housework. Multi-task: Travel should be combined with exercise..."

OK, the avoid housework has been a given for a long time now. And I didn't mention it before but I just got back from walking to the library (my only exercise) to see if they had a copy of "Elements of Typographic Style". (They do, but it's out.) Hanging out here, combined with the new job have reawakened my interest in typography.

The nerd club, hmmm. Do we have a secret handshake?

Oh, and the supportive spouse? My sweetie is a creative person too and understands.

I have found this information by Spiekermann to be very helpful regarding H&J

http://www.spiekermann.com/iblog/C1109747452/E911593965/index.html


"Any particular rules? Hundreds! All written down in my books. One particular one quickly: Text is usually set too tightly and with too big a wordspace. One of the legacies of Quark Xpress. Always set the optimum wordspace (under H&J) to 80% or less in unjustified setting. And increase tracking by at least 2 units (that

The only thing I was ever taught to do concerning H&Js in Quirk was to change the amount of hyphens in a row from unlimited to 3.

Infact, because that's the only thing I ever did in Quirk, I never realised that there were any other settings until seeing the screen cap Kent posted.

:hangs head:

My solution to typesetting problems are first I won't use Quark ever again after using InDesign and second when I have a problem fitting text I don't play with the scaling of the type I go to my trusty copy writer and he/she will rewrite the problematic line in a way that fits better. See two heads are better than one.

I don't agree with Erik's guidelines. I do agree that Quark's defaults often yield too tight letter spacing and too loose word spaces. But I think Erik's guidelines overcompensate.

I don't think 80% word space in unjustified is always a good idea. It really depends upon the space and fitting of the typeface. For some, maybe; but not as a general rule in my book.

I would especially *not* combine +2 units tracking with 80% word space. Again depending upon the typeface, the individual words generally start to pull apart and the word space between them starts to close up and approach the size of the intercharacter spacing. You'll get a more even gray texture on the page this way, but I don't think this is necessarily good for the readability. For instance, I think that the text of Stop Stealing Sheep is a bit too loose in this way. (I'm looking at the first edition.) Perhaps he gets away with it in this case because the style of writing is short and bite-sized, but I wouldn't want to read a novel or lengthy essay set like this.

I agree that with smaller sizes, positive tracking is a good idea, just as it is a good general rule to use tighter tracking in display sizes. But my threshold for opening up the character spacing is usually down around 8 or 9 point, not Erik's 12. Again, it depends upon the typeface. Most reading faces are designed and fitted for use around 12 point. As a general rule for the average graphic designer, I would not recommend varying from the designed fitting until you get down into the smaller extremes.

As far as that unlimited hyphen thing goes -- I have found that even when I allow Quark to use unlimited hyphens, it rarely provides more than three in a row anyway. Sometimes, four hyphens is the only way to achieve acceptable spacing in the paragraph.

Again, I let the program have some leeway to find what it thinks is the best solution, knowing that I will take the time to review everything anyway and can reject those that are not necessary or not acceptable.

I personally don't think hyphenation is the worst compromise in justified text, but it requires human judgement to determine when a hyphenation is cumbersome and impedes the sense of the text and when it is perfectly natural with the phrasing and cadence.

But there's nothing more frustrating than trying to tweak a paragraph by forcing a hyphen and not being able to figure out why it isn't allowing me to adjust it ... until I finally realize that there are already three hyphens and Quark is blindly obeying the three hyphen rule. After all, that's what machines do.

-- K.

I think generally rewrite is a last resort, depending upon the material being set.

When I'm working on a book, I feel that the author has said what she means in the way she means it and it's my job to make that work. In this case, I will generally only request a rewrite when there are no acceptable alternatives. But sometimes, that's what's required.

If it's average everyday copy, though, then rewriting is usually not a big deal. Sometimes its the most expedient solution.

-- K.

Hey! Post #500!

I have a problem with rescaling type and it has to do with time and other people. What if you have a project that gets reprinted quarterly and you happen to be on vacation. The person making changes doesn't have a clue about what tricks you made to make the type fit. I think unless its a one-time project the operative word is KISS

One last thought on rewriting, I don't mean change the context of the sentence, what I mean is to change one word or eliminate a word. I had a creative director years ago that loved the word concept, I would change it to idea and save 3 characters.

I'm really enjoying this exchange on H&J's, a subject which I've been immersing myself in for the last month or two. I'm about half way through James Felici's The Complete Manual of Typography, and have just started rereading Bringhurst's Elements (which is even better now than 4 years ago).

Felici suggests (with many examples), that flexing character spacing with care and respect, along with consideration for how much the word spacing is flexing, is fine and even recommended. Whereas Bringhurst seems less open to that idea. My questions are these: are there any typophiles that allow more flex in character spacing than what Bringhurst and Kent suggest? (I'm talking about a max of +/-5%) Also, Kent, are you speaking about book typography, or do you include magazine typography, with its tight columns and fast turnaround?

I'm aware that flexing only word spaces is more 'purist,' but I rarely rely on purism for anything, and would rather settle for reasonable, functional, and attractive. Felici seems to contradict Bringhurst frequently in respect to technique, particularly since Felici is speaking specifically about digital typesetting, and Bringhurst is covering all typesetting. Also, Felici feels more realistic about the environments that typesetting is generally done in, wherease Bringhurst strikes me as describing typesetting ideals. They seem aimed at the same end results (solid typography), but have varied methodologies and philosophies.

Like Chris, I was taught next to nothing about H&J's, and my current employer couldn't care less, so I've taken on the responsiblity myself for a while now. I'm trying to figure out what should be discounted at face value, what should be accepted readily, and what should be used with care and judgement. My questions aren't meant as a challenge to anyone's assertions, I just really want to get a solid footing.

I too am checking in occasionally and enjoying the different viewpoints. I'll post a much longer reply tonight or tomorrow, as time allows. Gotta go to work now. :-)

> I don't agree with Erik's guidelines.

I think the main problem with them is the same problem with any canned rules: they over-ride the more important reason for the rules. And this isn't just an "acedemic" issue - specifically in this case (and as Kent points out) if you happen to apply both of the rules Yves reiterates (set smaller sizes with 80% wordspace and looser tracking) it'll backfire! In fact the way Spiekerman says that his font should be exempt for the loosening rule (like what about other loose fonts - how do you recognize them?) is a hint of the cardinal axiom: It''s Not So Simple.

Is it that hard to try to teach people the WHY?

hhp

Kent's setting is very much the old hand-composition aesthetic.

The crazy flexing that software does was really difficult with foundry type, involving tricks like shaving individual pieces of type.

So it makes sense to use Kent's old-school H&Js with an old-style face like Bembo, which was designed for that kind of setting (and no fancy kerning either!)

However, for something like ffDIN, a digital version of signage lettering, or Officina, open-fitted offspring of the typewriter, a different H&J strategy would be OK. Even for a present-day oldstyle like Minion, which one would assume Robert Slimbach designed with kerning in mind

The thing is, when you first start to use a face in a project, you can experiment with the vartiables. Size and leading, most obviously, but H&J too.

> Minion ... kerning

Speaking of which, the other day I noticed that Minion has a kern for "Yq" but not "Yg":


Tsk, tsk, tsk... ;-)

hhp

I don't think it is Minion (check this).

Although it seems you're right anyway.
MyFonts' Minion sample has too much space between Y and g too.

Oh. And look at the cupped serifs. Must be Adobe Garamond.

I guess that either means:
1) I'm so obsessed with kerning I can't even tell what font it is.
2) Slimbach needs to diversify a bit. :-)
Or both.

> MyFonts' Minion sample has too much space between Y and g too.

He probably just reuses kerning pair lists, which is a good idea for designs that are so close (but it does assume the base list is solid).

hhp

Kerning is a complex animal - best done by somebody who thinks like a programmer.

It's not just a matter of having a big list of kerns, it's a matter of having a tight list of kerns.
For example if you're going to kern "Ya" and "Yg", you should really kern like "Y

Yes, I believe it's Adobe Garamond.

Adobe Garamond and Minion are both designs in the same broad genre, and by the same designer, so it's not surprising there are similarities. But there are also a lot of differences.

T

> Slimbach needs to diversify a bit. :-)

http://www.daidala.com/17mar2004.html
:-))

In response to the other recent thread on sentence and word spacing, I got a copy of Dowding's 'The Finer Points of Spacing'.

I think his views are interesting in the context of this discussion also. He begins the book by railing against excessive word spacing. This was apparently common (in the '40s and '50s) because typesetters were paid by how many 'ens' they set. Thus it paid the typesetter just to end a line quick and dirty and let the machine space it out.

Dowding shows an example of this loose and sloppy style, and it makes the case that the line breaks up with excessive spacing. He says you can counteract the problem with more leading, but tighter word spacing is better. He quotes Tschichold as recommending 1/4 em and recommends this.

Now we could note that Tschichold, supervising book production, had an economic interest in keeping down the number of pages, and hence going for tight spacing. Dowding's book itself is too tightly set in my opinion. It is in 9/12 Times New Roman set to a measure of 23 picas. Bringhurst says that for small type sizes 1/3 em may be better, and from Dowding's book this seems to be so.

I think that in the '60s, with the advent of photo type, the 'close but not touching' school went overboard the other direction, of tight spacing, in my opinion.

Now that we have such experts as Bringhurst, Spiekermann, Felici, and here Kent Lew and Nick Shinn all not agreeing tells me that people still haven't figured out the new technology of flexible word and letter spacing and glyph stretching, and how it relates to an optimal reading experience.

I think Spiekermann's refutes his own case by immediately saying that you shouldn't follow his own rule with 'Unit'. Surely other faces respond differently, as others have said.

Since InDesign is new technology, it would be interesting to hear what Adobe has done in testing out and developing its technology on this issue.

It seems that no one has tied set width to the H&Js. Certainly setting narrow columns justified (most newspapers) will require different H&J values than setting wider book makeup.

>Also, Kent, are you speaking about book typography, or do you include magazine typography, with its tight columns and fast turnaround?

My primary frame of reference is book typography, since that is what I am most involved in. But some of the books I work with are two- or three-column formats as well. I think the general principles apply broadly.

For magazine composition, I will use different settings than for book composition. But I will be using the same principles to evaluate my decisions. I judge my work against high standards, even though I rarely meet them.

And yes, I have to make the same compromises as anyone else when confronted with difficult situations or time constraints. Narrow columns will require more compromises than traditional book pages, generally. Not every piece of copy is worthy of the same attention or effort to get it just perfect. Sometimes you have to let less-than-preferable hyphens or somewhat loose lines go. I try to be practical.

I think the important thing is that one should know *what* it is you are compromising.

Periodically, I will review my assumptions. Here is a controlled test I do now and then with a new face:

I'll set several lines of sample text unjustified (not greeking, but actual text, something representative of the type of work I'm currently doing) with hard returns at the end of each line.

Then I take every third line and apply progressively tighter tracking in -.5 unit increments -- i.e., the first and second lines are set normal character spacing, the third line is -.5 units, the next two lines are normal, the sixth line is -1 unit, etc. (Sometimes, I'll even mix them up more randomly. And sometimes I'll do the same thing with progressively looser tracking, too.)

Then I set this aside and come back to it later. I try reading the text and note when the words become noticeably tight (or loose). It is difficult to be a reader and at the same time try to pay attention to the reading itself; this is by no means a perfectly objective test. But by this method, I establish my parameters and tolerances.

+/- 5% is really excessive -- note that this is equivalent to 10 units in Quark. I find that I cannot abide more than 2 units (1%), and am more comfortable with 1 unit.

My bottom line is that word spacing is more flexible than character spacing. As a reader, I can tolerate greater variation in word spacing, much more than I can tolerate distorted word shapes through even seemingly modest variations in character spacing.

I am, perhaps, admittedly old-fashioned in my aesthetic, as Nick points out.

I think the important thing is to keep looking and analyzing and learning. Notice good composition; notice bad composition; what works and what doesn't. Consider what makes it so. Develop and refine your own discrimination and judgement.

Then you won't so much need Rules.

-- K.

BTW, Hrant, I think you'll be hard pressed to find very many fonts with a Yg kern pair. I will confess that Whitman does not have a Yg pair, even though it has Yq (and Ya, Y

>

Oops, your right. My apologies.

Kent, thanks for the comprehensive response.

It's funny, I'll never alter tracking more than +/-3 units in Quark when I manually adjust type, so in that context %5 (10 units) adjustment starts to seem pretty ridiculous. I need to jot down the cross-program and mathematical equivalents so I don't overlook the obvious again.

I'll give your method a try with some faces I've purchased recently. I like having a method, even if I end up deviating from it (and I usually do).

"I think the important thing is that one should know *what* it is you are compromising."

I completely agree.

> I judge my work against high standards, even though I rarely meet them.

Now that's a golden quote (in the immaterial sense, of course :-).

> you'll be hard pressed to find very many fonts with a Yg kern pair.

Certainly, and I don't think it's a big deal to miss it... as long as you also exclude "Yq"!
Being heavily programmatical, kerning makes most sense when it's consistent, although it can get complicated, especially in terms of linguistic sensitivity. For example, Mana-16 has a kern for "vv" (because it would look like a "w"), but not "ww" (because the undesirable looseness wouldn't be worth it since the string wouldn't generally be confused for anything else), but it also has a kern for "yw" but not "wy" (because of the orientation of the tail of the "y", in terms of filling the space).

BTW, these days lc-UC pairs need a lot of attention too.

hhp

BTW, Kent, I forgot to mention something:
Since Whitman I've been an official user of your emdash-emdash closing-up kerning trick!

hhp

Hello Kent, Yves, Stephen, Nick, Gerald, Chris A., Daniel, Chris R., Hrant, Eduardo, Thomas, William, Tiffany and James. Thank you all very much for the education, and congrats on making this a handpicked thread. Some questions and comments:

"Make sure that in your Document Preferences > Character tab, the box for Standard Em Space is checked."

Can anyone explain what this does? Why it is a good idea. (I like to know why too.)

Some books were mentioned that I will be looking for. Tiffany, I haven't forgotten that I owe you a book review. Soon.

InDesign? Someday maybe. My use of Quark is and has been dictated by interaction with other Quark users. It is still good to be informed, though, and I am taking note of the comments.

To help with future projects my plan is to set up a sample document with several repetions of the same two paragraphs in the same type box and then create a variety of H&J styles based on the numbers discussed in this thread. When I have a new project I should then be able to quickly set the type in the desired font and get a quick look at a couple of options.. That's the plan anyway.

Thanks again.

>the box for Standard Em Space is checked

1. An em space is two en spaces.

2. This width is not derived from or related to the width of the normal space character.

3. Type "option space" to insert an en space

4. If the box is checked, the width of the em space will equal the point size of the type. (And the em space will be an em "square". In fact, a pica is merely a 12 pt. em. In hand composition the length of a line of type, or a column of type, measured in ems, is significant. In particular, typesetting was priced by the em. Also, knowing the number of characters per em in a font aids in copyfitting --and for this reason line lengths are also measured in ems. Furthermore, the em square, being a little piece of metal with a symmetrical cross section, has a strong aesthetic presence, enabling the typographer to conceptualize and design space in a scalar, proportional manner, rather than through absolute distances. The zen of the em.)

5. If the box is unchecked (Quark default), the em space will be the width of two "0"s -- ie the en space will be the width of the figures in a font with tabular (monowidth) figures. So this may be useful for some kinds of tabular work with figures, and I would imagine that's the reason for this Quark option.

6. The difference between the options is more apparent in a condensed or extended typeface.

7. In 15 years, I've never used the em/en space in Quark. Maybe that's my loss -- but does anyone have examples where it's been useful?

Alyce, Nick gave a good explanation of the difference between the standard and Quark default em spaces. But let me summarize: a standard em space is a relative unit equal to the point size; the Quark default em space is variable defined as twice the value of the width of the 0 (zero) in a font. So the standard em space is size-dependent, while the Quark default is size- *and* font-dependent.

(I believe this default was inherited from some phototypesetting systems and may have some validity when working with tabular matter or extremely condensed fonts. But not as the default, IMO.)

The reason I recommend checking the Standard Em square is that it makes some things more predictable.

Although the word spacing is based on percentages of the spaceband character width in a given font, character spacing is determined in units which are 1/200th of an em space. I believe that Quark calculates this differently based on the preference setting for the value of the em space. (There's a slim chance that I'm wrong and that it uses the point size value regardless of whether the standard em space is checked; but I don't think I am. I'd have to go back and do controlled tests to confirm.)

Nick, I use the en and em spaces periodically. One example is when I have a run-in C-head. Frequently I will decide to style the head in a bold sans serif, set it without any terminal punctuation, and use an en or em space (depending on the line measure) before the beginning of the paragraph proper.

Sometimes I will use an en space following a graphic bullet, when the list is not being set hanging.

I will also frequently use an en or em space before an end-of-story dingbat.

The most important reason I choose to check the standard em space is that I use the Quark flex space as a thin (or hair) space, and the flex space is defined as a percentage of the en space (which in turn is determined by the definition of the em space). I like to define the flex space at 25% (which yields an M/8 space).

I then frequently use this flex/thin/hair space to slightly separate em dashes -- a compromise between the two competing camps of closed or open em dashes, and again, a convention that I've derived from traditional setting.

I also like to use this flex space to separate the multiply symbol in dimensions, or sometimes in botanical hybrid names (depending upon how the typeface sets and where I'm taking the multiply from).

-- K.

Alyce, your idea for a sample comparison document is not a bad one. But I will say that so much depends upon the quality of your actual text (is it punchy journalism with a lot of short terse words or is it academic text with a lot of long jargon, for instance) and the line measure (narrow column vs. full-page measure) that a standard document may not be adequate.

A slightly different approach that I might recommend is to create a resource document in which you develop several H&J styles -- for instance, your Standard might be 80% / 100% / 133%, but then you define a Tight at 75% / 100% / 133% and a Tighter at 67% / 90% / 125%, or some such (these are just random examples).

Then when you start a new project and you're establishing your line length and type size and leading, you can import ("Append," in Quark lingo) these H&J styles from your resource document and quickly apply them to your actual text.

(Once you've made your decisions and the work begins in earnest, you may wish to delete any unused H&J styles to prevent confusion later.)

-- K.

Kent -- where you use the flex space for adding a little separation, I use kerning. In the other instances you mention, I set tabs, or horizontally scale the space character. The advantage of this method is that by tapping the keyboard command which varies the value incrementally, one can adjust a space until it looks just right.

I am also very fond of "command-backslash" (as an alternative to paragraph formatting) for hanging indents, although it can really befuddle other people working on the same file!

There are so many ways to do everything with software...

(Alyce, every time I see a post from you, I hear that old Jam tune...)

>where you use the flex space for adding a little separation, I use kerning.

True, but with a flex space, I can search and replace an entire document's worth of em dashes. Or I can search an entire document's worth of [inchmark - space - multiply] combinations. Without messing with custom kerning tables or anything.

I also use the command-backslash to hang paragraphs, but you're right, others following can get confused. I also have been known to use non-breaking spaces following things like "I" or "a" or "an" in unjustified text in order to keep them from hanging out at the end, but to allow the text to flow if the paragraph gets rewritten (as opposed to what happens with a forced break). Or to keep measurement numbers with their units.

But again, this can mystify colleagues who follow after. (Although, it's less problematic, in my opinion, than hard breaks that get bumped because of an AA.)

-- K.

>I can search and replace an entire document's worth of em dashes.

I see what you mean. My m.o. is derived from working mostly on short documents, but I will try and remember your technique next time I tackle something longer.

I use the non-breaking space a lot for web type.

"(Alyce, every time I see a post from you, I hear that old Jam tune...)"

OK, I'm skipping over the additional typographic education (thanks Nick, Yves and Kent) to get right to the most important question. :-) Song title, please... even better, lyrics please.

"Town Called Malice", by The Jam. Search for lyrics here.

> "A town called Alice, ooh ooh yea-eah"

Hehe. Something funny for you (scroll down).

Being compared to a town called Malice doesn't seem so bad when you consider the alternative...

Misheard Lyrics:
Time is short and life is cruel
and its up to us to change this Tart called Alice
Correct Lyrics:
Time is short and life is cruel
and its up to us to change this Town called Malice

Sing it, Yves! :-)

Eduardo, I really like that site. From there, in the same vein:

"White Rabbit"
Misheard Lyrics:
Tell them a hooker and a smoking character
Has given you the call.
Correct Lyrics:
Tell 'em a hookah smoking caterpillar
Has given you the call.

The Jam song puns Nevile Shute's 1950s novel "A Town Called Alice" (Alice Springs, Australia).

Ah, but the grandaddy of all mistaken lyrics sites is:

http://www.kissthisguy.com/

As in the Purple Haze lyric "Scuse me while I kiss the sky"

This is a great thread. Good practical information about H&J from Kent Lew and others.
William Berkson said earlier that he thought that Dowding's word-spacing (in Finer points ...) was too extreme. It certainly looks too tight when you first read the book but after a while it seems just right, at least to me .
Here is the quote from Tschichold from the Penguin Composition Rules that Berkson mentions: "All text composition should be as closely word-spaced as possible. As a rule the spacing should be about a middle space or the thickness of an 'i' in the type used."

Glad you brought the conversation back to type, jbs. I've been rereading the entire thread and am even more appreciative of the willingness of typophile.com members to comment and contribute.

I like the command-backslash tip, and will be looking for a place to use it soon. Any others that might not be common knowledge?

(Nick, coincidently I just finished rereading "The Chequer Board" by Shute. "A Town Called Alice" is an old favorite too.)

Doh, I know I shouldn't post before my morning coffee. The correct book title is "A Town Like Alice". Oh and BTW, I got my T-shirt today. :-)

About the flush zone, forgive my ignorance but what does it even do? Everytime I change it to zero, I don't see a difference in my stuff.

Em