I would go small-caps whenever possible, proper nouns or not. Some caps/small-caps clashes will always be inevitable.

Pick a face with better small caps. They are obviously too low.

Even better: a face with multiple sizes
of smallcaps, such as Atma or Ernestine.

hhp

«"Allentown EMS," which, when EMS is set in small caps, seems inconsistent because Allentown is capitalized»

It's hard to say without seeing this in your actual setting, but I don't think things like this need to look «consistent» (because otherwise, the only sensible solution would be to just use the full caps for acronyms). I suspect you may be having a hangup that you don't need to have.

If you're going to hang it on a museum wall, do what typophiles like. If you're expecting a technical audience, do what they're familiar with -- unless a major part of the work is to teach an audience what typographers like...

> do what they're familiar with

But never ignore what people are "familiar" with
in their unconscious, often without even realizing.

hhp

Charles_e, yes I agree with your point. Standards, even if they're just de facto standards, are there for a reason.

Hrant, you may have jumped the shark here. Your post may haunt you.

In tech writing, capital letters are all over the place in the middle of sentences: spec acronyms, company names, company acronyms, people names, descriptive letters after names, units of measure, etc. And how do we keep the reader from wanting to stop at the capital letters? Answer: we use a typeface that has small capital letters. This is the simple, clear solution.
Greta Text: http://www.typotheque.com/fonts/greta_text/
Gentium: http://scripts.sil.org/cms/scripts/page.php?site_id=nrsi&item_id=Gentium
Of course there are many others.

The complicated solution is not necessarily the best one.

You haven't really been following my online activity have you.
Haunt? The ghosts fully settled in my living room years ago.
We're OK now. Sometimes they make me shark soup.

--

It's not about simplification; the human brain cannot be
described as simple. It's about accepting the difference
between what people think they're looking at versus what
their brain is actually look at*, and not simply settling
for the way things have been done so far. Now, Ben might
like to play it safe in this case, but ignoring how the brain
works is never safe.

* For example it's been empirically proven
that people think sizes above 12 are easier
to read while in fact there's such a thing
as text that's too large.

hhp

It may be better to leave kHz and MHz in uc/lc, no matter how much it Hz. AFAIK, that's both the canonical way and the standardised way.

Hrant, it is about communication. A piece can be poorly written, or poorly presented (or both). It is not about eventual comprehension. I belong to another forum where some people have abandoned capitalization, punctuation and spelling. Too much texting, I suppose. I can usually figure out what they're saying, but my tendency is to put them on my "ignore" list so I don't see their posts.

J. Tillman and Michael Boyer raise an interesting point. A font with a large x-height, and esp. one where full cps aren't as tall as the lower-case ascenders, minimizes the difference between lower case and full cap letters (and full & small caps). The only question is how well such fonts work with subscripts and superscripts. My initial guess is they can work just fine, but you'd have to do a bit of testing to be sure.

Here is a grab from the Gentium Specimen booklet.

It does not seem to be the x-height that is most relevant here, but the cap-height that is much smaller than the l-heigh. Digits are also smaller.

If you want smallish caps, Malabar might just be perfect.

> It is not about eventual comprehension.

You actually wrote that. Wow.

> A font with a large x-height, and esp. one where full cps
> aren't as tall as the lower-case ascenders, minimizes the
> difference between lower case and full cap letters

I'm sorry, but: Duh.
Also: As Michael indicates, Gentium's x-height isn't at all large.

What are we arguing about here?
I'm saying that some conventions of technical typesetting are junk*,
and the careful addressing of how smallcaps should be working seems
to be overdue. It has probably never been done; people were probably
oblivious to typographic refinement (not the same thing as typographic
snobbery) and simply used TNR and whatever other junk was just
sitting there. Ben is trying to improve his/the situation, and he's doing
it in the best place possible - so let's be a bit lucid here.

* Just as some conventions of type design.

hhp

Hrant, this is not a reply to any of your posts here, but I am going to quote you to define what I'm talking about.

"What are we arguing about here?" Always a good question. The problem we're discussing is the one defined in the original poster's first sentence. "I have some technical text with tons of abbreviations and initialisms." In the link in that same first post Florian Hardwig said: "Recently it was my task to ‘calm down’ a text churned up by all-caps names, acronyms & abbreviations."

The reader's eye wants to stop when it gets to all these capital letters in the middle of the sentence. That's the problem. And in the internet age, we have all become technical writers and technical readers.

"and the careful addressing of how smallcaps should be working seems to be overdue." Indeed. After reading numerous posts on small caps in technical writing, I conclude that the "rules" are very complicated. (If you are interested, check out http://www.typophile.com/node/84025 where a bewildering number of options are offered, in the pdf, to be considered.) It seems that for ALL instances of capital letters in the middle of a sentence, the writer (or typesetter) should evaluate each one individually and make a decision:
-leave it as regular capital letters
-use small caps
-use small caps and lower case,as in the link in the original post
-use the bigger size small caps
-use the bigger size small caps and lower case
-try a number of typefaces and reevaluate all the options over and over
And if the typesetter goes through the book, thesis, or paper and makes a thoughtful decision for every instance of capital letters, is that a success? Do we need the 147 rules of small caps explained in the supplementary book to the Chicago Manual of Style? Or does the reader want a simple standard, that can easily be applied across multiple books in the universe, for easy-to-read technical material?

Let's be kind to the reader. Let's be kind to the tech writers. Keep it simple and understandable. For technical writing, just use a typeface that has smaller-than-usual capital letters. Michel Boyer kindly demonstrated how effective it is. Forget about any small caps. (This is heresy in this forum, of course.) There are lots of fonts that meet this requirement, in addition to the ones Nina and I have suggested. Just do it.

I'm actually a big fan of smallish (main) capital letters.
And I don't pretend everybody needs smallcaps all the time;
for one thing I don't believe smallcaps should come included
in (most) base fonts: most people will never use them, but
they're paying for them anyway. I'd rather charge extra for
them to people who know what they're doing.

There are however two important points you might be missing:
- Sometimes caps actually need help standing out. Especially
in technical writing. This is because -unlike in something like a
novel- the reader often needs to [go back and] scan the text
for important words, and these words are quite often proper
nouns (ie capitalized).
- Immersive reading is not about the reader making choices,
it's about the designer accomodating the human reading
mechanism that readers don't know about, and don't need to
know about. Readers are not "confused" by caps of different
sizes - they just read. But when caps are the wrong size (they do
need to vary depending on the contexts of density and meaning)
then they snap out of it, and pay for our avoidance of complexity.

So: You certainly don't encumber the reader's consciousness
with rules; but you do try to follow the quite complex "rules"
of how humans actually read (without them even realizing it).
Compare this to cars: engines are very complex, because they
need to be; but simplifying the usability for the driver is of
critical importance.

hhp

Hrant, regarding your two points...

If you want to capitalize a word, just do it. With a font that has smaller-than-usual capital letters it's not intrusive. You don't need small caps.

The reader wants to see small caps (or capital letters) used uniformly across different books. If the reader is looking at several different Python books, and small caps are used differently in each, it will be aggravating. You using small caps predictably in your book is not sufficient. Thus if the rules for small cap use are not clear, or if they are being made up on the spot by this forum or the writer, there is a problem. ("and the careful addressing of how smallcaps should be working seems to be overdue.") And this is the case. Technical writers will be clearer (and less aggravating) to the reader by NOT making up their own rules, but just by using an appropriate typeface.

Hrant, nuts, I mis-read point number 1. I thought you meant the reader would scan for words in all-caps or small caps. But you meant the reader would scan for a word starting with a capital letter, a proper noun. You may have a point. I'll have to pay more attention to how I scan. But I think I look for a particular word more than a capitalized word.

> The reader wants to see small caps (or capital
> letters) used uniformly across different books.

1) I don't believe a reader can, should or wants to
care to this extent. And this is exactly why [text]
typography is an under-the-radar sort of endeavor.

2) By that logic we should limit ourselves in all
kinds of ways. Like only using Times. No way José.

> I think I look for a particular word more than a capitalized word.

It does depend on the type of work.
But most scanning is for proper nouns.

hhp

Another problem is that we often perceive things differently than they actually are, and good fonts make the appropriate adjustments. Here is a small example (with Gentium, but it works the same with other fonts I have tried).

1. "It is not about eventual comprehension."

You actually wrote that. Wow.

It is a little dense. Sorry. I write more academic stuff than tweets. (Actually, I don't have a twitter account.) How about "Just because a reader can eventually figure it out doesn't make it O.K."

2. You're going to have to give me your list showing what's typographic snobbery & what's not. Apparently I see a lot more snobbery than you do.

3. An by the way, IMHO, Times Roman was a pretty good font until the modern generation of font fiddlers mucked it up.

> Just because a reader can eventually figure it out doesn't make it O.K.

I can buy this.

Snobbery:
In the microcosm of type design, the only thing I've seen that can
be construed as snobbery is something I can actually be accused of:
seeing text fonts as superior to display fonts; I get flack for that
sometimes, but I stand by it. Besides that the only mildly offensive
thing I've seen is better characterized as one-upmanship: "My font
has 137 more contextual alts than yours!" This is a problem Matthew
Carter once referred to: people are putting so many bells and whistles
into their fonts they don't spend the time thinking about the basics.
This might also be why people like Thomas Huot-Marchand are
complaining* that there's little real innovation in type these days.
We'll fix that this year... :-)

* See Ruxandra Duru's recent paper:
http://www.eina.edu/arxius/docs_2/Duru_Ruxandra.pdf

What snobbery do you see, specifically?
I'd like to help you fight it.

> Times Roman was a pretty good font

Metal Times with ink gain was/is indeed great. But my main
point above is that anybody who thinks it's a good idea to
use Times for everything (even within a given field) really
does not get visual communication. And by the same token we
need to not feel so constrained by what readers supposedly
mind so much that they can't handle any deviation from.

hhp

This is a further question about the use of a font with smaller-than-usual capital letters. Hrant said:
Sometimes caps actually need help standing out. Especially
in technical writing. This is because -unlike in something like a
novel- the reader often needs to [go back and] scan the text
for important words, and these words are quite often proper
nouns (ie capitalized).

And this raises the speculative question: How much would a typeface like Gentium interfere with scanning for proper nouns. There are two extreme views. 1. If the capital letters are smaller they're a lot harder to see, so of course there will be a heavy scanning hit. 2. Well, the initial capital letters may be smaller but they're still capital letters, which look different than lowercase letters, and they're higher than the lower case x-height; the reader knows what the letters look like and will still find them.

Would any speculators offer a gut reaction (brain reaction) as to which way you are leaning?

The question that was raided was about technical texts. When a student gives me a thesis to read, I ask for the pdf, preferably with hyperlinks for equations and references. I don't scan the paper, I search the pdf. More and more journals and even books are available in pdf format. So, for that market, I would not worry too much about proper names that are hard to find in a paper version.

Good point about the rise of Control-F.

hhp

Snobbery:
In the microcosm of type design, the only thing I've seen that can
be construed as snobbery is something I can actually be accused of:
seeing text fonts as superior to display fonts; I get flack for that
sometimes, but I stand by it.

I agree with you. Does that make me a snob?

What I'm really after is the notion that design should play an instructional role in someone else's writing. Example: Rich Hendel had a student explain to him that he'd chosen a sans at a small setting size for a work on Shakespeare "to challenge the reader."

I dunno what Rich said to him; my reaction was "if you want to challenge the reader, write your own damn piece. It's not your role to interject that challenge to an author's work."

The specifics of the goblet may change over time, but Warde got it about right.

We'll fix that this year... :-)

Next year for sure, sorry.

hhp