http://www.typophile.com/node/38756

I read the link. The only person who got it right was Pattyfab.

. . . it’s a matter of house style. Some publishers italicize punctuation directly following itals, some don’t. There’s no across-the-board consensus. It’s very frustrating actually to try to keep track.

We keep a list of publishers with their preferences. One publisher lets the designers decide; even within one publishing house they don't agree.

BTW, publishers don't all agree on dictionaries (& hence hyphenation) either. In the States, many use Merriam Webster's -- the order of use is the Collegiate, backed up by the International. Note these don't always agree, which is why the Collegiate is the "first stop."

However, others use the Random House dictionary. You want to tell them they're wrong?

Worse, some of us learned to hyphenate words when the Second International was the latest dictionary. In the Third International, not to mention the many subsequent editions of the Collegiate, hyphenation has changed. "Standard" is one example. But you can't be looking up every word to see if someone has changed the preferred hypenation.

If you don't italicize the subsequent period/comma, you will get gappy text.
Note how the comma is practically midway between words, which is silly.
When both word and mark are in the same font, the font's built-in kerning is operative, as shown.
Otherwise, you are more or less committed to manually kerning (or applying optical kerning) on an individual basis for all words ending in f, r, v, w, y--if you want to do proper typography.

«Otherwise, you are more or less committed to manually kerning (or applying optical kerning) on an individual basis for all words ending in f, r, v, w, y—if you want to do proper typography.»

Which is, if you ask me, the only proper way to do it.

If the punctuation is not a logical part of the content that is being italicised, there is, to my eyes, absolutely no justification for italicising the punctuation.

(I notice that the International Reading Association [shall we just call them the IRA?] commits another of my stylistic pet peeve faux-pas, too: that of opting not to employ the Oxford Comma.)

I italicize just to piss off the editors, who inevitably mark it when proofing.

Oisín

"If the punctuation is not a logical part of the content that is being italicised, there is, to my eyes, absolutely no justification for italicising the punctuation."

––––––––––––

Yes,yes.

Giampa

I italicize just to piss off the editors, who inevitably mark it when proofing.

you cad

you cad

Something like that.

If you don’t italicize the subsequent period/comma, you will get gappy text.

Why not add properly kerned upright punctuation marks to the italics and a feature to use them?

Some house styles are simply wrong. Longstanding use does not make them right. There is, additionally, no such thing as a definitive “Webster’s” dictionary, since anyone can use that term. People working in the 21st century use Oxford, which has a range of credible dictionaries for national variants, including British, American, and Canadian.

I am waiting for someone to discuss the exceptional cases of bang and paren/bracket following an italicized segment (worst common example: d+bang+paren/bracket).

—
Joe Clark
http://joeclark.org/

Some house styles are simply wrong.

Tell an editor that and lose a client. "My way or the highway" designers usually don't get hired again.

«I am waiting for someone to discuss the exceptional cases of bang and paren/bracket following an italicized segment (worst common example: d+bang+paren/bracket).»

The obvious answer to that is to always make sure the last word inside the parentheses/brackets is not italicised. There is no other way to avoid gappy and odd-looking combinations.

The obvious answer to that is to always make sure the last word inside the parentheses/brackets is not italicised.

Not easy to do for instance in a dictionary where italics, parentheses and brackets may have conventional meaning. Here is an entry from my Collins/Robert English-French dictionary; parentheses are used for synonyms, partial definitions and syntactical information; square brackets are used for typical subjects or complements.

All parentheses are upright, but with brackets, the logic seems to have been bent!

Some house styles are simply wrong. Longstanding use does not make them right. There is, additionally, no such thing as a definitive “Webster’s” dictionary, since anyone can use that term.

Yes, some house styles are wrong. But not about this. There are good reasons for going either way. Writing is communication, and in communication, long-standing use does indeed make many things right. On the other hand, I'd argue that a specialist audience such as type designers and their aesthetic theories do not make make things right.

As to your "Webster's" comment, if you are referring to my post, please note that I said Merriam Webster's, not simply "Webster's." There is a difference.

 If you must have italicized words in parentheses, either use a font that has upright parentheses in the italic version and proper spacing there, or an application that can override the font’s spacing when opening or closing parentheses.
—Joel

«All parentheses are upright, but with brackets, the logic seems to have been bent!»

I’d say almost the opposite. In the cases in your snippet, at least, I would have italicised the parentheses, too. Looks silly like that.

Here is an analysis: If you use italic for brackets (& parentheses), you remove the need for hand-kerning between, say, and italic ell and the bracket.

[pill] as opposed to [pill]

But if the word spacing is tight, and the next letter has small or negative left sidebearing and an ascender on the left, all you do with the italic close bracket is to move the kerning problem over one character.

. . . [pill] languishes . . . opposed to [pill] languishes . . .

The problem is even worse with a footnote call.

As a book compositor, I use to prefer commas and periods following italic to be in roman. But see Nick Shinn's post, and remember that there are 7,000+ lines in a book, and hand kerning all that punctuation is not something most publishers will pay for. It also exceeds my boredom quotient.

I also apparently disagree with Bringhurst. There are times following a bold word/phrase when putting the punctuation in roman makes the text look odd.

hand kerning all that punctuation is not something most publishers will pay for. It also exceeds my boredom quotient.

Hand kerning in a dictionary where every page is filled with italics, parentheses and brackets is not something I'd like to have to do. On the other hand, if you add upright parentheses, brackets, commas etc to your italics, kern them once and for all, and add a feature to use them as alternates, would that not help at least for editing a dictionary?

Fiction wouldn't be much work. But non-fiction,with all the notes and citations, has a lot of italic. Moreover, some of the punctuation should be in italic, all you'd want to change would be when the italic ended or began, with parens & brckets). I don't know how to write a feature that could determine that sort of context. Eventually, I'll write something to use in pre-processing a file to be imported into InDesign -- we already do most of our coding & make an InDesign-tagged-text file to place as our import to InDesign..

We did this with TeX, using Vedit. Vedit allowed multiple buffers, so you could have your text in buffer 1, general search & replace routines in buffer 2, and font-specific routines in buffer 3. All the TeX codes are low-order ASCII, so for example, you could search for \IT{, move the cursor to the open curly brace, jump to it's mate, & overstrike it with a uniquely assigned high-order ASCII character as a scratch character. That gave a unique string to search for, & then insert a hard kern in the file.

We automated all this, the only downside was those hard kerns in the file that had to be cleaned up if you needed to later make an XML file.

If you can figure out how to specify the context of begin & end within a font, I'd love to know how.

If you can figure out how to specify the context of begin & end within a font, I’d love to know how.

I don't know if it is even possible. What I had in mind was much more simple and could not handle parentheses embedded in italic text. To handle the general class of texts you describe, I would also resort to preprocessing. So far, with LaTeX inputs (sometimes obtained using rtf2latex2e), simple sed (or awk) scripts have been enough. I have no idea how to process the Adobe .inx format.

Thanks for the extensive comments in this matter, it really helped.

It's not the italic comma you're gotta watch out for, it's the italic semi-colon that will haunt you forever!

Thanks for your insights folks, I've always half-wondered about this too!

On the other hand, I’d argue that a specialist audience such as type designers and their aesthetic theories do not make make things right.

Give some credit to designers.

Héctor

I doubt that adding upright punctuation marks plus a bunch of stylistic Set Features like
   01 -- upright comma
   02 -- upright semicolon/colon
   03 -- upright parenthesis/bracket/brace
would really serve designers and typographers well. It looks like a terrible workaround to deal with the OT limitation that no spacing adjustment is possible across fonts. I think this is something which is better addressed by applications, not by fonts.  ;-)

I doubt that adding upright punctuation marks plus a bunch of stylistic Set Features [...] would really serve designers and typographers well.

Let me state the problem differently. Assume I have to edit a one thousand page book containing lots of italics, in the main text, footnotes, endnotes and margin notes. I am asked to give out and InDesign file. The fonts I am using allow "adding functionalities". Let's assume I am experienced in adding characters to a font, kerning them to my pleasure, and adding features. Let's also assume I can write InDesigns cripts (which in fact I have never done). What next?

If my only resort is InDesign scripts, the script would essentially have to automatically kern the italic against the regular; is there any good algorithm I could implement to do that? If not, I would need to provide some kerning data, and all that looks like too much work for me.

I could also add a few upright characters to the italic (parentheses, brackets, braces, colon, semicolon, comma), kern them using standard tools meant for that job, and add a single feasure, say S01, that would just mean "upright" (that is simply the converse of the ital feature).

The rest of the job looks simple if I assume that InDesign scripts can do simple standard tasks. One task might be, given a selected text, to find the italics, and give the "upright" feature to parentheses, brackets and braces that begin and end a sequence of italics. Another script might just give the "upright" feature to punctuations. Maybe the script could be applied only to paragraphs of a certain style, for instance, paragraphs that are essentially in "regular" and where the italics are considered as embedded, and not governing the shape of adjacent punctuation.

Do you have any better idea?

[added] I have seen InDesign script that select the italics. I just guess the rest can be done.

Well, yes and no. The first problem is most foundries don't allow the end user to modify fonts, so with those fonts, what you propose is unlawful, illegal or something like that.

But there is always Adobe, which does give the end user such a right. Thinking this through, you had better standardize on which stylistic set you are going to use -- it will be called by number in the script -- and only use, say, SS19 for that purpose. I've had to go back into a lot of fonts & change more than one number as, over time, something became a standard at our shop rather than one on my whims.

Finally, everything looks right. The publishers who want terminal roman (& initial) have it, publishers who don't, don't. Until you export an XML file for one who want them roman, because those nice terminal "roman" characters are really in the italic font, which isn't what they wanted.

But there is always Adobe, which does give the end user such a right.

As I understand it, so does the commercial license of Canada Type.