I know that this type of punctuation was used for English in the United Kingdom, but I'm surprised to learn it is also used in France - and doubly surprised that it is also used in Québec, where I would have expected the practice from the United States as well as English Canada to be adopted.

If I typeset a text in LaTeX using \usepackage[french]{babel} that's the spacing I get automatically. See the babel documentation (pdf 2Mb).

It’s still prescribed usage. There is a very recent thread about it with some useful links.
Maybe someone could retrieve it.

By the way, Quebecois are often more strict than French themselves (if that’s possible ;-)

There was a recent thread about spacing for the slash: http://typophile.com/node/94281

Here are other relevant threads:
http://typophile.com/node/86513 (2011)
http://typophile.com/node/77959 (2011)
http://typophile.com/node/63720 (2009)

You can have a look at the pdf version of one of the UdeM Weekly News to see actual practice, for instance number 32 (pdf).

As a Canadian, I design and do production for many projects for one client that need to be done in both French and English. We use an extra space before a colon and after an opening guillemot (quotation mark) and before a closing guillemot, but not in any of the other instances you’ve mentioned. For a while we were kinda sorta also putting extra space before a question mark and an exclamation mark, but that was dropped; I don’t see much (if any) of that usage in Canadian French.

David

There is a large list of pdf documents from the Quebec ministry of health and social services you can consult just to see. Here is one with plenty of space before the question marks (pdf).

After reading more carefully the Government of Canada link I realize there is one major difference with the corresponding information on the OLF site: on your noslangues.gc.ca link, it is explicitly said that the thin space is about 25% of a normal space; on the corresponding link of the Office québécois de la langue française (types d'espacement), it is simply said that the thin space is "une espace insécable réduite", a reduced space; nothing is said on how much it is expected to be reduced. On a site in France I have seen that the thin space should be about 50% the width of the normal space. With LaTeX, I get a thin space that is 0.167 of an em, which may be quite large for some fonts; for instance with Arno Pro, I get a normal space of less than 0.2 of an em so that the thin space is about 83% of a normal space, which is too much for me. Here is a test with pdfLaTeX using Arno Pro installed for LaTeX. First the input file (using texlive 2012 and my own arno package)

\documentclass[twocolumn,10pt]{amsart}
\usepackage{arno}
\usepackage[french,english]{babel}
\begin{document}
\newlength\iws%
\setlength{\iws}{\the\fontdimen2\font}

\selectlanguage{french}
| e; e | e! e | e? e |\par
\def\thinspace{\kern 0.5\iws}
| e; e | e! e | e? e |\par
\def\thinspace{\kern 0.25\iws}
| e; e | e! e | e? e |\par
\selectlanguage{english}
| e; e | e! e | e? e |\par
\end{document}

and now a grab of the output:

The first line uses 0.16657 of an em, the second 50% of the normal space, the third 25% of a normal space and the last line uses the English spacing. The first line is much too loose. The second and third look fine to me.

Michel

PS. Sorry, I don't know how to do that in InDesign.

I know that this type of punctuation was used for English in the United Kingdom, but I'm surprised to learn it is also used in France - and doubly surprised that it is also used in Québec, where I would have expected the practice from the United States as well as English Canada to be adopted.

Such spacing used to be standard practice in the US too, I believe.

This is DeVinne, printed in New York in 1902.

[Craig — Note also that DeVinne’s opening quotes in that sample appear to be turned commas and very likely not cast opening quote marks. Hence the lower position and the wider set of the opening pair. Not that that has any bearing on the spacing convention, as they are obviously spaced before the capitals.]

There are a few examples in Karsten Lücke's notes on spacing issues (pdf, 4 pages).

As an aside, when I add the morse (Unicode spaces, dots and dashes) to the fonts/faces I use, I set the thin space to 75% of the character space (#32) and the hair space to 25% of same. Maybe not perfectly kosher, but works for me.

And here is a grab from Daedalus, an MIT press journal; the grab contains a colon, a semicolon and an question mark:

If you want to see more in details, there are free pdf files you can download from the Daedalus most downloaded articles.

PS. If you wonder how I found it, I just remembered the post http://typophile.com/node/87629?page=1#comment-485104 which referred to that journal, in which I had noticed the spacing.

Michel — It’s difficult to tell whether the spacing on those punctuation marks is from inserted hair/thin spaces or not. In the PDF sample that I downloaded, I could not detect any inserted space characters in the underlying text string.

The native fitting for those punctuation marks in Cycles Eleven happens to be fairly loose by design. Here is an unadulaterated Cycles Eleven setting of those same components, set without any preceding spaces:

Interesting, but the consequence is that no additional spacing seems to be required if I rely on your grab (I see no room left for additional French spacing if I rely on my grab).

Notice as an aside that in TeX, there is no "thin space character"; \thinspace is a kerning command. What babel does with the french option is that it first removes all space between the last letter and the punctuation mark (semicolon, question or exclamation mark) and then it positively kerns the mark, the kerning being done (indirectly) by the \thinspace command that I, maybe unorthodoxly, redefined above.

By the way, I heard that inDesign has a "grep" command. Is it then not possible, using a properly chosen pattern, to add some positive kerning before punctuation marks?

I just checked with the pdf and you seem to be right. It looks like no spacing was added before the punctuation; it is the spacing after the punctuation that appears to have been reduced, at least in the case of the semicolon in the line "possibility; this is more likely to produce".

it is the spacing after the punctuation that appears to have been reduced,

Or not:

As I said, Sumner has fitted these sorts rather generously. And asymmetrically, I might add:

Wow! The space is only 0.180 em! With a left bearing of 0.101 em, the semi-colon already has, I guess, about 0.075 em more than lots of other fonts; 0.075/0.180 is more than 40%. That is significantly more than the 25% suggested for French on the Canadian government site.

...Sumner has fitted these sorts rather generously. And asymmetrically...

My practice is to put the ; and : on the same advance width as the , and . but to offset them 15 or 20/1000 units to the right. Similarly, I always give more space on the left of ! and ? so that they stand away from the preceding letters a little.

I generally set semicolon and colon wider than comma and period. The potential pitfall of offsetting the colon to the right is that it will not be centered between figures when setting times, as in 10:00 for instance. But I suppose adding a standard positive kern for colon-@figures to offset the offset is not too burdensome.

Hi,

I am French and a former typography teacher and I can confirm that French typography rules still apply:
• you should use a thin space before semicolon, question mark and exclamation mark
• you should use a fixed width nonbreaking space after an opening French quote« and before a closing French quote» and before a colon

This has been super helpful -- many thanks for the detailed feedback and consideration of this issue.

One final question (for now): Is it standard practice to use hyphens (-) between number ranges instead of the standard (in English) use of an en dash (–)? Also is standard usage to use an en dash (–) in place of an em dash (—) in offset statements and interjections? I am seeing some conflicting advice on this. The manuscript I received from our editor includes things like: "1965-78" and "such and such – and such and such", which, to my eye, looks incorrect.

For dates, La typography – Année specifies that the two years must be written full length and separated by a trait d'union (hyphen) (1965-1978 instead of 1965-78). The same rule is given in my 1976 edition of the "Lexique des règles typographiques en usage à l'imprimerie nationale" (p. 108).

According to both the .gc.ca link Les signes de ponctuation et leurs espaces and the .oqlf.gouv.qc.ca link La punctuation, French distinguishes trait d'union (hyphen) and tiret. The entry Tiret : généralités of the OQLF simply states that it is longer than the trait d'union. Both sites encode their tiret as an en-dash; my 1976 edition of "Lexique" would have used an em-dash instead.

Things get tricky when there are dialogues, where the em-dash seems to be preferred. I have found sites (a full course on typography in Belgium, links in France and Québec) stating that the em-dash is to be used for dialogues and the en-dash for other instances of tiret. This blog in lemonde.fr however makes fun of that practice, arguing that there is just one tiret in French and you choose either the en-dash or the em-dash. The same blog states that use of the em-dash has almost disappeared from the press. I had a look at a few of my books and found five from different editors (two from Québec, three from France) that use the em-dash, and two others (one from Québec, one from France) that use the en-dash (consistently, so far as I could see).

Here is the opinion of Druide informatique, a respected local (Montreal) company that specializes in dictionaries and language tools, about des tirets plus ou moins étirés.

I'm I the only one who thinks this is how it should be done anyway?

Ryan, the same way the length of spaces can vary a lot between fonts — the space in Cycles eleven is 180/1000, which is ten percent less than a thin space if it is defined as a fifth of an em — so do as well the relative lengths as the absolute lengths of en-dashes and em-dashes:

(click to enlarge) For careful editing, I personally see no other solution than to fix or have fixed the lengths of those dashes to one's needs.

It just looks better with more spacing, that's all I'm saying.

Ah, I agree.