All 4 books by Edward Tufte now in
paperback editions, $100 for all 4
Visual Display of Quantitative Information
Envisioning Information
Visual Explanations
Beautiful Evidence
Paper/printing = original clothbound books.
Only available through ET's Graphics Press:
catalog + shopping cart
All 4 clothbound books, autographed by the author $150
catalog + shopping cart
Edward Tufte e-books
Immediate download to any computer:
Visual and Statistical Thinking $2
The Cognitive Style of Powerpoint $2
Seeing Around + Feynman Diagrams $2
Data Analysis for Politics and Policy $2
catalog + shopping cart
Edward Tufte one-day course,
Presenting Data and Information
Tampa, March 1
Boston, March 14, 15, 16
Oakland, April 20
San Jose, April 21
Palo Alto, April 24
San Francisco, April 25, 26
Punctuation typography

First some refugees from another thread:

And, by the way, in the top line of the iPhone screen, the colon in the time stamp on iPhoto (10:36 PM) should be changed to a period (10.36 PM), just like the times in good railroad timetables. Also AM and PM should be in lower case. So instead of 10:36 PM use 10.36 pm. These and similar typographic delicacies for time-stamps and timetables are discussed in Enivisioning Information, 104-105. -- Edward Tufte, January 11, 2007

"." or ":", using a 24-h clock must be superior to am/pm? -- Peter H (email), January 12, 2007

In re: ET's inclination towards the "period" instead of a colon for time-stamps, I have begun to see (and use):

i) the "period" in phone numbers as well - 650.555.1212, and

ii) only a small "a" or "p" after time numerals - "....will meet you Tues. at Peet's on campus, 7:15a."

These seem tidy; any reason/s to think either confusing or unappealing? -- David J (email), January 20, 2007

-- Edward Tufte

a for am and p for pm are unconventional and attention-attracting. The 24 hour clock is excellent and if the United States ever gets on the metric system . . .

For phone numbers, might this form 607 544-7031 be preferred to indicate the area code part? Any advice about European phone numbers, which are sometimes confusing in their mash-up of country and local area codes?

The point of the period instead of the colon in timetables is (1) to avoid making a units distinction look like an active element and (2) to preserve the little space between hours and minutes as a space. This is particularly important for small type (timetables) and for lower resolution type (computer screen). If this is the case, then for phone numbers the hyphen and the period would both seem to work fine.

-- Edward Tufte

I'm from Germany and a compositor by trade, so I can give you some advice on our phone number system.

First of all, German area codes and phone numbers vary in length depending on the size of the town or region but usually have a total length of ten digits. For example, Berlin's area code is 30, Dresden's is 351, and my hometown's (really a village) is 35873. Instead of having a fixed-length format like the American XXX-XXX-XXXX, we separate all digits by spaces in groups of two from the end (or groups of three or even four, where appropriate):

1 23 45
12 34 56
333 456
5678 1234

Area codes are usually put in parentheses but can be separated by a period or slash, and extensions are appended after a hyphen or period. (I prefer the parentheses notation because I find it to be the clearest.) Furthermore, when dialing within Germany, area codes are prefixed by a zero:

(0 30) 12 34 56 78
03 51/1 23 45 67-89
03 58 73.1 23 45

In international notation, area codes are not put in parentheses. The leading zero is replaced by the country code 0049 when dialing from outside but can be put in parentheses if the phone number is published both in Germany and abroad. The double zero itself is usually replaced by a + sign, which is more general because the US, for example, use 01149 as the German country code:

+49 (0)30 12 34 56 78
+49 351/1 23 45 67-89
0049.3 58 73.1 23 45

If you want to be really fancy, you can use half-spaces (the correct English typographical term escapes me) instead of normal spaces for separation. However, this usually only matters to people like me who use layout software like QuarkXPress and InDesign.

I hope this helps and is of interest!

-- Martin Winter (email)

en-space and em-space are probably the terms in English, but with Quark those terms have disappeared from my point of view.

-- Edward Tufte

I doubt if am/pm will ever be dropped and 24 hour times adopted in the US even if you did metricate. Australia became a fully metric nation 30 years ago and 24 hour times are pretty much restricted to (mostly airline) timetables. They are never used elsewhere.

One interesting thing with timetables is the use of leading zeros for minutes. In Envisioning Information ET discusses the New York to New Haven timetable (pp104-105). The 1913 timetable has zeros before the minutes. Whereas in the UK and Australia, leading zeros seem to have only been used from the late 1960s onwards. So 7.00am was shown as 7. 0am. Was it because of digital clocks?

Also are leading zeros almost like chart junk before hours and days? Looking again in EI in the discussion of the Vietnam Veterans Memorial (p43), birth and death dates have a leading zero for the day. I feel it detracts from ease of reading. 4 MAR 1945 is easier to read that 04 MAR 1945.

-- Andrew Nicholls (email)

The 04 for the fourth day of the month is surely military style, an appropriate form and tone for the Vietnam Veterans Memorial. Had the leading zero been omitted, then some bureaucrat would want to write "this little space left intentionally blank."

-- Edward Tufte

The International Telecommunications Union's Recommendation E.123, "Notation for national and international telephone numbers, e-mail addresses and Web addresses" is one international standard for formating phone numbers and web addresses. The standard can be downloaded (currently for free) from Wikipedia also provides some historical information on telephone numbers. See

-- Keith Dishman (email)

European business cards often provide the country and city code with a "plus" symbol to indicate the need to dial an international access code (such as "011" from the US, but frequently "001" in other countries):

+44 1223 XXX XXX (Cambridge)

+41 1 XXX XX XX (Zurich)

+49 69 XXX XX X-X (Frankfurt)

+33 1 XX.XX.XX.XX (Paris)

Most countries require that you add a "zero" to the city code when dialing in-country. Each country also seems to have its own punctuation and breaks for numbers. The French example (which often has people giving the paired digits together: would be spoken as "twenty-four, sixty-six, twenty-two, ninety-nine") has carried over in modern punctuation here in the US, but the pairing has not, it seems. I recite each number separately, but many don't.

I tend to write my telephone number as (+1 XXX) XXX-XXXX in most correspondence. I put the number first and then the function (office, telefax, mobile) after.

-- Claiborne Booker (email)

UK numbers add up to ten digits total, but how they are organized depends on how old the exchanges are (my layman's understanding: someone with knowledge of UK telecoms history may correct me). Cambridge, following Claiborne's example, is

1223 XXXXXX or 1223 XXX XXX

but Central London is

207 XXX XXXX or 207 XXX-XXXX

(the space between the first and last three digits in the old six figure codes is purely a matter of legibility, while the one between the first three and last four digits in the new codes represents a real structure of dialling areas. A hyphen replacing the space is frequently allowed in the latter case, but would be considered inappropriate in the former. Presenting all six digits unpunctuated is considered normal in the former but presenting all seven digits unpunctuated would be frowned on in the latter.

Those familiar with the UK will have noticed I missed out a leading eleventh digit, the zero. But that's not really part of the number, but only a signal to the phone that you are about to enter a Standard Trunk Dialling number, which is why it's dropped for international dialling.

-- Derek Cotter (email)

It's a little more complex than Derek Cotter suggests: 0207 is not the dialling code for central London. 020 is the code for all of London, it just happens to be a historical accident that 020 7XXX XXXX numbers are in central London and 020 8XXX XXXX in the outskirts; 020 3XXX XXXX numbers are just coming on stream now. The preferred way to present a London number for an international audience then is

+44 20 7123 1234

For a domestic audience it should be

020 7123 1234

You can combine the two in the ITU stand shorthand

+44 (0) 20 7123 1234

The spacing between the first and second chunk of local digits is for legibility but leaves an unbalanced look to the number (the French usually break their numbers into pairs: eg

+39 12 34 56 78 90

Which looks nicer.) The reason we split them that way in London is again a historical accent: London exchanges used to be know by a three *letter* mnemonic - WHItehall 1212 getting you the Prime Minister. So there was a semantic difference between the first three numbers (which stood for letters) and the latter four.

In the late 1980s BT ran out of London numbers and so changed the universal (01) prefix into an inner and an outer ring pair (071 and 081). Then, in order to fit with the ten digit scheme a 1 was added so London became 0171 and 0181). The we lost our iconic 01 again and the inner/outer code was moved to the local portion of the number. It would be churlish to suggest that changing the phone number of your capital city three times in 15 years suggests a lack of foresight.

Incidentally, the 24 hour clock is not a metric standard.

-- J L Smith (email)

Claiborne Booker:

The French example (which often has people giving the paired digits together: would be spoken as "twenty-four, sixty-six, twenty-two, ninety-nine")
If only. It would actually be spoken as "twenty-four, sixty-six, twenty-two, four-twenties-ten-nine"). For taking down numbers spoken on the phone it's worst with numbers in the seventies, such as sixty-twelve, where one has already written down a 6 before realizing it should be a 7. They order these things better in Belgium, where they have a proper word for seventy.

J L Smith:

In the late 1980s BT ran out of London numbers and so changed the universal (01) prefix into an inner and an outer ring pair (071 and 081). Then, in order to fit with the ten digit scheme a 1 was added so London became 0171 and 0181). The we lost our iconic 01 again and the inner/outer code was moved to the local portion of the number. It would be churlish to suggest that changing the phone number of your capital city three times in 15 years suggests a lack of foresight.

Didn't two of the changes occur within about two years of each other, showing real lack of foresight?

Nonetheless, I prefer denying the capital city its iconic 01 to renumbering the entire country in order for the capital to maintain its pride. Consider what is going to happen to car licence plates in France in the next year or two. When they run out of numbers for Paris, which will happen quite soon, but at least 20 years before they run out of numbers anywhere else, they are going to change the system for the whole country. They could maintain the existing system in Paris for a long time just by adopting the solution they chose in Corsica when they divided the single Département into two: the old 20 became 2A for one and 2B for the other. (The A and B don't actually stand for Ajaccio and Bastia, but it's a convenient mnemonic all the same). If they don't like 7A for Paris I think there are still some numbers in the nineties that are not used, for example I don't remember ever seeing a 97 or a 98.

-- Athel Cornish-Bowden (email)

When writing phone numbers it is important to make a distinction between the number proper (which in its full expression should be unique), and the number as it is dialed (which typically depends on the location from which one is dialing). To dial, one adds digits designed to indicate interurban/long distance/international access (or, in PBXs, simply the public network), or to indicate toll or other special charges. These digits vary from place to place: for example, in Spain the international access code is "00", whereas in the US it is "011" -- thus it is best to adhere to convention, using "+".

Thus spaces are needed for readability as well as to separate the area/city code from the main number. Separating the exchange (in the US: the first the digits of a seven-digit number, 555 in the example above) from the rest of the number is now quaint, as the letters it represented are in disuse (e.g., 256-1234 was AMHerst 1234). But introducing spacers to facilitate reading is also non-universal: in some languages it is customary to read numbers in pairs, or triples of single digits, or as numbers in the 00-99 range (as indicated above for the French case).

It may be best to use spaces to separate meaningful groups of digits, and periods for readability, according to local custom (pairs, triples). For example: +1 212 555.1234, or even +1 212 555.12.34 This also takes care of variable-length area/city-codes.

Unfortunately, Caller-Party-Pays for mobiles is sometimes mishandled, witness the case of Argentina. A landline Buenos Aires number may be written: +54 11 4555.1234, and from (say) Mendoza (within AR) it is dialed: 011 45551234, and from New York it is dialed 011 54 11 45551234

However, a mobile Buenos Aires number is: +54 9 11 4555.6789 (the 9 indicates a special rate), but from within Buenos Aires it is dialed 15 4555 6789 (15 indicates a special rate) and from Mendoza it is dialed 011 15 4555 6789. Then the ITU recommendation of listing "National/International" is inevitable.

As the ITU notes (see docs referenced above), this is made difficult by the fact that country-code "1" can be confused with the direct-dial long-distance access code 1 (the "trunk prefix"); consider 1 212 5551234. By the way, the prevalence of the digit "1" in both area/city codes and trunk access is rotary phone technology legacy. A 1 is faster to dial (at the telephone) and process (at the switch), as it is done by counting pulses. Larger cities received shorter-to-dial area codes: 212 for NYC, 213 for LA, 11 for Sao Paulo, same for Buenos Aires, so that most-dialed area codes were shortest.

-- Cris PM (email)

On the interrobang and the irony mark, two unusual punctuation marks, see

Along with the irony mark, perhaps a joke mark--more subtle than a smiley face--would be useful.

-- Edward Tufte

While it lasts, google for ‽. Nothing; not "Your search - ‽ - did not match any documents." Nothing.

-- Niels Olson (email)

On the other hand, interrobang yields .32 "Richard Feyman" Google result units (299,000/934,000).

-- Edward Tufte

In at least some major books during the late 1500s and early 1600s, the page numbers on individual pages (or, often, the page number which labeled a double-page spread) were enclosed in brackets. The Museum of Modern Art, in this month's calendar of events, revives the bracketed numbers in a table of contents. What for?

-- Edward Tufte

The changes to the UK numbering plan occurred five years apart, not two, there's a rather nice if redundant graphic at the Wikipedia page:

In general I would counsel information designers to stick with just spaces separating out the numbers and not introduce extra glyphs such as hyphens and points. (I think this is what the ITU spec means when it says: "Grouping of digits in a telephone number should be accomplished by means or spaces unless an agreed upon explicit symbol (e.g. hyphen) is necessary for procedural purposes. Only spaces should be used in an international number."

The reason this thread has struck a particular resonance with me is that I'm responsible for a website that captures telephone numbers from around the world and it's my job to deal with all the exceptional cases and try and work out what the ISDN version of each human readable number is. And I've just spent an afternoon fixing a piece of code becase US users insist on separating their phone numbers with hyphens.

-- J L Smith (email)

J L Smith wrote:

The changes to the UK numbering plan occurred five years apart, not two,...

Thanks for the correction. It just seemed like less, probably because two of the changes occurred during a six-year period when I was going to London rather often.

However, for your later point...

I've just spent an afternoon fixing a piece of code because US users insist on separating their phone numbers with hyphens.

... I have to disagree with your criticism, because it implies that humans need to learn to communicate in ways that computer programmers find convenient, whereas I think that computer programmers need to cope with the diverse ways of expressing the same information that people have always used. It annoys me a lot that some web sites require credit card numbers to be written as four groups of four digits, and can't understand 16 digits without spaces, whereas others require 16 digits without spaces and can't understand four groups of four.

An even worse example, which anyone who has bought iWork from Apple will recognize, is to insist that the user distinguishes beween 0 and O when entering the serial number, but to supply the serial number in a printed form in which 0 and O look exactly the same.

-- Athel Cornish-Bowden (email)

More UK phone number neepery

Thanks for the correction from J L Smith, who is of course right that the appropriate form for the new London numbers (and some other new city codes?) is XX XXXX XXXX and not XXX XXX XXXX. I've been corrected on this point before, so why do I keep forgetting? Probably because this fact means I have to deal with three different forms in my contact spreadsheets:


Although the third form is strictly correct, I find only a few people object if I present it as if it were of the second form; but then of course I forget that I was cheating, and come to believe the cheat is real :-)

Incidentally, while I mentioned previously that the first form strictly has no breaks in the last six digits, the presentation shown here above shows why I include the space: because this way, all my phone numbers, without exception, have ten digits and two spaces, leading to a neat right-justified appearance no matter which form, and no matter which type face, monospaced or proportional

(eleven digits counting the conventional leading zero)

ET probably would rightly point out that I am destroying information by doing so, since, if I had a ragged-right column, it would be a virtual bar chart indicating the location and prevalence of small town codes versus big city codes.

-- Derek Cotter (email)

Dropping the hyphen in "e-mail" seems an easy call. See Don Knuth

-- Edward Tufte

E-mail, e-mail, mél, courriel, mail, e'mail

It is interesting that most of the style guides AMA, APA, CMS
still seem to use the form e-mail. I remember hearing that it
is supposed to follow form of x-ray but I don't really
know why.

In France the debate is over mél or courriel

Typographers in France prefer mél to balance tél

tél  +33 x xx xx xx xx

-- Tchad (email)

The New Yorker and The New York Times, both with excellent typographic and editorial taste, use e-mail not email.

I like the way e-mail looks, and I don't pause over the word when reading it, so maybe Knuth's efficiency argument should not carry the day.

I forget what I did in Beautiful Evidence and don't have the book here in my hotel room.

-- Edward Tufte

In the Internet marketing of the 15th edition of CMS, the University of Chicago Press uses both e-mail and email.

-- Niels Olson (email)

Robert Bringhurst advocates email.

-- Steve Sprague (email)

One issue to consider is the ease of memorising a phone number when putting in spaces. Breaking the rules sometimes helps memorising numbers.

In Australia all fixed phone numbers are an 8 digit local number + 2 digit area code. There are only 4 area codes for the entire country so most people only ever use the 8 digit local number. Mobile phone and special service numbers are always 10 digits long.

The standard format is 08 9999 9999 for fixed an 0499 999 999 for mobile or special service.

But compare 9936 1361 to 99 361 361 for ease of memorising. Similarly a free call customer service number 1800 102 030 compared to 1800 10 20 30.

-- Andrew Nicholls (email)

For some reason to me which defies logic ("Military Intelligence?") is in the military the "Date Time Group" (DTG) are written in the bizarre format of

06 08:00 JUN 44 06 09:00 JUN 44 06 09:15 JUN 44

For those seperate times but the same date. I find it jarring every single time I read it. For me, the more obvious approach is from small to large numbers i.e.

08:00 06 JUN 44 UTC 09:00 06 JUN 44 UTC 09:15 06 JUN 44 UTC

This becomes easier to scan and read for me and everyone I talk to agrees but it is NATO military format.

-- Bill P (email)

I've found the usual ISO 8601 format to be best for general purpose, especially when sorting. It is a zero padded big-endian format:

YYYY-MM-DD hh:mm:ss

The spec requires a capital "T" instead of a space as the delimeter between the date and time parts, The time portion is 24-hr format. If you want to specify the time zone, represent it as an offset from GMT.

-- Matt Grimaldi (email)

The previous hyphen discussion on "e-mail vs. email" notwithstanding, do fellow contributors have any comments on shortening "e-mail address" to "eddress"?

-- Don Hottel (email)

Last I checked we didn't refer to postal addresses as pddresses. The word "email" or "e-mail" has two meanings to most people. It can mean email address as in "Whats your email?" or the actual message as in "Did you get my email?" In the former case, email is understood to mean "email address" while in the later it is understood to mean "email message"

-- Matthew (email)

How about the typography of punctuation adjoining web links? Specifically, if a link ends a sentence, like this: sample link.

I've included the period in the link following the rule that punctuation should be styled the same way as the text it is associated with. I believe that is the advice given in The Elements of Typographic Style, but I can't find the right page.

I've found two cases where I disagree with this approach: parentheses (do I include a closing parenthesis in a link?) and urls: My treatment of both cases makes sense to me, but I'm bothered by the inconsistency. Is there an accepted practice, and if not, is this a reasonable approach?

-- Rob Simmon (email)

As an engineer, using the decimal point in time punctuation (as in 7.00 am) strikes me as very misleading. The decimal point has a very specific meaning when it comes to numbers, and as its name infers, carries with it an implied base 10 system. Therefore, "7.30 am" could be misinterpreted as 0.3 hours after 7 am, when really it is 0.5 hours. The use of a colon (or space) avoids this problem.

-- Steve Levine (email)

Below, a good point about colons after headings from The Huffington Post. The fact that it is a heading for some upcoming words is already signaled by the title's location at the top, by the larger type, by the pattern of previous headings, and by the shortness of the line. No need for another signal, no need for a colon.

Sometimes these colons are the result of insecurities in writing mark-up language rather than doing typography the way it should be done (WYSIWYG)--since, in mark-up language, the architecture of the typography is not visible to the designer. I encountered endless colonized titles and colonized subtitles and colonized subsubtitles in trying to shape up IBM manuals (written in mark-up language) years ago.

Source: Decolonizing Slate

-- Edward Tufte

In France we use "h" instead of the ":" in time.

16h30 is your 4:30 PM, or rather 4.30 pm

We also put the day in front of the Month. This
removes the need for the comma and makes a more
logical progression of unit from smaller to larger.

16 mars 2007 is your March 16, 2007

Perhaps we could discuss it later, say 16h45 on 16 March 2007

-- Tchad (email)

The disambiguation hint is very good, a method natural to reading by humans.

-- Edward Tufte

In reply to Tchad. 16h45 16 March 2007 is not in lowest to greatest order. if you wanted that progression 45m16h 16 Mar 2007 would be the complete smallest to largest order. While this is somewhat pedantic, I finding it interesting that when speaking times, some people do start with minutes first.

While some might say "Four Fortyfive" others are likely to say "forty five minutes past four".

-- Matthew

The evolution of precision?

Well spotted - I'll get back in my box.

The treatment of time is strange isn't it. I wonder if
our "norms" have anything to do with evolving precision.

If we were using the sun, it might be good enough to say
that it is about 13h00 or that it is between 14h00 and 15h00.
If the minutes are unknown, then it may not be worth mentioning them.
Then again, perhaps we should look at look at arranging date or
time in decending order by the level of information embedded in
the word.

How would we order the following information to maximize utility:
Month of March
Day number 78
US week number 11
ISO or Europe week number 12
19th day of March
Year 2007
Between 5 and 6 o'clock
31 minutes after 5 o'clock

I suspect we might come to different conclusions.

-- Tchad (email)

Another cultural comparison for time type.

The form used in the Defense Department's messaging system, which dates back to the telegraph, is 1645R 16MAR07, where R is the time zone, counted from Zulu time. On navigation charts for restricted waterways, waypoints on the track are marked with a cross and a single line callout leads to the time of turn, down to the second, where the form is extended to 164530R 16MAR07.

-- Niels Olson (email)

Regards Matthew and Tchad's comments and discussions on "smallest to largest" when describing time display. Note the analogue clock; small "hand" is the hour, large hand, the minute. Therefore 16:45 is the right progression. <grin>

-- John Allred (email)

But a lot of analog clocks don't have a 16 on them, especially in the United States.

-- Bill Sharpe (email)

The timestamp format in the authoring interface of Movable Type blogging software starts with mostly right and gets more accurate, just like Arabic numeral system:

2007-03-28 10:45:28

-- Niels Olson (email)

The () indicates a part of the phone number that is either optional or situation-dependent. The (area code) might just as well identify the kind of service—cellular, fax, or land-line voice, for example—as a geographic area. But even that distinction is fading as we use our cell phones out-of-area. A phone with (212) can be anywhere in the country, not necessarily New York City. In many U.S. cities there are so many devices that it is required to use the entire ten-digit phone number as the dialing number.

Therefore the (area code) should not be treated as a separate element of the phone number and the () format should be dropped. In the seven years since my workplace began using a . separator for phone numbers on our Web site and business cards I’ve seen that usage become common enough to be considered normal, if not yet standard.

Is the . a period, a decimal, a separator, or a dot? We know by the context. (Really it’s simply a chunking mark no matter what the context.) In a timetable it’s recognized as a legibility/design element for the : separator between hours and minutes, not as a decimal. (In this context 7.30 is the same as 7:30; 7.30 does not even suggest a time somewhere around 7:20. One may like this or not, but one will most likely understand its usage.) In the context of engineering or math or money (7.50) it’s a decimal. As part of an IP number, email address, or URI, it is just a dot. In a sentence, it’s a period.

-- Chuck Pratt (email)

The ISO format for date and time is in use in most countries of Asia (Japan, China...). Its main advantage is that it follows the natural convention we use with other quantities, i.e. we start from the most significant piece of information and we finish with the least significant details.

So now is 2007-05-19 19:04:32.377 UTC.

The 377 milliseconds will be less relevant in 2 centuries from now than the fact that now was happening in 2007. Of course, the 377ms might be quite relevant in the context of a sequence of events, but that is not the point.

Note that the decimal dot (.) is used to indicate a true decimal point for seconds, while the column (:) is used to denote the base-60 separators for hours, minutes and second. Similarly the dash (-) separates years from months from days.

I must admit that the dot is less intrusive and easier to read though, and if all date stamps were always complete, we could easily write the previous time as 2007. On the other hand, the change in separator makes it easier to spot which part of the time is the date versus which is the time of the day.

-- jr bouvier (email)

Chuck Pratt is right that cell phones have weakened the link between geographic location and area code and have thus increased our need to dial them. I would further add that in an increasing number of American cities you have to dial the area code even when making local calls. Here in Boston I must preface a number with "617" even if I am calling across the street. Increasingly, the area code is thus demoted to a standard part of the number. It makes sense not to set off these debased codes with special punctuation.

This leaves us with several options: 617-555-1212, 617.555.1212 and 617 555 1212. I favor the dots, but I prefer all three to the now-inaccurate (617) 555-1212. The "617" is no longer optional for anyone; marking it as such is incorrect.

-- John-Paul Ferguson (email)

Steve Levine writes: "As an engineer, using the decimal point in time punctuation (as in 7.00 am) strikes me as very misleading. The decimal point has a very specific meaning when it comes to numbers."

Decimal notation is not tied to the dot. It varies from country to country and writing system to writing system; where I grew up, decimal notation used a comma, and the period was used in large numbers -- 1.000 denoting "one thousand", 1,000 denoting "one to three decimal places". Use in decimals is not a universally-agreed upon attribute of dots. Conversly, one could just as well make the complaint that the colon, in the context of numbers, is frequently understood to mean "ratio", or that "7:00 am" calls for an illegal division by zero!

-- Orbis Proszynski (email)

another formatting option...or not?

As a formatting for American numbers, what think you all of "123.456-7890"? I've been setting mine that way for some time. I believe it started as a way of cleaning out the parentheses, as I grew just graphically (though perhaps not typographically) conscious enough to start getting myself in trouble!

Being used to the distinct area code, I find it easier to read as a phone number. With matching spaces or even hyphens it looks like something else, a tax number or something. Using unique separators seemed to work nicely. (Although swapping them for "123-456.7890" looks wrong to me, too small of a pause in the middle of the 7-digit stretch now makes me read it as a math problem!)

-- n[ate]vw (email)

Orbis is of course correct that there is not, even within base ten notation, a one-to-one relationship between '.' and the boundary between the whole numbers and fractional numbers. However "7,30 AM" is equally grating to my eyes as "7.30 AM." Base ten time does exist; it was even the law in France once (albeit 200 years ago) and fractions of seconds are currently expressed in base ten, as a previous poster noted. Why use the same symbol to denote two different numerical bases? What's more, the use of the full stop strikes me as being a passing fad linked to the dot-com boom. But that might just be me.

As for the colon, it's third on the list of symbols that denote ratios, after the colon and the dash-between-two-dots, so the chances for misunderstanding are minimal.

I'm with Steve -- Let's restrict decimal point use to base ten and colons to base sixty.

-- Matt B (email)

Numbers are like words. We recognize what they are by their shape, not just their context. Messing with typographic conventions changes the expected shape, and can become a drag for the reader when it is done without justification.

Non-standard notation should only be used when it clearly makes things better (or when style trumps function). A printed railway timetable is obviously filled with clock times, so when tiny type on cheap disposable paper threatens to obscure figures through ink spread, and to reduce scannability due to clutter, using lining (monospaced) figures with the 0.00 form may make good sense. The vertically-aligned lighter space between figure groups helps align figures, reducing the need for column rules, while the period still visually joins them, allowing for a relatively tight, economical character spacing. When we look at the iPhone display twenty times a day, we know that's a clock at the top, so perhaps Helvetica bold is best decluttered there.

In other contexts, this kind of thing may not be such a good idea.

Messing with conventional punctuation in numbers can be as dangerous as it would be in prose. Imagine if we reduced commas to periods, and apostrophes to hyphens, relying on the context for their meaning.

Is the . a period, a decimal, a separator, or a dot? We know by the context.

But often there is no obvious context, especially when we're quickly scanning material in a web page or an email. 6:00, 6:00 pm, or 18:00 are clearly clock times, and not quantities (e.g., 1,800), dates (2000-01-02, or the no longer excusable 01/02/00), dollar amounts ($6.00), or batting averages (.180). Reducing meaningful punctuation to the status of anonymous separator is confusing: I have to read the numerals and fully parse out 2007. to figure out that it is a date-time string, and it is real work to extract just the time (um, is that 19:19 or 04:32 o'clock?). On the other hand, the identity of 2007-05-19 19:04:32.377 -06:00 is obvious and the hour can be read at a glance.

For U.S. and Canadian phone numbers, the surviving traditional form is (000) 000-0000. This is neither "inaccurate" nor accurate, but a typographic convention. For clarity, it should be used for North American audiences. For an international audience, the ITU form of +1 000 000 0000 is acceptable, where the plus sign flags a phone number.

(Although the area code is necessary to uniquely identify a number within the North American Numbering Plan, it is not "no longer optional for anyone". Where I live, it is always dialled for out-of-town numbers but never for local ones, and a single area code covers the entire province.)

Thanks to globalization, folks making business cards in North America have discovered international phone numbers, and they are playing around with typography, perhaps to appear more Euro-cool. This may be acceptable in the corner of a business card where we know that ten digits represent a phone number, but to start using a half- dozen different forms like 000 000 0000, 000.000.0000, 000-000-0000, or 000.000-0000 everywhere is diluting the convention, not following some new one.

Perhaps the hyphen was used in e-mail to differentiate it from email, a pre-Internet word for enamel, and it is appropriate for conservative editors to continue to do so. To me "e-mail" looks okay in fine print typography, but the wide hyphen in many screen fonts makes it look clunky on the computer display, and it can break awkwardly at the end of a line. Everyone including Oxford is now acclimated to the spelling "email" for electronic mail. How long until "mail" only refers to email, and not to something in a letter carrier's sack?

[Dear webmaster: editing HTML here is a touch frustrating because the preview function adds line breaks in Safari. Eventually, my paragraphs disintegrate. Thanks for listening.]

-- Michael Zajac (email)

As for russian phone numbers standards, I recommend an article A short history of telephone numbers by Artemy Lebedev.

-- Nick123 (email)

I'm curious if there exist any strong preferences or rules around the punctuation of professional designations, such as "M.D." versus "MD".

-- Michael Cusack (email)

Michael Cusack,

The Chicago Manual of Style states (?? 15.21):

In conservative practice, periods are added to abbreviations of all academic degrees (B.A., D.D.S., etc.). Chicago now recommends omitting them unless they are required for tradition or consistency.

It may be proper to use small caps as well:

In general, I don't think there's a "proper" way to do things any longer, as long as you're consistent in the method you use fits the over theme of your typography.

-- David Magda (email)

The interrobang is now the symbol for a new unit of measurement, the cuil. I'm guessing art economisting would probably be somewhere in the vicinity of 1.2‽

-- Niels Olson (email)

*** You can combine the two in the ITU standard shorthand

*** +44 (0) 20 7123 1234

The above number format is NOT valid.

The international format should contain only the digits that must be dialled after the access code.

It must not contain any digits that will cause the call to fail when dialled from abroad.

Since the 0 'trunk code' must not be dialled when calling the UK from abroad, it must be omitted from the international format.

Don't fool yourself that including the (0) in parentheses means 'exclude this when calling from abroad', because the ITU standard says 'items in parentheses should be dialled when calling from outside of the area, and omitted when calling from within the same area'.

That's describing area codes of course, but the standard is clearly disallowing the (0) too.

It also says 'parentheses should not appear in the international format'.

The London number can be written +44 20 7123 1234 OR (020) 7123 1234. Anything else is not following the standard.


-- Roger (email)

Could you tell me what is the typographic convention for duration? For exemple :

"Bratislava", video, colour, sound, 3'27"

or "Bratislava", video, colour, sound, 3 min. 27 sec.

Thank you for your answer.

A poor lonesome french graphic designer ...

-- isabelle jego (email)

Threads relevant to typography, book design, and publishing: