On Oct 24, 2006, at 3:17 PM, Noel Stoutenburg wrote:
There is, though a third choice: that some fonts had fixed width "1"s, while others had proportional width "1"s, and if the typographer setting up the page knew that there would be a need to maintain a strict vertical alignment of columns, he chose a face where the "1" was the same width as the other digits.
That's what we did on the phototypesetting machines used in the 1970s and 1980s. There the "font" came in the form of a memory cartridge inserted in a slot (for the widths) and a negative filmstrip to be attached to a drum (for the images).
In many typefaces we had more than one font, one a text cut and the other a display cut. The text cuts always had same-width numerals. They were not only the same width as each other, but reliably the same width as numerals in any other font, always one en (ie, one-half the point size). Certain other characters were always one en wide: notably the en-dash and the en-space, but also the bullet and the leader dot (but not, curiously, the letter N itself). In a display cut, the numerals might be any width. In some typefaces the difference was dramatic. In Avant Garde, for instance, the zero is flattened and the 1 has a wider serif in the text cut, while in the display cut the zero was round like the O.
Numeral width wasn't the only difference between text and display cuts. There were other differences in styling and kerning. The basic idea was that display faces were expected to be used in titling and probably at larger sizes, while text cuts were expected to be used for blocks of text.
As I noted, a number of faces (according to the typeface references I consulted, _do_ have a fixed width "1".
In the phototypesetting days, I don't think there were any non-display fonts that had variable-width numerals. My impression is that the phototypesetting fonts imitated the lead fonts that preceded them. From what my brother tells me, I gather that fixed-width numerals were standard in any text font, with the variable-width numerals used only in display fonts. I'd be curious to see if anyone can find non-fixed-width numerals in a text block in any font from about 1850 to 1950.
As I gave more thought to this problem, attempting to consider it as I imagine a typographer would have done, I realize that the number of cases the 1 has to be the same width is really somewhat small. It is not necessary, for example, in a table of contents, or in an index, or in many other types of tables. In fact, it is only necessary in circumstances where there is a need to add a column, such as in an accounting table.
There's more to it than that. It's certainly useful when you've got a numbered list. In today's desktop publishing, the numbers are automatically tabbed off, but in the old days, it was much more hassle if the numbers weren't the same width. Same idea for numbered outlines, or even paragraph labels with a series of numerals. For instance, if you're setting a legal text where the paragraphs are numbered 101.5.01, 101.5.02, 101.5.03, etc, it doesn't look as good with irregular numerals.
Page numbers look fine with variable widths if they're centered, but if they're flush to one side or another, the associated inconsistency of the non-flush margin is aesthetically displeasing, I feel. (And I've seen that in recently published books.) I disagree with you about the table of contents, too. I think they look bad with unequal-width numerals. This was even more an issue in the old days, when the dotted line had to be created with a string of leader dots; if the numerals weren't the same size, the dots would be misaligned and it'd look like crap.
mdl (who was a typographer before he was an engraver) _______________________________________________ Finale mailing list [email protected] http://lists.shsu.edu/mailman/listinfo/finale
