Re: [Groff] explicit hyphen and numbers

[Ted Harding]

On 10-Aug-06 Steve Izma wrote:
>> From: Keith Marshall <address@hidden>
>> Date: Thu, 10 Aug 2006 14:28:29 +0100
>> 
>> I won't!  If I write a numeric range, such as `200-400',
>> I *don't* want a line break to occur within it.
> 
> Well, I'll disagree. I can't find any reference to this in either
> "Words into Type" or the "Chicago Guide ...", but when setting
> indexes on short lines (e.g., two columns on book page, which
> gives you about 10 to 12 picas, usually indented), any place
> where you can get a line break is very important. By the way,
> most style guides recommend using an en dash here, which is
> subject to the same .hcode and .cflag issues as a hyphen. So I'm
> keen on the idea of having another .cflag code to increase our
> options in such situations.
> 
> Even in the middle of a regular text block, I don't think anyone
> is going to confuse a range of numbers broken at the end of a
> line with a hyphenated long number, which essentially is never
> needed.

Yes, that does make something of a case for it, and indeed it
does occur. But even so I think it can be avoided in most cases.

Fr example, I just checked in the index of a book "Multivariate
Analysis" (Mardia, Kent & Bibby, Academic Press 1979).

In one entry I see:

seemingly unrelated regressions, 203-
      5, 211

and in another:

simultaneous  confidence  intervals,
      144-5


(the interword gaps in the book being relatively wider than
shown in the second example).

Both cases could have been improved in appearance by not filling
the lines:

seemingly unrelated regressions,
      203-5, 211

simultaneous confidence intervals,
      144-5

especially since some index entries are short lines anyway,
leading to a ragged-right effect overall -- e.g. (and I've
chosen a series which mixes the effects):


canonical correlation analysis, 5, 281-
      99
  contingency table, 290-3, 299
  discriminant analysis, 320, 230
  prediction, 289-90
  ridge technique, 298
  score, 289-90

which I reckon would look better if both the range-break
and the filling were turned off, giving

canonical correlation analysis, 5,
      281-99
  contingency table, 290-3, 299
  discriminant analysis, 320, 230
  prediction, 289-90
  ridge technique, 298
  score, 289-90

What opinions do others have?

Best wishes to all,
Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <address@hidden>
Fax-to-email: +44 (0)870 094 0861
Date: 10-Aug-06                                       Time: 17:24:32
------------------------------ XFMail ------------------------------