Some preliminary remarks. I think we can all agree that there can be no universally agreed-upon optimal amount of "typographic color" or "overall grayness" of a page of text. What some might consider to be "good color" may strike others as unpleasantly light, or too dark. Nevertheless, for the sake of the argument, let's assume that (i) there is an optimal amount of color and (ii) the settings for 10pt Computer Modern set "single-spaced" achieve this optimal amount. (By the way, this does not imply that the optimal distance between consecutive lines is 10pt for text set in CM. The topic of what exactly constitutes "single-spacing" is a matter for a different discussion.)
The question then becomes: how does one achieve this amount of color when switching to a different (text) font, such as Palatino? Palatino and CM obviously differ in many respects. Not only are their x-heights different, but so are their cap-heights and ascender heights as well as the average stroke width (to name just a couple of additional factors). The upshot is that if text that consists of a group of paragraphs is set both in Palatino and in CM with the same nominal point size (say, 10pt) and the same interline distance (say, 12pt), the paragraphs and pages set in Palatino will look noticeably darker. I think most will agree that the Palatino paragraphs will be "too dark" -- again, assuming that the text set in CM have "correct color".
- Who exactly came up with \linespread{1.05} for Palatino? The authors of l2tabu or someone else?
- Is the factor of +5% just a rule of thumb, or is it the result of some (unknown to me) established typographic formula?
Regarding your first question: I have no idea who was the first to come up with this recommendation. Most likely, though, the need to increase the "leading" when typesetting text in Palatino was "discovered" immediately after people started mixing/matching Palatino and CM.
The +5% recommendation you noted in the l2tabu
document can be nothing more than a rule of thumb. My understanding is that it should be understood as a "lower bound" on the required line spread adjustment, rather than as a single number. The code below generates two half-pages of text; the upper half is in Latin Modern
, the lower is in Palatino nova
with a linespread factor of 1.05. (Both texts are set with nominal font sizes of 10pt, and the text itself is the usual lipsum stuff...) To my eyes at least, the text set in Palatino still has more color than the CM reference text despite the 5% adjustment. Playing with the line spacing setting, I'd say that ca. 7% achieves equal color.
% !TEX TS-program = xelatex
\documentclass[letterpaper,10pt]{article}
\frenchspacing\pagestyle{empty}
\usepackage{fontspec}
\usepackage[margin=1in]{geometry}
\usepackage{setspace,lipsum}
\begin{document}
\setmainfont{Latin Modern Roman}
\emph{Latin Modern}
\lipsum[1-3]
\vspace{1in}
\setmainfont{Palatino nova Regular}
\setstretch{1.05}
\emph{Palatino, linespread 1.05}
\lipsum[1-3]
\end{document}

palatino
package), i.e. Walter Schmidt.\textwidth
, the larger the\baselineskip
. Also the larger the x-height, the larger the\baselineskip
. The combination of line width and leading should make sure 1) the reader can find the next line easily 2) the reader can follow the current line without distracted by following lines. The relations between x-height and leading is more arguably. Some types are designed to be "economically", they look not too bad if you set them in long lines and relatively small leading, while their x-height is also large.\linespread
method is too coarse when dealing with size changing problems: it's not automatic that, with Palatino set 10/12.6 for the main text, it's right at 9/11.6 for quotations (\small
) or 8/10.5 for footnotes. But that would require writing a very detailed specification based onsize10.clo
.