Have any blind trials been organized to test to which extent "the average reader" prefers paragraphs typeset by the Knuth-Plass algorithm (or more advanced variants) over those typeset using the simple greedy algorithm? See the Wikipedia article on word wrap for more information.

I'm mainly interested in trials that test longer texts as opposed to (contrived) single-paragraph samples. In addition to the preference (binary) of the reader, it would be very interesting to compare the read speeds for the differently typeset texts.


2 Answers 2


Leyla Akhmadeeva and I (Boris Veytsman) made several experiments trying to determine whether typesetting niceties influence reading speed and comprehension. Some of them are described here:

Our experiments were inspired by the great paper by Chuck Bigelow and Gordon Legge Does print size matter for reading? A review of findings from vision science and typography.

Basically we found that despite typographic lore there is no much difference between serif and sans serif fonts at the same font size and similar shapes. This corresponds to Legge and Bigelow's finding that size does not matter (as long as it is within the "comfort zone").

We did perform some experiments with justified and ragged right texts (we did not try different justification algorithms). We did not gather enough data yet to be sure, but it looks like the difference, if exists, is very small.

My opinion is that typesetting is driven mostly by aesthetics rather than by actual differences in reading speed or comprehension. Human brain is very flexible and can read almost anything. Or, as Legge and Bigelow think, over the years of typographic history the really unreadable styles just died out, and everything that survived is basically equally readable.

  • 34
    +1 for "Huamn brian is vrey flixeble and can raed amolst anihytng".
    – Fran
    Oct 23, 2014 at 8:48
  • 7
    +1 Do you know of any experiments with dyslexic subjects? We are frequently told that the following assist students with dyslexia: (1) using sans fonts only, (2) not justifying text at all, (3) avoid italics completely. I am curious whether there is any evidence either for or against this for longer texts (i.e. not just things like street signs where 1-3 strike most people as unsurprising).
    – cfr
    Oct 23, 2014 at 18:33
  • 2
    @cfr We experimented with post-stroke patients, but we selected those who still could read easily. Never tried to work with dyslexic subjects (yet?)
    – Boris
    Oct 23, 2014 at 23:46
  • 1
    @cft Gordon Legge worked on reading by visually impaired subjects, but I do not remember anything about dislexic ones (thanks to Karl Berry for reminding me this. At the end of the day my memory works slowly...)
    – Boris
    Oct 24, 2014 at 1:37

Research on whether readers prefer Knuth–Plass over the simple algorithm is not going to be as prevalent as research on whether readers' comprehension and speed is improved by it.

The studies linked-to by Boris are good examples of the latter; I know of no specific studies on the former.

  • 1
    You talk about liked studies, but there are no links in your post. Did you forget them? Jun 12, 2015 at 20:29
  • @HenriMenke I take it the reference is to the studies linked in Boris's answer.
    – cfr
    Jun 12, 2016 at 13:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .