Science Of Spreed
Reading and Art
The standard layout of text on the printed page was not designed with reading speed
and comprehension in mind. The page began its evolution as an art form and page
layout mimicked art and architecture by using proportions such as the Golden Section
(a 14th century ratio that artists used to proportion their works). Generally speaking
it was pleasing to the eye. In the 20th century, these ancient practices were rediscovered.
In his 1928 book, The New Typography, considered one of the most important statements
in modern typographical design, Jan Tschichold argued:
Though largely forgotten today, methods and rules upon which it is impossible to
improve have been developed for centuries. To produce perfect books these rules
have to be brought to life and applied.
According to Tschichold, this revival of medieval typography formed the foundation
of the modern typographic grid which is ubiquitous today. Its reach has spread beyond
the printed page to online publications as well. Design director, Khoi Vinh, used
the typographic grid in the design of the New York Times online version.
While it is pleasing to the eye, the standard layout of text on the printed page
is a poor user interface in terms of reading speed and reading comprehension. This
should not be surprising as modern notions of interface design played no part in
the evolution of the printed page.
Neurological Compensations
There is a vast interdisciplinary body of scientific literature on reading which
amounts to a catalogue of the compensations and strategies the brain employs while
reading in an effort to adapt to the medieval artifact which is the printed page.
Eye movement research has established that reading a single page of text involves
approximately 400-500 eye movements called ‘saccades’. It also establishes that
there is no information perceived during the eye movement, only in the brief fixation
on a word or group of words. Saccades therefore slow down reading speed.
It is counterintuitive, but slow reading actually can have an adverse impact on
comprehension. Understanding a sentence or paragraph necessitates that information
be buffered in a person’s short term memory. One scientist has shown the impact
this can have on comprehension: “a slow reader's ... memory is taxed by the inability
to retain information in sufficiently large chunks to progress through a text with
adequate retention of the content.” It is sort of like traffic on a busy road. As
long as the cars move at a certain speed, traffic engineers have ensured that traffic
will flow efficiently. But as soon as one car slows down, the efficiency can be
prejudiced and traffic backs up. The brain has similarly been designed to absorb
information at a higher speed than the traditional typographic interface allows.
As the reader slows down, information “backs up” and can be lost.
We have to look at brain research to get a better understanding of how we read.Pelli
and Tillman demonstrated three processes that readers use to identify words:
Phonics (L) - a letter by letter sounding out of words. Contextual
Clues (S) - predicting upcoming words in text based on semantic and
syntactic content in order to skip them and save time. Holistic Word Recognition
(W)- recognizing the physical shape of words.
They were able to show that these three strategies for word recognition are additive
and also determine their respective contributions to the reading rate of the average
reader:
Average Reading Rate = 62%L + 22%S + 16%W
The preponderance of L which is the slowest of the three processes likely reflects
the fact that people learn to read phonetically and continue to apply this process
throughout life. Pelli and Tillman noted that the fastest readers in their study
had the highest contribution of the S process as they could effectively infer Information
allowing them to skip certain words.
So what does this mean for Spreed?
The inexorable shift of reading from print to computer screen (or mobile phone)
shows no signs of abating. More and more is read in electronic format, which creates
an opportunity now to define the new reading interface by leveraging the power of
the computer to focus on reading speed and comprehension instead of the medieval
focus on aesthetics. Processing power has been used to improve reading
speed before in two ways:
- Addressing the inadequacies in display technology
- Using screen based reading tools as training tools to increase print based reading.
Spreed is a new user interface for screen based reading. The patent pending Spreed
Algorithm parses text based on semantic and syntactic content to maximize contextual
inference (S) and then renders the text in a way that virtually
eliminates speed deterring saccades.
There are many estimates of the average reading speed. For paper based text, the
accepted range is 210-250 words per minute (wpm). Studies have shown that this speed
degrades by approximately 28% when reading directly from a PC screen. But the Spreed
application allows a first time user to read screen based text comfortably, and
with no loss of comprehension, at 350 wpm.
As the user acclimatizes to the new interface, this speed improves to 420~500+ wpm.
This has been correlated with an increase in Holistic Word Recognition (W)
with repeated use. These reading efficiency gains are even more dramatic for handheld
devices. On a handheld device, the screen size degrades reading speeds to approximately
110 wpm. Yet Spreed reading rates remains 350-500wpm.
Isn't it about time that someone tried to apply modern science and technology to
a 600 year old problem?