Off-topic: Magic Number 7 (WAS: Re: XML Editors - Word 2000??)

Rick Jelliffe ricko at
Sat Jul 3 09:38:36 BST 1999

From: Eric Bohlman <ebohlman at>

>On Fri, 2 Jul 1999, John E. Simpson wrote:
>> Couldn't find any mention of AT&T or Bell Labs (or for that matter,
>> occurrences of the string "phone" except in "phoneme") in the HTML
>> of the paper that Dan pointed to. That I heard of this 7-more-or-less
>> theory as "belonging to" AT&T about 20 years after Miller's paper was
>> published possibly reflects the embrace-and-extend philosophy: "Not
>> invented here... but if it's useful, we can say it was!"

Thanks to John S., Dan B. and all, for the references and info.

>Note, BTW, that the "7 items" theory refers to the limits of *working
>memory* only.  In particular, it is completely irrelevant to tasks
>involving picking a single item from a displayed list; such tasks
>only *one* item to be held in working memory, regardless of how long
>list is.

Miller uses "span of absolute judgement" and "span of intermediate
memory". I think these are applicable to GUIs and to structured editing,
but only with care.

For example, many structured editors give a context-sensitive menu:
you cannot use position on the menu to locate an item because it
changes. So you have to scan the list to find the information, and you
might have to figure out "which of the available choices are relevant
to this document?"

But if the list is scrolling, then we are using the list differently: we
are not
choosing from many, we are saying yes/no to a presented choice. And, in
any case, the stimulus from words may be a graphical stimulus not a
letter-based one: we may remember things like "top" or "bottom" or
"next to the menu line" or "has a little icon": this is what Miller

For GUI work, other factors such as the physical difficulty of the tasks
(Fitts' law's Index of Difficulty) and reading criteria such as the
of eye fixations also play an important part.

But the question is not "Do Miller's ideas apply here?" but "Why are
structured editors easier to use when the DTDs have a 7-ish number
of elements available in any context?"

All we need from Miller, Fitts, span of attention studies, eye fixation
studies, are rules of thumb. If  we take the magical 7 as one such rule
of thumb, then it has implications for GUI design of structured editors,
DTDs and workflow:

* structured editors should provide some other dimensions (glyphs,
colours, 2D, nested menus, shortcuts, "recoding") apart from the
scrolling  list for large,  loose content models;

* If an element type contains more than 7ish elements in a loose
structures, DTD designers may decide to favour generic elements
type to reduce the number; if their are fewer than 7ish the DTD
designer may decide to favour ungeneralizing some element
types (i.e., <ul> and <ol> rather than <list type="ul"> and
<list type="ol">)

* mangers may decide on a division of labour in which each
operator only deals with on 7-ish branch (i.e. a table
entry operator, a cross-reference and indexing operator, metadata
entry operator, a text entry operator).

In Miller's terms, I think that selecting element types by navigating
through style information (created by inverting the stylesheet) may
be a useful "recoding".

Rick Jelliffe

P.S. Here are some key quotes from Miller:

"Everybody knows that there is a finite span of immediate memory and
 that for a lot of different kinds of test materials this span is about
items in length. I have just shown you that there is a span of absolute
judgment that can distinguish about seven categories and that there is a
span of attention that will encompass about six objects at a glance.
What is more natural than to think that all three of these spans are
different aspects of a single underlying process? And that is a
mistake, as I shall be at some pains to demonstrate."

His conclusions include:
"First, the span of absolute judgment and the span of immediate memory
impose severe limitations on the amount of information that we are
able to receive, process, and remember. By organizing the stimulus
input simultaneously into several dimensions and successively into a
sequence or chunks, we manage to break (or at least stretch) this
informational bottleneck."
"Second, the process of recoding is a very important one in human
psychology and deserves much more explicit attention than it has
received. "
"Third, the concepts and measures provided by the theory of information
provide a quantitative way of getting at some of these questions."


'We give the observer as much time as he wants to make his response;
we simply increase the number of alternative stimuli among which he
must discriminate and look to see where confusions begin to occur.
Confusions will appear near the point that we are calling his "channel
capacity." '
'Let me summarize the situation in this way. There is a clear and
definite limit to the accuracy with which we can identify absolutely the
magnitude of a unidimensional stimulus variable. I would propose to call
this limit the span of absolute judgment, and I maintain that for
 judgments this span is usually somewhere in the neighborhood of seven.
We are not completely at the mercy of this limited span, however,
we have a variety of techniques for getting around it and increasing the
accuracy of our judgments. The three most important of these devices
are (a) to make relative rather than absolute judgments; or, if that is
possible, (b) to increase the number of dimensions along which the
can differ; or (c) to arrange the task in such a way that we make a
of several absolute judgments in a row. '

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list