which one tries to represent
objects, places, and abstract
thoughts in image form. It’s
a hard game, and there is no
reverse game. Because information density is drastically greater
in pictures than in prose, a
picture is indeed worth a thousand words, but only when they
describe a concrete visual like a
graph or a portrait. In the realm
of the abstract, pictures fail. How
would you pictorially represent
Marxism? You could try a picture
of Marx, but that doesn’t distinguish between the person and the
school of thought (and requires
your viewer to know what Marx
looks like; otherwise it’s just a
guy with a beard). Words can
capture abstractions that pictures
cannot because language has
an immense amount of descriptive and differentiating power.
Abstract thoughts are exactly
represented by the words that
give them names. It is this power
that comes to the rescue in specifying functionality.
Standard GUIs, with their drop-down menus, check buttons, and
tree-lists, cannot compare to
the range of options that a text
interface effortlessly provides.
With just five alphanumeric characters, we can choose one out of
100,000,000 possible sequences.
And choosing any one sequence
is, in approximation, as fast as
choosing any other (typing five
characters takes roughly one second). It’s difficult to come up with
a non-text-based interface that
can perform as well.
Using language to access functionality brings to mind the old-form command line, which is still
one of the most powerful interface paradigms we have for controlling our computers. Although
command lines are hard to learn
Linguistic Command Line Interfaces
Two current programs attempt to deliver linguistic command-line interface to users:
Humanized’s Enso, and Blacktree’s excellent Quicksilver. Enso uses a more natural-
language syntax, and works like this:
1. At any time, the user presses an activation key to call up a text-entry area.
2. Next, the user begins typing what they want to do. For instance,
“translate to Japanese.”
3. As the user types, Enso autocompletes to the most likely command,
and related suggestions appear below the typed text.
4. The user either continues typing until the command desired is specified,
or arrows to a preferred command.
5. The user dismisses Enso, and the specified command is executed.
Enso then takes the selected text, uses the Google translation service, and places
the results back into the text. With just one implementation, the ability to translate to
and from languages is available anywhere on the computer, always with the same
interface, and accessible in a few mnemonic keystrokes. Enso uses copy and paste
as the graphical equivalent to standard out and standard in, allowing it to speak to
almost any application in an implementation-agnostic manner. Because of the power
of language, adding a large number of commands scales well. It’s always easy to get
to the functionality desired.
January + February 2008