created Sunday, 2013-01-27 T 20:35:17Z
updated Sunday, 2013-02-03 T 18:25:28Z
Two problems with hyperlinks as we use them is that
they are only where the document's author thought they would be useful
& that they lead to where the author thought would be useful, nowhere
else. In fact, there are many reasons a person might
select a word or phrase. They include:
navigation: It's the name of a section in the
current or another document, & the person wants to read that now.
This is the automated version of slipping through the pages of a book
until one finds the page number that was indicated in the table of
contents. The links we have now do this.
reverse navigation: What links here or near here?
definition: Person wants the dictionary definition
of the word.
gloss: Read annotations attached to this word or phrase.
index: Where else is this word or phrase used in
this document? In any document?
cross index: With what other words is this word
used, & where?
edits: View other versions of this part of this
document.
more detail: Is there a more detailed discussion of
the phrase, paragraph, or section I'm reading? (Depending on how a
document is written, this feature can be implemented as navigation.)
We needn't limit selections to text...
3D image: Zoom & rotate a scene.
image in context: If it's a painting in an art
museum, see it where it's hanging.
animated edits. Run forward & backward through the
edits that have been applied to this document.
alternative representations: Braile? Audio? Word
frequency analysis? As a sequence of formally state logical propositions?
document's context: Other documents by this author?
Where this document fits in the timeline of documents by this author?
By anyone? How common is this topic, & what other documents address
this topic?
We can do some of these things with HTML links or
with modern search engines, but not all of them, & even those we can
are not done well.
Implementation ideas
Can we achieve these features with the link feature of existing HTML,
maybe by following some authorship conventions or with some automated
pre-processing? Or do we need another technology?
With drop-down context menus, a user could select a word or phrase
then select an item from the menu. There could be some standard items
such as definition, find elsewhere in this document, search the web, but
with current HTML, the author still must predict the words or phrases
that will interest the user. Also with current HTML, it'd be difficult
to distinguish between selecting a phrase or selecting a word in it.
What's more, I think such drop-down context menus
would require active content (Javascript).
People that have thought this before
At least gloss & cross index are part of Vanevar Bush's vision of
memex. This is clear when in
"As we may think",
Bush says:
When the user is building a trail, he names it, insets the name in his
code book, and taps it out on his keyboard. Before him are the two
items to be joined, projected onto adjacent viewing positions. At the
bottom of each there are a number of blank code spaces, and a pointer is
set to indicate one of these on each item. The user taps a single key,
and the items are permanently joined.
(You have to overlook the implementation & user
interface details that Bush mentions.)
As of 2013, selection is done by clicking with a cursor that is
controlled by a mouse, backball, or game controller, or by touching
with a finger. I say "select" because I don't know what
user interface the future will use, & I don't want to imply that
my idea is specific to current user interfaces.