Re: 1.4 c: Basic Terminology

The name "terminus" for one of the "things that are pointed at" seems
OK to me.  But I'll make another suggestion, based on the feeling that
terminology like "terminus", "target", "end", and "link end" is too
termin-centric; it's taking the link element's point of view, that the
important feature of these "things that are pointed at" is the fact
that they're being pointed at.  That's certainly true from the
link-machinery point of view, but not in terms of what the document is
trying to do.

I'd suggest "selection" as a possible alternative, because it focuses
more on the picking-out-and-relating function than the arrival-here or
pointer-points-here or address-maps-to-this functions.  It is blander,
but one problem with end-oriented terms is that they're too specific:
linking may not be end-oriented.  Suppose what you want to do is
display two texts in parallel: are you really thinking in terms of
"traversal" or of "going" from one text to the other?  The viability
of "terminus" lies in the fact that you can sort of think that way
about any linking; but it's also not the mot juste because what you're
really trying to do with a link may be more complicated than, or just
different from, what the railway metaphor implies.

"Selection" also has some familiarity from its use in the jargon of
some current computer systems.  This is an advantage in that the usual
sense does not conflict with what we want it to mean in
(Hyper-)?(X[HM]L-)?(Link)?: in particular, a selection does not have
to map exactly to the full content of one or more SGML elements, and
indeed it needn't be text at all.  But there could be some possibility
of confusion as well, of course (for example, in the way you often
can't have more than one selection active or even defined at a time).

The problem I've described with "terminus" is a problem with "pointer"
as well; but that feels more acceptable to me because it's more of a
behind-the-scenes thing.  For example, as a document author, you can
think about linking some termini without having to work out how
they're pointed at or addressed within the document's innards; that's
the part that a program could do for you (in response to your clicks
or pointing at the screen or whatever).  When document authors talk
about this stuff, I expect they will often talk about links and
termini, and not much about pointers and addresses.  (And even the
stuff about links will be in a more abstract sense than what we've got
in the draft: it's probably more focused on a "link relationship" than
on a particular link element.)  I expect they will also have reason to
often point out to neophytes that "terminus" has a more subtle meaning
than "endpoint as in the place where an HTML <a> link takes you".

John Lavagnino
Women Writers Project, Brown University

Received on Tuesday, 4 February 1997 17:16:03 UTC