[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
Re: FromNess, ToNess
- To: <acad!blitzen!rick>
- Subject: Re: FromNess, ToNess
- From: Bob Perez <acad!xanadu!bobp>
- Date: Wed, 24 Jan 90 15:07:11 PST
- Cc: <xanatech>
I actually didn't like "1Space/2Space" either and, as you may have noticed
from the most recent draft of the XSOD, I've started using "LinkEnd", a term
similar to Joel's "anchor points", to describe terminating EndSets.
I share your appreciation of the orientation provided by "From" and "To", but
I think the appeal may be illusory. What happens when developers start using
4-ended links (which I predict will be quite useful)? "From" and "To" don't
map logically to any specific ends without some context.
Note that this isn't just a naming issue. The more fundamental issue concerns
the proper mindset for thinking about these objects. For the limited purpose
of directionality, I think of links as the cable in a data network,
interconnecting information that can flow freely in any direction, with
no abstract directionality. In a given context there will always be a "From",
and there may be several "To"s. Naming them in the abstract is an invitation
to needless confusion, no matter how well you've advised developers not to
give names any particular weight. You'd be absolutely astonished to know how
many times Developers have asked Apple's Tech Support group exactly what it
is that the Finder finds...
Ravi and I have started discussing this and I expect that what will happen is
that some messages will be changed to create a more general approach to the
retrieval of EndSets. N-ended links are useful enough and easy enough to
implement that we should probably expect them sooner than later. If so, then
we'll need a sufficiently general interface to LinkEnds, one that takes into
account n-number of ends, any number of which might be "Froms" or "To"s.