First Class Links Make Second Strike

[Mark S. Miller <mark>]

Date: Thu, 19 Oct 89 15:23:58 PDT
   From: marcs (Marc Stiegler)

I think the presentational problems with embedded links are
interesting and unsolved, and I don't have much to suggest there.
However, I will reason from my intuitive notion of what an embedded
link *means*, and assume someone will eventually (yes, most probably
after shrinkwrap) figure out a presentational system that does justice
to my intuition.  Note that this is not an argument for embedded links
in our first generation front-ends.

As many of you have heard me say, a link should be though of as a
bundle of meaning in much the same way as a sentence.  Just as very
few documents consist of a stand-alone sentence, so will very few
documents consist of a stand-alone link.  A traditional document is a
structure consisting of a bunch of little statements about the world--
sentences--which are meant to be coherent when taken together.  A
Xanadu document is a bunch of statements about the world--sentences
and links--which are meant to be coherent when taken together.  From
this starting point, and in the absence of any experience of use, we
can (I claim) frequently reason about the proper place of an
individual link of a document by making an analogy with the more
familiar individual sentence in a document.  I will reason from this
analogy below.

Btw, the most vivid fact to me is our lack of experience with
hypertext as a form of literature.  The issues we're raising are
really those of literary conventions, which *must* undergo substantial
evolution after we create this medium, and whose evolutionary results
we cannot forsee.  Our goal should be to make as *evolvable* a medium
as we can, while seeding that evolution with something that is
initially good enough to be wonderful.  In general, we have all been
collectively terrified of creating new Qwerties (the plural of
Qwerty?)  I have been among the most terrified.  However, in this case
I am not so fearful.  As long as the docs & links layer stays the
same either way, what me worry.  This isn't one for which it is
*possible* for us to get it right the first time.  It doesn't matter
how smart we are.

   Unfortunately, whether the link is embedded or not, FINDING THE 
   RIGHT CONTEXT IS STILL AN UNSOLVED PROBLEM. In the scenario I 
   created to ask questions earlier, we had 4 berts, A1-A4, and 
   3 links, to A1, A3, and A4. A2 didn't have a link embedded in 
   it because it had branched off before we made the link. 

   Question for the reader: Was A2 left without an embedded link 
   because of the careful, methodical planning of the creators of 
   A1 and A2 and A3? Or was it left without a link because the author 
   didn't know about B until after A2 had spun off, or because the 
   author forgot that B was there until A2 spun off, or because 
   B was created after A2 spun off? If the cause were any of the 
   reasons except Cause Number 1 above (Number 1 being the careful 
   intentional planning of the author, the cause which I honestly 
   find the least likely), A2 is just as interesting as an endset 
   as A3 and A4.

Reasoning by analogy with sentences: "What if we have a document
(bert) A1, and we fork off a new version A2.  Then we go back to A1
and fix some typos and insert a new sentence.  We now fork off (from
A1) versions A3 and A4.  Was A2 left without the new sentence (and the
corrected typos) because of careful methodical planning ..."  I think
transposing the example this way pokes a sufficient hole in this
argument against embedded links.  In any version control system,
changes will accidently get left out of some versions.  Not being able
to read the user's mind, the best we can do is to try to let it be
always clear what the consequences of making a change to what are, and
provide tools like version compare through markUp diagrams for
manually merging divergent changes.  If there's a better way to manage
version control of links, could that solution conceivably apply to
sentences?

   ...

   Dean, if you aren't overwhelmed by my reasoning here, we probably 
   need higher bandwidth, i.e., let's nuke it out in the office.

Whew that's a lot of stuff.  After you guys nuke it out, please post
the resulting thoughts.  I can't tell you how happy I am that this
whole controversy is happening above the docs & links layer!  Good
luck.