Re: forwarded... [Explicit Linking is Impossible]

Tony Sanders <sanders@bsdi.com>
Errors-To: sanders@bsdi.com
Message-id: <9305052334.AA01420@austin.BSDI.COM>
To: www-talk@nxoc01.cern.ch
Subject: Re: forwarded... [Explicit Linking is Impossible] 
In-Reply-To: aronsson@lysator.liu.se's message of Wed, 05 May 93 23:24:26 +0200.
Errors-To: sanders@bsdi.com
Reply-To: sanders@bsdi.com
Organization: Berkeley Software Design, Inc.
Date: Wed, 05 May 1993 18:34:40 -0500
From: Tony Sanders <sanders@bsdi.com>

aronsson@lysator.liu.se's message of Wed, 05 May 93 23:24:26 +0200.
> might have.  My limited experience of writing HTML documents and
> converting old plain text documents to HTML makes me agree with Eliot
> Kimber's conclusions.
...
> automatically.  But manually keeping track of hypertext links between
> a text that I maintain and another text that someone else maintains,
> would be next to impossible.

Use indirection, make the anchors say "the part of the foobar manual about
blat" and then have a database that translates them to actual HREF's and/or
computes them on the fly.  There is no magic wand or silver bullet to
solve these problems, but you can make it easier to maintain when something
goes away.  In the local arena you can use backrefs to manage changes,
not so easy with global documents (though DOCid's and HTTP2 will solve
some of the problems).

There is nothing wrong with adding a layer of software beteen HTML
and whatever documents you are storing to make life easier.

Part of problem lies in the approach.  Hypertext is designed as a
distributed information management system (how else can you expect to put
a yottabyte of data online :-), and they are trying to do everything local
and then publish this whole big thing.  No doubt they are having problems.
No doubt that the problem they are trying to solve is hard.  The fact that
link complexity is non-linear isn't a new problem.  Every system that has
relationships has this problem, including human interaction.  This is the
big company/project problem and you can use some of the same techniques
to minimize its impact.

In summary, I agree that massive explicit linking is impossible, but it
isn't desirable either.  I like what Edward Vielmetti, just said about
it, he sums it up better than I.  Yes, there is a need for tools in this
area, keep in mind this is a VERY young technology (you can tell becuase
hypertext still means something, unlike older technologies like
Object-Oriented programming which have ceased to have any real meaning).
To me it's not hypertext if it's not interactive (authoring tools!) and
we aren't quite there yet.

> Thanks, Marc and NCSA, for X Mosaic.  Is there an NCSA HTML editor
> too?  Today I have GNU Emacs.  I would like two Mosaic windows, and to
tkWWW has editing functions (you can add links on the fly) but it needs
work before it's "production" quality.  The problem is that HTTP2 isn't
in place yet so you can't send the changes back to the server anyway.

An ex-IBMer also (moved 4 times in three years) ;-)

--sanders@bsdi.com