Referencing parts of docs (was Including files)

John C. Mallery <JCMa@REAGAN.AI.MIT.EDU>
Errors-To: listmaster@www0.cern.ch
Date: Thu, 9 Jun 1994 21:20:00 +0200
Errors-To: listmaster@www0.cern.ch
Message-id: <19940609191804.8.JCMA@JEFFERSON.AI.MIT.EDU>
Errors-To: listmaster@www0.cern.ch
Reply-To: JCMa@REAGAN.AI.MIT.EDU
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: John C. Mallery <JCMa@REAGAN.AI.MIT.EDU>
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: Referencing parts of docs (was Including files)
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
    Date: Thu, 9 Jun 1994 13:49 EDT
    From: "Alan (Miburi-san) Wexelblat" <wex@media.mit.edu>

    > It would be extremely desirable (but hard) to be able to resolve URL region
    > fragments without fully accessing the document.

    Indeed, it is quite hard under the current scheme.

    > One idea is positional specs like:
    >
    > #{char=24},{char=40}
    >
    > #{h1=1},{h1=2}

    But people don't (generally) want to jump to arbitrary points in text --
    they want to jump to points which have *meaning* and that meaning can often
    be expressed within an encompassing scope.

    ...

The point of have named position locator types is to allow them to be extended
trivially to other available naming conventions.

Of course, the naming junk you add the more resolution problems you create.

Best to have KISS first, then full NLP second.