In article <1991Aug1.142641.17
...@HQ.Ileaf.COM> jhar
...@HQ.Ileaf.COM (Jerry Harris x1586) writes:
>My question is: does anyone know of any papers,reports,etc on this
>subject? Specifically, problems relating to layout of objects (text
>and graphic) on the electronic display, hyperlink address indexing and
>resolution, SGML support for electronic layout, etc?
I have thought about the problem of generating links among documents,
especially when documents are created on separate systems and then
collected into a single database. The problem is how to get the
documents to link without tediously reading everything looking for
potential links.
The method I hit upon is to use what I call 'half links'. Instead
of directly linking documents, each document points to a common
subject table. Two documents that point to a common subject then
implicitly are linked, even if they were created independantly.
An example implementation would use a standard subject list. The
Library of Congress Classification System is a useful one to select.
As you create a document, you can give the entire document a
classification by subject (this is what librarians do with books),
but in addition, you can insert pointers to LC numbers internally
in the document wherever you want a subject pointer. When a set
of documents are collected onto a database, the computer can go
through all the documents once and build a table of subject pointers.
Then when you are reading one document, you can highlight exact
subject matches by simple table lookup. Another thing you could
do is loosen the tolerance on a match -- in effect asking "what's
in my database that's close to this subject".
Any standardized list of intermediate pointers could be used, as
long as there is widespread agreement on the list and on the format
for calling out the pointer in a file. The LC system is particularly
useful IMHO, since it automatically also references all the material
in paper libraries that has been so catalogued. The Library of
Congress is busy making all their catalog material into machine
readable form (magtape and CD ROM), so eventually you will be able
to ask at any given subject pointer "list all books on this subject",
as well as asking what is immediately accessible in your database.
The National Union Catalog is the mathematical union of the catalogs
of the 1000 or so largest libraries in North America. It thus
tells you where all the copies of a given book are located. When
this is available in machine-readable form, you could go from a subject
pointer, to calling up the nearest location of a specific book, to
dialing up that library and asking for a FAX, FedEx, or Inter-Library
Loan, all in a few minutes.
My questions for the newsgroup are:
(1) Are there published discussions of this type of link method?
(I am not aware of any, but I am not a hypertext expert either)
(2) What other intermediate tables could be standardized on?
Examples:
Dewey Decimal Classification System
Calendar Date
Unabridged dictionary
Encyclopaedia Britannica Propaedia and Index volumes
Britannica Great Books of the Western World 'Syntopicon'
(the latter two are in effect crude paper hypertext systems. The Propaedia
is a subject classification of the encyclopedia contents. The Syntopicon
is a subject outline with pointers to a set of the classics. They suffer
from poor backward links from the text, you have to already know what
subject you are reading to look up related items.)
Dani Eder
--
Dani Eder/Boeing/Advanced Civil Space/(205)464-2697(w)/461-7801(h)/#905, 1075
Dockside Dr.,Huntsville,AL35824/Member: Space Studies Institute
Physical Location: 34deg 37' N 86deg 43' W +100m alt.
***THE ABOVE IS NOT THE OPINION OF THE BOEING COMPANY OR ITS MANAGEMENT.***
