Open Hypermedia Systems Working Group

A Southampton scenario for OHS

A Southampton Scenario for OHS

Wendy Hall and Hugh Davis


Extracted from the video script of "Making Links", inaugural lecture given
by Wendy Hall on 28th January 1997

The scenario begins with the film 2001: A Space Odessey. It is the vision
of building a real-life version of HAL: the ship-board computer in 2001. In
fact, according to the book of the film, HAL became operational on 12th
January 1997, which is pretty timely as far as this scenario is concerned.
In today's terminology HAL would be called an "intelligent agent". At
Southampton our work on open hypermedia systems is evolving into the
development of  communities of "intelligent" agents that can do some of the
things HAL could do, in particular to help us interact with digital
information. This scenario is illustrated by an imaginary dialogue with AL,
a cut-down version of HAL!

Imagine that AL has access to vast banks of digital information.

Our Motivation - the digital version of the Mountbatten archive

As we described in our ECHT'90 paper, one of the motivations that led us to
developing Microcosm was the acquisition of the archive of the Earl
Mountbatten of Burma by the University of Southampton in 1987.

The archive contains about 250,000 text documents, 50,000 photographs, many
of his speeches recorded on 78 rpm records, and a large collection of film
and video. It spans his entire lifetime (1900-1978) and essentially mirrors
British history in the twentieth century.

The multimedia nature of the archive makes publication impossible by
traditional methods. We became very interested in the prospect of building
a digital version of the archive. This was not something that would fit
neatly into a database. There was no linear sequence to the material other
than that of chronological date, and users would want to move seamlessly
from one document to other related documents. Every user would approach the
material from a different perspective, so there was a need to create
different 'views' for different users. To us this meant different sets of
hypertext links, and the integration of hypermedia technology with database
management and information retrieval systems to even begin to tackle the
issues.

Here is an example of a dialogue with AL to illustrate the type of
environment we were trying to build.

WH: Al, lets see some examples from the period 1947/48 when Mountbatten was
Viceroy of India

AL: Could you be a little more specific?

WH: How about something typical of the period - a picture of Mountbatten in
dress uniform

AL: (slight pause) Will this do?

SLIDE 9 - Mountbatten in dress uniform

Thats fine. The database and information retrieval queries that AL will
have had to perform to retrieve that photograph from the archive could be
extremely complex. He will be searching the picture catalogue for pictures
taken between 1947 and 48 and either using key-words that describe the
picture, or some sort of image processing techniques, to find the ones of
Mountbatten in uniform. Anyway, lets look at some more information in the
archive.

Through working with the historians, we know that when Mountbatten arrived
in India, one of the many problems he had to deal with was that of the
Indian National Army. AL,  can you find us some text documents about the
Indian Army?

AL: This one looks like it might be of interest

SLIDE 10 - Viceroy's interviews No.22, April 1947

WH: Ah yes, this is interesting. Here Mountbatten is describing his utter
despair at ever being able to achieve independence for India and mentions
his interview with Claude Auchinleck about the Army. AL will have found
this by doing a text search of the catagolue or the documents.

AL Do we have any other documents that refer to the Indian National Army
Problem and Auchinleck?

AL: Yes, here are the minutes of a meeting at which the problem was discussed

SLIDE 11 - Minutes of 2nd Miscellaneuos Meeting

WH: There AL could have followed a pre-defined hypertext link or again
undertaken a text search of the documents.  Auchinleck is clearly a
significant person in this debate. Can you tell us a bit about him, Al?

AL: Field Marshal Sir Claude Auchinleck was the Commander in Chief of the
Indian Army during 1947 after a long distinguished career serving in the
British Army. I have some pictures of him if you would like to see one?

WH: Yes please, one with Auchinleck and Mountbatten would be good

SLIDE 12 - Auchinleck and Mountbatten

So again, AL has made links, that is associations, between relevant pieces
of information using a mix of database, information retrieval and hypertext
techniques.

What We Learnt

We started with the period when Mountbatten was in India, and we quickly
learnt that the biggest issue was that of copyright, but we'll  leave that
to the lawyers to resolve. Working with the archive and trying to build a
system that would support these types of queries led us to build an open
hypermedia system, Microcosm, in which there was a minimum of three levels
of access to information.

 Structured access to the information through the use of document and
database management systems to answer questions like
	"List photographs taken by ...."

Information retrieval techniques to do similarity matching across indexed
sets of documents to answer questions like
	"Find photographs similar to ........"

Browsing/hypermedia techniques that could be used whatever the format of
the information to answer questions like
	"Tell me about ......"

Voyager 2001

Microcosm is in essence a set of communicating processes (filters). As we
move beyond today's version of the system, our work is naturally evolving
to incorporate the use of intelligent communicating agents to provide
access to distributed information in ways that are completely transparent
to the user.

The vision is epitomised in science fiction by HAL, but other examples
well-known to televison viewers would be ORAC and ZEN from Blake Seven and
HOLLY from Red Dwarf.  The following dialogue is intended to represent the
way we might interact with computers in the future to find the information
we are looking for.

We start with Cell Biology since that was one of the first topics we used
at Southampton for developing a hypermedia application.

Suppose we want to  learn a little bit about how cells move. AL, tell us
something about Cell Motility

AL: Here's a tutorial on the subject used by biology students at Southampton

SLIDE 12a - Cell motility image

WH: OK, Al lets see the Introduction

SLIDE  13-  Intro to Cell Biology

WH: Thats good. Some links have been highlighted by the biology tutor to
tell the student that there is more information in the system about this
subject that they should look at. Lets look at Tissue Cell Movement AL

SLIDE 14 - Tissue Cell Movement

WH: Now the word embryo is not highlighted but I'd still like to know more
about what an embryo is. AL can you tell me about embryo's

AL: How about a dictionary definition?

WH: That will do fine

SLIDE 15 -  Embryo definition

WH: OK. Now lets switch contexts to computer science. Here's a challenge
for you AL. Can you find me some relationship between embryo's and computer
science?

AL: (Pause) Hard but not impossible. I have a paper by Alan Turing, one of
the pioneers of computer science, which mentions embryo's.

WH: Yes lets see that

SLIDE  16- Turing Paper

WH: Interesting.  Now lets stay with Computer Science and think about the
Department here at Southampton, since thats where we are. AL what
information do you have about the Department?

AL: Here's the information that's available on the Internet

SLIDE 17 - ECS info + map

WH: And almost the first thing you see is that the Department of
Electronics and Computer Science is housed in the "Mountbatten" building.
There's a familiar name. But what's the association between Mountbatten and
Computer Science AL?

AL: Well, Lord Mountbatten was once the President of the British Computer
Society.

WH: So that could be it. Now lets change context again to history and World
War II. What information do you have about Mountbatten's role in World War
Two AL?

AL: Would you like to start with some pictures. Here's a typical group
photo taken in Casablanca in 1943.

LEFT SLIDE  18 - Mountbatten and Churchill group

WH: OK, thanks AL, we won't continue down that trail now. But it is
interesting that we've ended by making a link to Churchill, whose archive
is also a major feature of research work at Southampton.  Lets finish with
the film of Churchill's  "Iron Curtain" speech where he talks about the
need for making links, or special relationships as he calls them. Can you
find that clip AL?

VIDEO 8 - Churchill's speech

So maybe that's how we'll be finding information in 2001. What's going on
behind the scenes?  Well, there's a complicated mix of database queries,
pattern matching, hypermedia linking and knowledge-based techniques that is
all transparent to the user, as is the location of the information that is
being made available. Al is creating associations between different pieces
information to answer my queries - its all making links!

There's also an awful lot of natural language processing and speech
recognition, and in order to do all this Al is a supercomputer because the
processing power required for him to perform these tasks is quite
phenomenal.

This is all rather far-fetched considering the state-of-the-art today but
whatever proposals we might come up with for an open hypermedia protocol,
they should take us another step forward to achieving this vision.