Almost 1,000 years after the original, the BBC has produced a new version of the famous Domesday Book - using the latest video disc and computer technology to compile an electronic portrait of the UK and the life of its people
in the 1980s.

To Peter Armstrong, leader of the Domesday Project, this venture points the way to new forms of interactivity which could
revolutionise our attitudes to and uses of television in the future.

Re-printed from : Television: Journal of the Royal Television Society January/February 1987

Transcript :

i ,
Peter Armstrong (BBC)
Almost 1,000 years after
the original, the BBC has
produced a new version of
the famous Domesday Book
— using the latest video
disc and computer
technology to compile an
electronic portrait of the
UK and the life of its people
in the 1980s.
To Peter Armstrong,
leader of the Domesday
Project, this venture points
the way to new forms of
interactivity which could
revolutionise our attitudes
to and uses of television in
the future.
Beyond Domesday
An idea for interactive broadcasting
There seems to have been little
disagreement that it was right for the BBC
to undertake the Domesday Project. Even
though it apparently had no direct connec
tion with broadcasting, it did seem to
follow the public service tradition of pro
viding education and information for the
general public. The lack of a direct broad
casting link meant only that it was not
financed by the licence fee.
In essence the project consisted of the
creation of a multi-media database on
British life in the 1980s to mark the 900th
anniversary of the Domesday Book. It
seemed to me obvious that if William the
Conqueror had been offered the same
opportunity he would have wanted to
include photographs and maps of the
Kingdom as well as text. He would have
used official statistics, like the latest
census results, and he would have wanted
examples of television news coverage.
The only medium which offered us the
possibility of storing all these different
types of material was interactive video disc.
However, the Laservision disc format
had to be modified in order to include all
the digital information — statistics, text
and computer programs — that we wanted
to include. This led us to create with
Philips, Eindhoven, the A.I.V. (Advanced
Interactive Video) format which can store
digital information in the portion of the
waveform normally reserved for sound.
The total capacity of an A.I.V. doublesided
disc is 108,000 image frames and
648 megabytes of error-checked data (or
sound where required).
From the user's point of view the essen
tial feature of the Domesday Discs is that
all the material can be viewed interactively
— you decide all the time what is shown
on the screen. So, for example, a user can
go into a library — in this century or the
next — and point anywhere on the map of
the United Kingdom. He or she then
zooms in through closer and closer maps
to pictures, text or data as fancy or a
particular research topic dictates.
Beyond Domesday and the other inter-
30 Television: Journal of the Royal Television Society January/February 1987
active video discs being developed around
the world lies the larger concept of
interactive television: the extension to
broadcasting of the principles of inter
activity between the viewer and the
programme. I am sure that well before
the beginning of the next century
television must become more interactive.
Notice that I say more interactive — it's
not an all or nothing state. From the
beginning there has been some inter
activity — as people like Mary
Whitehouse have always pointed out —
via the on/off switch. Then the viewer had
the power to change channels — more
powerful when it could be done remotely.
Now broadcast television needs to develop
the next stages in interactivity.
What I tend to think of as slump
television in which the viewer collapses in
front of the set and gives himself or herself
over into the hands of the programme
makers — this will seem increasingly
intolerable. My children are now so used
to interacting with the screen — in video
games or bank machines — that they will
demand something more relevant than
slump TV.
Moreover, as long as interactive pro
gramming is confined to video disc publi
cation, it will always be regarded by
broadcasters as peripheral — at best a
support medium for television pro
grammes in the same way as books or tee
shirts. In the BBC it will never qualify for
licence fee funding and all that goes
with it.
Broadcasting is also better than video
disc for the sustained sequences of linear
video — for example a Shakespeare play
— which overflow the capacity envisaged
for video discs in the next few years.
In the short term the only answer seems
to be cable. Some cable operators are
already talking about equipping cable
heads with interactive video disc systems
that subscribers could log onto from
home. This is a useful start, but I suspect
that speed of access is going to be very
slow indeed unless there are a lot of
players being commanded by not many
subscribers.
What is really needed in the interactive
broadcast. Whenever I talk about this, I
am told by engineers that it is in effect a
logical impossibility. There is no reverse
channel from the viewer back to the
broadcaster without something like cable;
and in any case what is being broadcast
cannot be varied to suit different viewers.
I have always felt that this argument was
too simplistic, so I sat down recently to
work out how it might be done.
Let me make clear that what follows is
no more than my personal scenario: I do
not speak for the BBC. Nor is this
approach immediately practical — I
suppose it must be at least five years away.
The point is that it represents a different
way of thinking about the problem — and
as we found with Domesday, seeing
far,
_i ' <r ..A .317
1 PortMow,
Portreath-'
.'/An b.'y ?odfl
0.'. .v/
"'ViU'
SJ-' .
63S 5» 1
VU
U n J
V GborchtowoE'"®
" J V - Ut
< ' Help' ' o - Text r 
."■'r' FALMOUTH
patterns for future systems is the most
important first step towards bringing them
into being.
In essence my idea is that we decouple
the act of transmission from the act of
viewing by introducing a form of inter
mediate recording. This introduces a timeshift
of variable length for different pur
poses, which is under the control of my
personal micro computer, taking instruc
tions from me and from data within the
broadcast.
Since I'm aware that nothing catches on
without an acronym, I've christened my
proposed system TRANSIT — Trans
mitted Interactive Television. To make
TRANSIT work requires only one step
beyond the present technology in the new
BBC video disc player: a separate record
ing hetid ctipabic of writing pictures and
data onto an ertisable video disc.
Assume then that by 1991 we have such
;i mitchine at domestic prices. Now we can
begin to broadcast intenictively. Only one
channel is required, dedicated either all
Any part of the UK map can
be selected for close
examination, with successive
displays — accompanied by
text and photographs in
increasing detail — of areas
scaling down to as little as 4km
television: .lournnl ol'tlie Royal 'television Society Janu;iry/l*ebruary 1'X7 31
or part of the time to interactive broad
casting. The programmes it transmits are
not designed to be watched as they go out
— they might look very odd seen in that
way. They are designed to be recorded
and watched after a delay of around 15
minutes. Here is my variable time-shift
machine:
REPLAY
Figure 1: Variable time-shift
optical disc player
I'm assuming that this disc holds 60
minutes of recorded pictures with sound
or data. The recording head records every
thing that is broadcast on the interactive
channel: as the disc is filled up, the head
moves back to the beginning and records
over the earlier material. Meanwhile the
replay head is free to play back any part
of the recorded material that the viewer
wishes.
Such a system suddenly gives us real
control over what we are watching. Take
first the case of a programme made in a
linear way as at present.
Imagine, for example, we are watching
an ordinary transmission of Whistle Test
and we want to see a particular song
again, without missing the interview that
follows. This is how it would work:
TIME 8.45 8.50 8.55 9.00 9.05 9.10 9.15
RECORD A B C D E F
REPLAY
Figure 2: Instant replay
B
-15 -20
Whistle Test is advertised on the inter
active channel as beginning at 9.0. Our
time shift disc recorder has been running
all evening, so it has already recorded the
first 15 minutes of the programme which
actually began being broadcast at 8.45.
Sections A, B and C are already on my
disc and the record head is beginning on
section D.
32 Iclcvision: Journal of ihc Royal Television Society January/l-ehruary 1987
The replay head, however, goes to the
front of the programme and plays me
section A. Playback is running 15 minutes
behind recording, but as a viewer I neither
know nor care — it feels to me as though
the programme is going out now. Section
A turns out to be a really good song that I
want to hear again. I select REPLAY via
my computer handset, and the replay head
moves back to play section A again. After
I've watched that, I'm happy to see what's
next in the programme, so the head moves
straight on to B.
The extra time I have in effect gained
(my viewing of Whistle Test is going to be
five minutes longer this week) has been
absorbed by the system in the fact that
replay is now 20 minutes behind record
ing. Again I am blissfully unaware of this.
I will probably restore the balance by
interacting in the opposite way.
Using the same principle in reverse, it
is easy to envisage how the viewer could
choose to skip ahead if he or she came to
a boring part of the programme (or an
offensive part, or — dare I say — a com
mercial break). It would work like this:
TIME 8.45 8.50 8.55 9.00 9.05 9.10
RECORD B C D E
REPLAY A C
-15 -10
Figure 3: Skip ahead
Our viewer turns on for a magazine
programme at 9 o'clock and starts watch
ing section A from the replay head, while
the record head is by now recording
section D. Then section B proves boring
so we decide to skip on to C. Everything
functions as before and we have simply
iost' the boring five minutes we would
otherwise have to have lived through. The
system has absorbed this by shortening
our time shift delay from 15 to 10 minutes.
So we are free of the tyranny of real
time viewing. But what about fuller inter
activity? Let's take a practical example —
an interactive broadcast of the series
Civilisation, which allows the viewer to
leave the main narration to explore pic
tures and places in more detail. So if
Kenneth Clark is standing in front of a
picture, the viewer should be able to call
up close up details of the picture, exam
ples of the painter's other work and more
detailed interpretative text or commen
tary. If Kenneth Clark is walking round a
building, the viewer should be free to
wander off on his or her own to explore
the rest of the building.
How is this possible using TRANSIT?
Let's take the walk as an example.
Somehow we must transmit the 2()(M) or so
still frames covering the rest of the build
ing, without the viewer seeing them
sequentially. We also have to transmit the
associated data structure that tells the
computer how to move spatially between
these frames in a surrogate walk. Finally
we must send the background text and
compressed voice-over that the viewer
may want to call up.
Here are two possible methods. I call
the first 'compulsory skip'.
TIME 8.45 8.50 8.55 9.00 9.05 9.10 9.15
RECORD A iii C D E F G
REPUV/ A C WM D
s
-15 -10 -13
Figure 5: Reserved disc space
Figure 4: Compulsory skip
The programme advertised for 9 o'clock
begins transmitting as usual 15 minutes
earlier and is picked up by the recording
system. The first five minutes, which I
have called A, offer no interaction. Then
a skip marker is transmitted (s on the
diagram) indicating how many frames
follow that cannot be viewed sequentially.
Following the skip marker 7,500 still
frames with associated data are transmit
ted — section B. At 8.55 the main line of
the programme continues with sections C
and D. These contain the sequences of
Kenneth Clark describing the pictures
and buildings covered by the still frames
in B.
At 9.0 our viewer switches on, the
replay head goes to the start of the pro
gramme at A, and as far as the person
watching is concerned the broadcast is
just beginning. At 9.05 when the recording
head is beginning section E, the playback
head reads the skip market at B and jumps
forward the number of frames indicated
— to the beginning of C. The viewer is
not aware that anything has happened,
although he is now 10 minutes behind
broadcast time rather than 15 minutes.
At the end of C (it could have'been
anywhere in the middle equally well) our
viewer decides to take the option to walk
round the building on his or her own —
looking at other rooms, and reading up
some background history that is of
particular interest.
To achieve this the computer uses the
data structure it has been sent to find the
right frames and text within section B.
Let's suppose this takes 3 minutes and our
viewer now wants to return to the main
narration. The replay head moves to
section D and everything continues
without a break. The slack time simply
means that the record/replay interval has
now increased again — to 13 minutes.
Nothing has interrupted the recording
head which has recorded every frame of
the broadcast and is at the moment
finishing off section F.
STARTADDRESS
SENT WITH DATA
0000
0000000
OOOOTTUOOO
00000
SINGLE
FRAMES
MOVING
VIDEO
A different approach would be to
transmit the split point between the two
types of material at the beginning of the
transmission. A reserved area of the disc
(say at the centre) would then be kept for
interactive material and the replay head
would move there whenever the viewer
wanted to leave the main narrative line.
Here the simplest approach would be to
transmit all the interactive material in the
first five or 10 minutes (which would give
between 7,500 and 15,000 still frames with
up to 90m megabytes of data). The record
ing head could then move smoothly five
to 10 minutes ahead of viewing. Having
filled the disc once it could happily con
tinue by writing over the opening. If the
user had spent so long with the walks or
close up pictures that material he or she
had not yet seen was about to be overwrit
ten a warning would be given.
The example I have used, based on
Civilisation, only takes interactivity so far.
The Domesday Discs would not be
suitable for broadcasting as they stand,
because there are not enough linear
sequences to mask the still-frame
sequences. The strength of TRANSIT is
combining interactivity with more moving
film than a single-sided video disc could
hold. Then again if you'd spent two and a
half million pounds making Domesday,
would you want to transmit it?
If you did want to transmit more
complex interactive programmes, why not
just transmit the programme for recording
at night and viewing interactively later?
Clearly this would work — and with a
simpler video disc player — but it would
seem to me a much less attractive option
Television: Journal ol ihc koval lelevisitui Soeielv January/I*ehruary IVS7 33
for the viewer. TRANSIT allows you to
dip into programmes, watch when other
people are watching and catch news when
it is fresh. The alternative means one
decision to record and another decision to
find time to watch later. That's why I find
I watch such a small proportion of the
programmes I record at present on
cassette.
It's easy to think of the kind of broad
cast programmes that would gain from
interactivity. Language-learning could
proceed at the viewer's own pace. Sport
could put slow motion replay under the
viewer's, not the director's, control.
Magazine programmes could offer more
items, knowing that viewers would pick
and choose which to watch. The Budget
Special could include all the text and back
ground figures with software to let you
work out the implications for your own
finances. Travel films could allow you to
choose which way to go when the river
divides.
My own idea of the ultimate television/
computer programme is the interactive
documentary. This is more difficult than
anything I've described because it seems
to need a wide choice of alternative
moving film sound sequences. But as the
technology develops this too may turn out
to be only a difference of scale, as
TRANSIT becomes a much more power
ful — presumably solid-state device at the
heart of every home's entertainment and
information system.
Of course some channels — for news or
event broadcasting — will always need to
be live, but I believe there could well be a
place for an interactive channel if the
advantages I've described seem suffi
ciently attractive. And the best way for
the broadcasting organisations to prepare
for that day is to develop interactive
programme-making techniques through
the medium we have at present — the
interactive video disc.
Of course, all this is a dream at the
moment — but two years ago so was
Domesday n
TRAIN
FOR THE FUTURE
Television is the fastest growing industry in thr world.
The Royal Tele/ision Society is a unique, independent
society concerned with every aspect of television's
development including recognising the need for
professional training.
So if you want to train for the
future in:
• Television Engineering*
• Digital Techniques
• Shooting and Editing with
Video
• Production Management
• Screenwriting
• 3-Machine Editing
• Television Sound
• Lighting
• Re-Training
why not call Liz Rankin on
01-387 1970 and ask for details?
The course lectures are available as
a book, in two parts, entitled
Television Engineering: Broadcast
Cable and Satellite'.
R('l
Television
Society
Tavistock House East
Tavistock Square
London WC1H9HR
34 Television: Journal of the Royal Television Society January/February 1987