1
00:00:00,000 --> 00:00:05,120
30 years since the birth of the World
Wide Web ?

2
00:00:05,120 --> 00:00:12,480
I believe it. So that would, if it ....
reassure me that we are now in 2021 ?!

3
00:00:12,480 --> 00:00:15,679
So therefore we must be talking about 1991?

4
00:00:15,679 --> 00:00:22,240
That was a weird era ! I mean, in my
pronouncement so far, I think we've got

5
00:00:22,240 --> 00:00:27,279
to the stage of the mid-80s where you had

6
00:00:27,279 --> 00:00:32,559
Macintoshes becoming available. You had
other workstations beginning to

7
00:00:32,559 --> 00:00:36,880
become available, from
people like SUN and Apollo Domain

8
00:00:36,880 --> 00:00:42,840
I think it was something like that.
So, things were moving on but yet,

9
00:00:42,840 --> 00:00:48,480
weirdly, the World Wide Web which 
came after the internet, remember,

10
00:00:48,480 --> 00:00:52,480
is basically how should these Internet

11
00:00:52,480 --> 00:00:58,399
capabilities be harnessed and
packaged to be Sean-and-cameraman

12
00:00:58,399 --> 00:01:06,879
friendly. so that ordinary mortals
can do it, you know. And it was not obvious

13
00:01:06,879 --> 00:01:11,040
In those early days. I seem to recall
there were information

14
00:01:11,040 --> 00:01:14,479
services weren't there? Like Archie and WAIS

15
00:01:14,479 --> 00:01:19,680
Wide Area Information Services. People were
really trying to find out ... well if we do

16
00:01:19,680 --> 00:01:23,119
have interconnected networks on computers

17
00:01:23,119 --> 00:01:26,960
how do you know who's got what and where
everything is?

18
00:01:26,960 --> 00:01:30,640
And i think the idea about WAIS was making

19
00:01:30,640 --> 00:01:34,799
basically catalogues of what's where and
all this kind of thing.

20
00:01:34,799 --> 00:01:39,119
But i think, to be fair to him,
[Tim] Berners-Lee's great achievement was

21
00:01:39,119 --> 00:01:42,320
being able to think out of the box. 
Perhaps it was the

22
00:01:42,320 --> 00:01:45,439
fact that he
was trained as a scientist and not as a

23
00:01:45,439 --> 00:01:49,280
computer scientist and i don't know ...,
that he could think outside the box.

24
00:01:49,280 --> 00:01:54,479
But basically the whole
thing about, y' know, the http

25
00:01:54,479 --> 00:01:58,719
and the internet protocol (IP)  and all this
kind of stuff. Really he managed to work

26
00:01:58,719 --> 00:02:02,799
out pretty
something pretty robust about how this

27
00:02:02,799 --> 00:02:06,880
would work. In the very early days, I
think we touched before, y' know, on the

28
00:02:06,880 --> 00:02:11,520
fact that the gap between something becoming

29
00:02:11,520 --> 00:02:16,400
vaguely possible and something becoming
commonplace could be anything up to

30
00:02:16,400 --> 00:02:21,040
10 years? Did we once agree? And here's a
classic example of it.

31
00:02:21,040 --> 00:02:25,120
I mean, in 1991, yes i was very slightly aware of
it but i didn't see it

32
00:02:25,120 --> 00:02:30,160
i.e. the Internet such as it was and
any development. So i didn't see it

33
00:02:30,160 --> 00:02:34,400
as being massively essential [for me].
>> Sean: Well, of course, we didn't actually have the

34
00:02:34,400 --> 00:02:39,519
Internet so the news wasn't travelling
fast at all! You had to read about it somewhere 

35
00:02:39,519 --> 00:02:44,400
in print, or [in] some kind of 
mailshot. 
>> DFB:  Yes, that's right.

36
00:02:44,400 --> 00:02:49,440
In fact my very earliest usages
of the Web, when it became available,

37
00:02:49,440 --> 00:02:53,920
were for really
boring things like working out that

38
00:02:53,920 --> 00:02:58,720
so-and-so had got a set of files that
implemented something on another Unix

39
00:02:58,720 --> 00:03:02,239
machine so we didn't necessarily have to

40
00:03:02,239 --> 00:03:08,000
make a magnetic tape or
burn a CD, or whateve. We could actually

41
00:03:08,000 --> 00:03:12,239
go to that site, not just to pick up a
bit of email because that was

42
00:03:12,239 --> 00:03:16,319
fairly straightforward, but also to
actually, with permission,

43
00:03:16,319 --> 00:03:19,599
download a whole suite of files. 
What about

44
00:03:19,599 --> 00:03:25,040
photos then? My memory about early
attempts to download

45
00:03:25,040 --> 00:03:30,000
photos down an early Web [link] thing.
I mean, honestly ...

46
00:03:30,000 --> 00:03:34,560
>> Sean: My memory is of being at University
and me and a friend of mine going on BBS's

47
00:03:34,560 --> 00:03:38,080
[Bulletin Board Systems]
and this was after the World Wide Web

48
00:03:38,080 --> 00:03:42,480
existed,   but
we had some sort of fractured connection

49
00:03:42,480 --> 00:03:46,400
to the Web through a BBS.
And yeah, occasionally, you

50
00:03:46,400 --> 00:03:49,200
might be sent some photos and it took an age

51
00:03:49,200 --> 00:03:52,319
to download them. And by the time they
downloaded were they what you wanted in

52
00:03:52,319 --> 00:03:55,599
the first place. Possibly not!
Do you go off and spend another

53
00:03:55,599 --> 00:03:58,080
several minutes waiting for the next one
to download.

54
00:03:58,080 --> 00:04:01,360
Yes i think this is where we would
benefit, if

55
00:04:01,360 --> 00:04:06,080
people wanted us to say more on this, from
having got Julian Onions as part of the

56
00:04:06,080 --> 00:04:08,799
Computerphile team, as it were, because he

57
00:04:08,799 --> 00:04:13,680
was he was with me i think,
as I remember, in that first phase of his

58
00:04:13,680 --> 00:04:16,639
career at Nottingham.
I think he was with me till the early

59
00:04:16,639 --> 00:04:20,799
90s and he would be pretty sound on which

60
00:04:20,799 --> 00:04:25,840
machine was it that first had
a Web browser on it, capable of even

61
00:04:25,840 --> 00:04:28,720
visiting another site let alone 
doing anything serious.

62
00:04:28,720 --> 00:04:34,720
The ability to
actually have what - should we say

63
00:04:34,720 --> 00:04:40,479
interactive video material - made
available at ultra slow speeds over the

64
00:04:40,479 --> 00:04:44,960
Web may have been in the early to mid-90s

65
00:04:44,960 --> 00:04:48,320
among the ultra keen people who could
just about

66
00:04:48,320 --> 00:04:51,680
bear getting it to work. But how long
does it take to become

67
00:04:51,680 --> 00:04:55,280
commonplace? The answer, as ever, is 10 years

68
00:04:55,280 --> 00:05:03,039
My major concern, at the time,
late 80s early 90s, exactly this era,

69
00:05:03,039 --> 00:05:08,160
for me, was a transition from PostScript
to PDF.

70
00:05:08,160 --> 00:05:14,080
Now, we've done videos on
certain aspects of this but I think

71
00:05:14,080 --> 00:05:16,800
perhaps what i haven't explained about my

72
00:05:16,800 --> 00:05:22,960
great adventures in this regard,
was that in the middle 1980s and

73
00:05:22,960 --> 00:05:30,400
with the development and announcement of
PostScript basically as a

74
00:05:30,400 --> 00:05:34,360
graphics language available on
Macintoshes and,

75
00:05:34,360 --> 00:05:42,400
notoriously, available by having
a [Apple] laser printer running that language,

76
00:05:42,400 --> 00:05:46,240
requiring enormous amounts of memory, for
the time, [and] costing a fortune.

77
00:05:46,240 --> 00:05:53,120
fortune. It couldn't be got any cheaper
than $7000 but it was touch and go.

78
00:05:53,120 --> 00:05:56,960
I mean, the hardware was only just fast

79
00:05:56,960 --> 00:06:02,080
enough to render that [300dpi] page at 
that time.

80
00:06:02,080 --> 00:06:05,600
What struck me at the time, and I was very
much in the era

81
00:06:05,600 --> 00:06:08,800
of knowing that my future, in some sense,
would lie

82
00:06:08,800 --> 00:06:12,240
in documents d- Digital Documents,
Electronic Publishing.

83
00:06:12,240 --> 00:06:16,160
I thought, y' know, there is something
about this [PostScript] language because, of course,

84
00:06:16,160 --> 00:06:19,360
PostScript is
Turing Complete. It's a weird language.

85
00:06:19,360 --> 00:06:23,199
Remember, it works on a stack. You overtly program

86
00:06:23,199 --> 00:06:28,400
in Reverse Polish [postfix].
You don't say 'a + b' or even '+ a b'

87
00:06:28,400 --> 00:06:33,600
You say 'a b +' and in a much more
verbose way, that is the way

88
00:06:33,600 --> 00:06:37,039
PostScript works. On top of it being an excellent

89
00:06:37,039 --> 00:06:42,479
2D graphics language, moving beyond what
most printers would have regarded as the

90
00:06:42,479 --> 00:06:47,120
"toy" of the [Apple]
laser printer people began to ask: "Would

91
00:06:47,120 --> 00:06:49,759
it be possible to put this PostScript language

92
00:06:49,759 --> 00:06:56,960
on real typesetters?" Yes, it would. 
And a bit bemused i think by the runaway

93
00:06:56,960 --> 00:07:03,759
success of the Apple Laserwriter,
yes it was Linotype

94
00:07:03,759 --> 00:07:10,000
that took the initiative of saying
"yes" to an idea internally about

95
00:07:10,000 --> 00:07:14,000
"How do we try and migrate PostScript
onto a real typesetter?"

96
00:07:14,000 --> 00:07:18,000
Because i've talked about my adventures
with the [Linotronic] 202 typesetter. But it was not

97
00:07:18,000 --> 00:07:21,120
PostScript based, but they tried one out just

98
00:07:21,120 --> 00:07:24,639
as an idea and people said: "Oh that's
wonderful - we want it!

99
00:07:24,639 --> 00:07:28,319
Doesn't matter about the cost *too* much but
you'll have to get it down quite a lot

100
00:07:28,319 --> 00:07:30,080
from [what you're asking] if you want to sell 
lots and lots of them".

101
00:07:30,080 --> 00:07:34,560
So, in the end then, that convinced me

102
00:07:34,560 --> 00:07:39,680
that just in my narrow world of journal
publication I wanted to create

103
00:07:39,680 --> 00:07:46,160
an electronic document learned journal which 
I did. It didn't run

104
00:07:46,160 --> 00:07:50,560
for very long because the
Electronic Documents 'learned conference'

105
00:07:50,560 --> 00:07:53,919
[now ACM DocEng] was far more popular than the 
journal so we started off with

106
00:07:53,919 --> 00:07:57,840
Electronic Publishing.
But then, gradually, it morphed into being

107
00:07:57,840 --> 00:08:02,400
a vehicle for publishing the Proceedings
of the Document Engineering

108
00:08:02,400 --> 00:08:06,560
Symposium, as it were, and it's gone [om] to
[be] great guns from there.

109
00:08:06,560 --> 00:08:11,520
But anyway, I was messing about then,
at an early stage, with doing

110
00:08:11,520 --> 00:08:18,160
proceedings for conferences all
in PostScript and thinking "Well this is

111
00:08:18,160 --> 00:08:21,120
good because you're actually archiving a

112
00:08:21,120 --> 00:08:25,759
[machine independent] program
and this could run, you know, ...

113
00:08:25,759 --> 00:08:29,039
although the hardware will be very
different and maybe they're emulating a

114
00:08:29,039 --> 00:08:32,719
layer on top of it, there will always be
a way you can write an interpreter for PostScript."

115
00:08:32,719 --> 00:08:36,959
So, here's my electronic archive that

116
00:08:36,959 --> 00:08:41,680
is potentially machine independent. Yes, of course 
printers [i.e. printing companies] archived

117
00:08:41,680 --> 00:08:45,120
but they always archived on the basis
that the world would be

118
00:08:45,120 --> 00:08:51,040
ruled totally by Monotype and Linotype

119
00:08:51,040 --> 00:08:53,440
They are in charge and what they say goes.

120
00:08:53,440 --> 00:08:57,040
And [at times] I was treated like a pariah for
saying "I'm sorry,

121
00:08:57,040 --> 00:09:00,720
it's not going to be like that. It's all
about [fonts and] computer graphics.

122
00:09:00,720 --> 00:09:03,760
And it's computing companies that are
going to win out of this -

123
00:09:03,760 --> 00:09:07,440
not just purely graphic typeset companies

124
00:09:07,440 --> 00:09:11,279
And i was right but [almost] nobody would 
listen to me at the time.

125
00:09:11,279 --> 00:09:16,720
So yes, this is
perhaps another anecdote of "How i met

126
00:09:16,720 --> 00:09:18,959
John Warnock" Because people ask me. And
[thry] say well

127
00:09:18,959 --> 00:09:24,800
"How did Adobe get to know about your
conviction that PostScript would be a

128
00:09:24,800 --> 00:09:28,560
good archiving mechanism."
And the answer was that I was working

129
00:09:28,560 --> 00:09:32,240
with John Wiley UK, the journal publishers 
at the

130
00:09:32,240 --> 00:09:36,320
time and suddenly my
contact at Wiley's [UK] down on the south

131
00:09:36,320 --> 00:09:39,839
coast [of the UK] got a "Dear Mark Bide letter 
[that was the name of the Publishing Director"

132
00:09:39,839 --> 00:09:43,920
"We understand
from our contacts that you are

133
00:09:43,920 --> 00:09:48,160
considering doing a future journal related to

134
00:09:48,160 --> 00:09:51,120
Electronic Publishing
and that it might be under consideration

135
00:09:51,120 --> 00:09:54,959
as to whether to use PostScript as the
typesetting language and as the archival means?

136
00:09:54,959 --> 00:09:58,959
Would you care to join me

137
00:09:58,959 --> 00:10:05,600
and my colleague Dr Chuck Geshke
for tea in London ? "

138
00:10:05,600 --> 00:10:11,200
[at the Park Lane Hotel i think it was]
signed: Yours sincerely John Warnock".

139
00:10:11,200 --> 00:10:14,880
The two co-founders of Adobe [no less]
So Mark rang me up in great excitement

140
00:10:14,880 --> 00:10:19,360
[and] said: "Well you know
this madcap idea of yours, of using the

141
00:10:19,360 --> 00:10:22,560
most expensive typesetting software solution

142
00:10:22,560 --> 00:10:26,079
seems to at least have got the big boys
interested!" 
And, yes, it was absolutely

143
00:10:26,079 --> 00:10:31,680
hilarious i mean
John Warnock s a good friend as a result

144
00:10:31,680 --> 00:10:33,920
of this largely because it was clear he and

145
00:10:33,920 --> 00:10:37,680
Chuck were such good friends
and so much 'singing from the same hymn sheet'

146
00:10:37,680 --> 00:10:40,880
There was nothing put on about it. They

147
00:10:40,880 --> 00:10:44,560
just said: "We're delighted you're doing
this and we'll give you every help".

148
00:10:44,560 --> 00:10:47,839
Obviously, y' know, i think at the end of
the talk they did

149
00:10:47,839 --> 00:10:51,519
indicate that there were works in progress

150
00:10:51,519 --> 00:10:56,560
that might come up with a sort of
"flattened" version of PostScript

151
00:10:56,560 --> 00:11:03,920
that was much more suited to
doing electronic documents.

152
00:11:03,920 --> 00:11:08,959
I think the feeling was: 
"Well PostScript's wonderful

153
00:11:08,959 --> 00:11:12,320
but it is Turing Complete. In a sense
it's too powerful.

154
00:11:12,320 --> 00:11:18,320
There is nothing in PostSscript to [prevent you from s
saying] "Here is my code for rendering this page

155
00:11:18,320 --> 00:11:23,200
It will be beautiful but - tell you what - just 
to annoy you all,

156
00:11:23,200 --> 00:11:28,079
before i do this [i.e. render the page] 
i'm going to use Postscript to calculate Ackermann(4,2)

157
00:11:28,079 --> 00:11:32,000
which as we all know takes [more than!] 27
quadrillion years.

158
00:11:32,000 --> 00:11:35,600
Or even Ackermann 3,1] which might take 15 minutes
[on a 1990s machine].

159
00:11:35,600 --> 00:11:40,640
Of course you wouldn't do that - but
somebody malevolent might do that to you.

160
00:11:40,640 --> 00:11:44,640
So in a sense Postscript being Turing
Complete was perhaps a bit too

161
00:11:44,640 --> 00:11:48,120
powerful for something that only needed a

162
00:11:48,120 --> 00:11:52,320
declarative statement of "I want this to happen

163
00:11:52,320 --> 00:11:56,720
exactly this. I  will leave the
implementation of it to you."

164
00:11:56,720 --> 00:12:03,040
Now,  in a way, it rang bells when the Web
became available because here

165
00:12:03,040 --> 00:12:07,680
was yet another declarative solution but
coming from a slightly different

166
00:12:07,680 --> 00:12:11,600
direction . So this Adobe thing 
[initially codenamed Carousel]  
was eventually

167
00:12:11,600 --> 00:12:16,160
revealed to me as PDF and the other

168
00:12:16,160 --> 00:12:22,800
declarative solution which was not as -
how shall we say - dot for dot accurate

169
00:12:22,800 --> 00:12:26,480
at the time as PostScript,
I saw them very much as being

170
00:12:26,480 --> 00:12:31,200
complementary. You know, if you want
ultra-super Vogue fashion magazines

171
00:12:31,200 --> 00:12:34,880
electronically over the Web
yes it's going to be PDF. If you just

172
00:12:34,880 --> 00:12:39,839
want some pretty decent graphics
although it's developed a lot,

173
00:12:39,839 --> 00:12:43,360
The World Wide Web, but using a declarative
model

174
00:12:43,360 --> 00:12:48,639
of the the HTML language.
I think as we've discussed the

175
00:12:48,639 --> 00:12:51,760
difference between
an imperative language that can do (say)

176
00:12:51,760 --> 00:12:55,040
Ackerman's function and a declarative
one that can't

177
00:12:55,040 --> 00:12:58,720
is that this Turing Complete [imperative way]
that can do anything

178
00:12:58,720 --> 00:13:02,880
needs to be able to ask for more memory
and to get it.

179
00:13:02,880 --> 00:13:07,839
There is no way in PDF for overtly
asking and saying

180
00:13:07,839 --> 00:13:11,760
"Give me lots of memory because i want to
mess about and make you unbelievably slow".

181
00:13:11,760 --> 00:13:16,160
And that's a great advantage

182
00:13:16,160 --> 00:13:22,000
to have a restricted language
that will try and stop people playing games!

183
00:13:22,000 --> 00:13:25,360
If you could have hidden things in PDF

184
00:13:25,360 --> 00:13:27,920
would it have ...  yeah, would it have taken off.

185
00:13:27,920 --> 00:13:32,240
I don't think it would.  And i think
Ii remember hearing that there'd been a

186
00:13:32,240 --> 00:13:36,320
robust discussion at Adobe
to the effect that one or two people

187
00:13:36,320 --> 00:13:41,199
were starting to use
PostScript preludes to do all

188
00:13:41,199 --> 00:13:44,800
sorts of weird and hairy things 
and wasn't this a 

189
00:13:44,800 --> 00:13:49,120
fantastic selling point for [PostScript] as a

190
00:13:49,120 --> 00:13:52,399
wrapper format for anything and everything?

191
00:13:52,399 --> 00:13:57,760
And the answer came back
from somebody on the team there saying

192
00:13:57,760 --> 00:14:01,839
"Basically this is not a feature it is a bug!" 

193
00:14:01,839 --> 00:14:06,079
And we have got to sit on it
firmly. People are now of course utterly

194
00:14:06,079 --> 00:14:09,360
aware of Trojans and so on and so forth, in a way

195
00:14:09,360 --> 00:14:14,399
that they weren't in the kind of early-late 80s early 90s,

196
00:14:14,399 --> 00:14:18,959
when this was just seen as
"Oh wow! isn't this incredible" 

197
00:14:18,959 --> 00:14:24,160
So, that was the situation then. You'd got
like a bit of a chasm between the two - 

198
00:14:24,160 --> 00:14:28,880
between HTML and PDF.
One of the things that fortunately

199
00:14:28,880 --> 00:14:34,480
happened, around about the early 90s,
was that Tim Berners-Lee and others

200
00:14:34,480 --> 00:14:40,000
headed off at the pass any attempt to

201
00:14:40,000 --> 00:14:45,920
make, as it were, the Web [and HTML] be the property
of either Apple or Microsoft or whoever.

202
00:14:45,920 --> 00:14:50,160
He [Berners Lee] set up the World Wide Web consortium
whose you know ... 

203
00:14:50,160 --> 00:14:53,920
one of his founding tenets was to try
and be platform neutral fair to everybody

204
00:14:53,920 --> 00:14:57,360
and to have clearly published standards.

205
00:14:57,360 --> 00:14:59,760
The Wworld Wide Web consortium that everybody could

206
00:14:59,760 --> 00:15:03,839
agree to and nobody would try and pull a
fast one on anybody else.

207
00:15:03,839 --> 00:15:07,120
Because i think even Microsoft had to admit [that]

208
00:15:07,120 --> 00:15:10,880
they'd try and rule the world with it
but they wouldn't succeed. But latterly

209
00:15:10,880 --> 00:15:15,040
of course, what has changed everything and

210
00:15:15,040 --> 00:15:20,160
again this business about just faintly
possible to send a photograph or a slow

211
00:15:20,160 --> 00:15:26,560
moving movie in the early to mid 90s
[versus] everything totally routine and above

212
00:15:26,560 --> 00:15:32,320
board in YouTube [of the] 2004-5 [era].
Something like that. What was needed of

213
00:15:32,320 --> 00:15:35,839
course, in those 10 years, was not just
any further realization about the

214
00:15:35,839 --> 00:15:40,399
[markup] language it was just
the infrastructure needed to be faster.

215
00:15:40,399 --> 00:15:43,759
So there of course you're up against
Telcos and

216
00:15:43,759 --> 00:15:46,800
Telecom Companies have to be
sufficiently enough of computer

217
00:15:46,800 --> 00:15:51,759
companies, or to have big
clients saying: "We need more bandwidth;

218
00:15:51,759 --> 00:15:58,399
we do not want 38 kilobits per second
out of AppleTalk, thank you very much

219
00:15:58,399 --> 00:16:03,279
we would like it to be at least
38 megabits, right, afactor of 1000 faster.

220
00:16:03,279 --> 00:16:06,639
Now get on with it!" But, of course, once

221
00:16:06,639 --> 00:16:10,000
it's been established that serious
people with serious money

222
00:16:10,000 --> 00:16:13,759
will pay for that then development will
follow.

223
00:16:13,759 --> 00:16:19,839
But there is always a bit of a
a lag - a bit of an "induction period" - there

224
00:16:19,839 --> 00:16:23,839
for people to say:
Yeah! you know, that idiot Brailsford is right [for once]

225
00:16:23,839 --> 00:16:26,240
This is going to take off; we'd better make plans.