Name: Bell Labs' Research (Prof Brian Kernighan) - Computerphile
Uploaded: 2020-04-13T16:22:13.000Z
Duration: 7 min 40 s
Description: Thousands of YouTube videos with English-Chinese subtitles! Now you can learn to understand native speakers, expand your vocabulary, and improve your pronunciation...

The comparative

>> Sean: Just to remind people here who, perhaps, haven't seen our earler videos with you, Brian.

Past simple

So, you worked at Bell Labs and at AT&T. I mean you did research there. You

Modal verbs - A2

were looking to the future! How right were you? How wong were you? Can you tell us? >> BWK:Yeah, so I was

at AT&T.  I went there for a couple of summers, starting in the in the late

1960s and I stayed there, essentially full-time, until 2000. I was in Bell Labs.

So at the time AT&T was a very big company. It was well over a million

people. It was the biggest company in the country and it provided communication

services - telephone - for essentially all of the United States. And Bell Labs was

the research arm of that. So that's the part that in theory is looking at the

future, trying to improve the services that they have and build the

technology, in all kinds of ways, that will make it easy for -- or make it

possible to improve -- telecommunications services. And so AT&T did things like

the transistor and the laser and miscellaneous other things that were

useful. And also a lot of computing-related  things as they gradually

realized that computers were here to stay. And that a lot of mechanical things

like relays, and so on, could be replaced by electronic devices and then

controlled by programs, running on general-purpose computers. In some sense

the Golden Age for me was probably the 1970s. I had just gotten out of School

and with the early days of computing there were a lot of really interesting

work going on and that was when UNIX was developed, in the  early 1970s, and the C

programming language and a variety of other things. And it's also when I played

with programming languages as well. So that was definitely a good time. Sort of

behind your question, I think, is: "Well, how good were you at seeing the future?"

And the answer is: "Pretty awful!".  I think most people are pretty awful at seeing

the future and I'm quite - I could hardly deny it so I'm perfectly willing to

admit it - I don't think we had and I wouldn't say all my colleagues and

friends at Bell Labs. had much of a clue of how the world would change. We I guess

knew about Moore's Law. Moore's Law, I think came from I think he published

that paper in 1965 based on a fairly a short period of data, like maybe five

years. And I don't think - certainly not I - but I don't think many people realized

the implications of an exponential rise in capability at a fixed price, if

applied for 40 or 50 years. I don't think we realized any of that  >> Sean: So, nobody believed it?

Nobody took it that seriously for this period of time? >> BWK: I don't know whether I would

say: "Nobody believed it", but I don't think people realize the implications.

Certainly I did not realize the implications and we were, I think, often

surprised when we discovered that something we had done had actually been

noticed in the outside world. I remember at one point Dennis Ritchie saying to me

something like: "We have arrived!",  because he had found in the New York Times, which

was still running classified ads for programmers, they'd wanted a

programmer who knew UNIX. And this was, I would guess, in maybe late 70s or

something like that. And this was us [saying]: "Boy, what we've done has actually had some

influence in the outside world", in a way that was completely unpredictable.

>> Sean:  So is there anything you can think of that you think should have made it but didn't?

I'm kind of putting you on the spot here! >> BWK: Yeah! that's definitely on the spot.

It's hard to say. One gets the feeling that a lot of things that were obvious

to us have been kind of lost in going forward. One [is] UNIX, the operating system;

C the programming language. Many of the tools. Those all had a flavor of

minimality, of being small and compact and good for their particular purpose,.

and not with too many bells and whistles and very carefully written, and so on. And

as Moore's Law came along and we got more and more processing power and more

and more memory capacity people sort of forgot the merits of simplicity, perhaps.

And so things became rather more baroque or rococo or - pick your architectural

period - and I think that that is in some ways to everyone's detriment that

systems are very complicated. The other thing is that at one point there was

really only one - let's call it UNIX - system and so things were fairly

compatible. And another thing that I think we've lost is

compatibility -  that it's harder to get things to work together perhaps than it

was. But in some ways perhaps that's necessary. You think the big change in a

lot of things has been the rise of networking: the idea that computers are

not self-contained things any more but, rather, they are devices that talk to

other devices using a variety of networking technology - Internet itself.

And of course increasingly phones. The phone system. And in fact there's no real

difference between the phone system and the Internet in some sense, it's kind of

an accidental separation that will disappear over time. And as different

things talk to each other what happens there gets more and more

complicated. There are more ways that things can break. You need standards but the

standards aren't necessarily there. And so, yeah, I think that's the place where we

need the most improvement in some sense. It's ways to make things work better

together and make them simpler and I guess those are related.  >> Sean: I think there's

a kind of irony, as well, that you've got this massive phone company

effectively pushing forward on computer technology, and now the many computer

people use is  the phone in their pocket!  >> BWK: Yes, right, I think that's again something

that people didn't predict. There's a famous story that a consulting company

in the United States, McKinsey, did a study for AT&T and told AT&T that there

was basically no market for cell phones; that people didn't want portable phones.

And this I think is a triumph of how consultancy can go wrong and yet not

suffer for it. in any sense. But AT&T didn't back that stuff as well - a lot of

the early work on mobile telephony was done at AT&T.

Absolutely. They figured out this the notion of cells and how you would pass

conversations from one cell to another something that's completely invisible to

people today but all of that was in fact invented originally at AT&T. And then of

course evolved tremendously by other people over the years. And again cell

phones today profit from Moore's law. That you've got incredible power in a

very, very compact device, would not have been possible

thirty years earlier. >> Sean: And the last thing they're used for       is phone calls >> BWK: It's hard to say.

You're hinting that people don't use their phones to talk and I think that's

true. As far as I can tell students that I deal with in the United

States use their phones primarily for texting each other and perhaps for

checking their Facebook pages - or whatever the modern equivalent is.

And this is so foreign to me that I don't actually know how to use a phone in many

ways. I had an extended discussion with Dave [Brailsford], earlier this morning, about how you

deal with with phone numbers and whether you need a 44 and not a zero or is it ... ?

Hopeless! And that was the reason I had trouble

communicating with him this morning. So culture shock in some sort of way.