>> Sean: [reading out a viewer question] Did the dominant linguistic theories of the time consciously or unconsciously

influence the design of certain programming languages? >> BWK: I suspect the

answer is 'yes' in some ways. I'm not sure that I'm enough of a historian to know

for sure, but for example languages like C derived from languages like, in

particular BCPL, which of course were influenced - sometimes positively,

sometimes in a sort of reactionary way to Algol, which was really really

fundamental. But also Fortran, which was, I would say, less theoretical in every

respect but very much more pragmatic in a way. So things are influenced by what

went before and sometimes it's: "Wow we need more of

that!", and sometimes it's "... Hmmm, we need less of that", and do something different. So

those would be examples. C++ would be another one, influenced by the wonderful

engineering that went into C -- efficiency, obvious mechanisms, and so on, but at the

same time taking what came from Simula, the idea of classes in particular. Simula

is a language that, I think, didn't get the recognition that it [deserved]. When Nygaard

got the Turing award for it, that was recognition, finally. But C++ [was] clearly

influenced by those two streams coming together. More modern example: Go, which

the pragmatic part of it, the syntax, the minimalism and so on very strongly

influenced by C, but on the other side the module structure influenced by all

of the things that came from Algol through [Ni]klaus Wirth -- Modula, Pascal,

Modula, Oberon, all of those things. And then the 'communicating sequential

processes' work of Tony Hoare. So those three streams come together and

certainly Tony Hoare's view of this was, let's call it a theoretical approach. So

all of these things come together. So, it's not like things come full-blown

from the head of Zeus. They are based on what has gone before. >> Sean: I've got a few

people saying just pass on thank you. So, thank you! >> BWK: Thank you all

>> Sean: What are your thoughts on writing 'secured by design' software in C today?

Is it better, as a result, for inexperienced programmers to use something

higher level - something else? >> BWK: I think writing secure software you should do it

by design. If you don't [i.e.] glue it on afterwards -- that won't work. You have to

think very carefully about what you do. I think the problem with C is that the

language itself doesn't provide you with many mechanisms for making sure that

your code is secure. So, there's lots of ways in which you can inadvertently make

a mistake and the compiler is not able to help you. And so for many purposes, a

higher-level language would probably be wiser because some of them may offer -- array

overflow is an obvious example. buffer overrun is one of the common

things that goes wrong with C programs. It doesn't go wrong, at least in that

sense, with any language which is manipulating strings or arrays for you,

in keeping track of the bounds of them. So, it's possible to screw up in any

language - absolutely - but I think C is by design a very, very sharp tool. And you

can cut yourself as well as cutting through things. >> Sean: Tabs or spaces? >> BWK: Tabs or spaces.

I started out with tabs because that's the way that UNIX was done. Ken

Thompson and Dennis Ritchie used tabs, everybody used tabs, there was no notion

of spaces, and this was partly the minimalism -- why type 4 characters

when you could type one? But the problem is that things tend to get wide after a

while, and especially in a language like Java which is intrinsically wide.

I think spaces are a better fit for most people, and so I think now, when I'm

writing my own code, I mostly use spaces and every once in a while I get bitten

because there's a tab when there should have been a space, or vice versa. Some

languages, in particular Go, enforce a standard format, period. And the standard

format uses tabs, and that's it. But you can display the tabs as any

number of spaces, so that the visual effect is right.

>> Sean: "Sinful [Citrus]" would like to know what you're involved in now? Innovations or

recreational ... Comp Sci? >> BWK: I'm glad you qualify it because

recreational, my wife and I are on vacation in England and having a

wonderful time, thank you. Technically, at the moment I am

trying to gear up for a class in the fall, where I'm going back after a several year

layoff to the class that I've been teaching for non-technical people. And I did

that for a long time, probably fifteen years or more. I can put in a plug for

the book, called "Understanding the Digital World"; please buy many copies for

yourself and your friends (!) But I had a three year layoff on that and so I'm

back to doing that in the fall. And of course the world changes and so there

are things that I didn't cover properly, I think, or enough, in previous go-arounds

that I want to include this time. The obvious things are things related to

machine learning, big data, natural language processing, and so what is

all of that stuff? How do you explain that for people who are not technical

but are probably going to be making decisions on behalf of those who are

technical? And so on. That's one of the things that I'm worrying about at the

moment. The other thing that I've been playing with off and on, really as a

dilettante, is issues in and around the digital humanities. We had talked earlier

about how computing spreads into lots and lots of different areas. So digital

humanities is basically doing with [a] computer the kinds of things of

analyzing data and drawing inferences from it and making it available to other

people and so on that you do with any other kind of data. But the data comes

from really, really interesting sources, often somewhat elderly. This goes back to

my interest in history. And so I have been over the last couple of years, I had

run computer science seminars, that is, seminars for computer science students

who are doing independent work in digital humanities, studying datasets

from all kinds of weird things. And I'm in addition supervising independent work --

one semester and senior thesis work -- for undergraduates on digital humanities. And

so I keep going on that, but very much as a dilettante. I could talk for a long time

about it but it would become obvious that I didn't

what I was talking about. So that would be fine and so that's another aspect. And

of course the same kinds of things about machine learning and natural language

processing and so on show up very much in digital humanities. There's a wonderful

thing here in England. It's called oldbaileyonline.org and

it's basically some university consortium, I guess, digitized all of the

records of the Old Bailey from the late 1600s until the early 1900s.

These had been court records, right; transcriptions had been taken from, as I

say, the late 1600s right into the early 1900s and they were all digitized, put

together in XML format, and now you can search them and find out things like,

well there was probably some guy named Sean Riley in the 1700s who was hanged

for stealing a pig, or something like that. It's just absolutely a wonderful

amazing database. It's close to two hundred thousand cases, very carefully

documented with the names of the perp[etrator]s and the victims and the witnesses and the

judges, and the what happened to the people when they were convicted? One

of the things I had not realized until I saw this was -- you've heard of transportation --

you know if you had been, if you'd done something bad, you were transported and

of course everybody thinks transportation meant Australia. No.

Transportation before 1776 meant you were shipped off to the United States, which

wasn't the United States at that point; it was just the colonies. And that might

explain all kinds of things.