I had the good fortune to have had Dr. Dewar as my advisor when I was at Courant. He was an amazing , if underrated Computer Scientist. He had a practical view toward programming and compiler development and his word revolved around situating languages and computer systems in the real world. He is without doubt the best lecturer I have ever seen and his classes were amazing.
I love this lecture so much. There is a story behind Ada and where this language exists in the current ecosystem that is just *begging* to be told. This lecture is the primary source material. But so many others like Tuck Taft and Randy Brukardt and many others that I could mention… could tell this story so well.
I especially liked the sequence beginning @21:48 - Where he details the costs, in orders of magnitude, of bugs getting by compile time to initial run-time to beta and even production deployment. Gotta Love Python, Perl, and other languages that let us claw our way through keyboard to deployment, but one must always consider his point about, and I paraphrase, "dificulty in getting code to compile translating into stable run-time". One particularly humorous moment following that sequence is how he referred to junk programming languages and then used C as an example of something that will let you incorporate bad design, and then checking himself by saying that he shouldn't editorialize on that point lolz.... He obviously likes structure to the point of tighty-whitey's, but by the same token, most contemporary systems level developement for drivers and even OS development (i.e., LInux) makes use of C (and for the reasons that both he and Linux Torvalds have elaborated on - NOT C++), which is superior in many ways if we just adhere to proper testing and oversight. PLI, and ALGOL also were quite stringent, often cited as producing code that, unlike C, is less prone to entering production with vulnerabilities and such. @AdaCore05: You say the lecture was filmed in early 2006, but the professor cites Ada 2005 as being hopefully on track for certification by 'that time'. It appears from his presentation that this lecture was actually taped before 2005 - so are you sure it was filmed in 2006?
Having spent over a decade with C and now a number of happier years with Ada. C is not superior to Ada for low level devlopment in any way. Quite the opposite. Ada makes handling memory register and network protocols safer in many ways whilst removing much of the complexity and obfuscation that C programs inevitably create. There is even built in validity checking.
I like his comment “mostly unclever code that is easy to read, easy to understand is what we are looking for”. You definitely see intelligent programmers trying to do the opposite. This talk was given before 2005, as Ada 2005 was a future event. And given after after 1995 because Java is available.
"I will, and won't, be inspired to learn how to write it."? I think it's a good name, and the name is not that important for anyone that is serious about using a tool such as a programming language. Likely only hobbyists and predjudice/ignorant people would overlook something of use to them just because of it's name.
I had the good fortune to have had Dr. Dewar as my advisor when I was at Courant. He was an amazing , if underrated Computer Scientist. He had a practical view toward programming and compiler development and his word revolved around situating languages and computer systems in the real world. He is without doubt the best lecturer I have ever seen and his classes were amazing.
Never thought an Ada programmer would be so funny
Bjarne Stroustrup is so much funnier.
I love this lecture so much. There is a story behind Ada and where this language exists in the current ecosystem that is just *begging* to be told. This lecture is the primary source material. But so many others like Tuck Taft and Randy Brukardt and many others that I could mention… could tell this story so well.
compare list of ADA usage @ 35:12 to the reality that F 35 is using c++ and is barely flying.
I especially liked the sequence beginning @21:48 - Where he details the costs, in orders of magnitude, of bugs getting by compile time to initial run-time to beta and even production deployment. Gotta Love Python, Perl, and other languages that let us claw our way through keyboard to deployment, but one must always consider his point about, and I paraphrase, "dificulty in getting code to compile translating into stable run-time".
One particularly humorous moment following that sequence is how he referred to junk programming languages and then used C as an example of something that will let you incorporate bad design, and then checking himself by saying that he shouldn't editorialize on that point lolz.... He obviously likes structure to the point of tighty-whitey's, but by the same token, most contemporary systems level developement for drivers and even OS development (i.e., LInux) makes use of C (and for the reasons that both he and Linux Torvalds have elaborated on - NOT C++), which is superior in many ways if we just adhere to proper testing and oversight.
PLI, and ALGOL also were quite stringent, often cited as producing code that, unlike C, is less prone to entering production with vulnerabilities and such.
@AdaCore05: You say the lecture was filmed in early 2006, but the professor cites Ada 2005 as being hopefully on track for certification by 'that time'. It appears from his presentation that this lecture was actually taped before 2005 - so are you sure it was filmed in 2006?
Having spent over a decade with C and now a number of happier years with Ada. C is not superior to Ada for low level devlopment in any way. Quite the opposite. Ada makes handling memory register and network protocols safer in many ways whilst removing much of the complexity and obfuscation that C programs inevitably create. There is even built in validity checking.
Somewhere during this talk, he managed to surreptitiously fixed his collar and tie.
LOL! Great presentation. Funny.
I like his comment “mostly unclever code that is easy to read, easy to understand is what we are looking for”. You definitely see intelligent programmers trying to do the opposite.
This talk was given before 2005, as Ada 2005 was a future event. And given after after 1995 because Java is available.
17:40 golang LUL
Great lecture ❤
yay, i guessed right! approx. 1 line per day constructive code in a safety critical application.
What was this stack-based "amp chip" he said was used extensively in aerospace?
Rockwell Collins Advanced Architecture Microprocessor, probably.
RTX144 probably
❤
"I will, and won't, be inspired to learn how to write it."? I think it's a good name, and the name is not that important for anyone that is serious about using a tool such as a programming language. Likely only hobbyists and predjudice/ignorant people would overlook something of use to them just because of it's name.
Incorrect; the RM has always allowed for Garbage Collection, but never required.