The Classroom

Ben I.

4:09 AM

@ScottRowe Eleven!

Holy cow!

@ScottRowe I agree that we can (and do!) move past it, though I think it sets the tone for the first few years.

I started in BASIC on my Apple IIe, and I never feel even a little tempted to use GOTO anymore 😂

Rusi

9:21 AM

@VictorEijkhout Sorry about the infelicitous notation. Since I've anyway de-ASCII-ized my notation (eg changed `<-` to `∈` etc so likewise I thought: What the heck! Anyway its not literally cut-paste executable anymore. So Ive changed the type `[a]` to `ℒ.t` ie `ℒ. ...` is the `List` type.constructor and the generic (template in C++ terminology) changed from `a` to `t` should hopefully be more suggestive. At a more broad philosophical level: Re. My haskellish-notation : Its important to think of it as math-by-design and executable-by-happenstance

@ScottRowe Ravishing yes! You got it! But like much 'classic' stuff it takes a bit getting used to. I used to be crazy about Bach in my younger days; would make students listen to it if they showed any inclination. I remember after one such session: Asking "How was it?" V quietly, hesitantly the young teenager opened her eyes and said: "It hurts my ears!"

Organ at hi volume, big speakers can be some treatment for the uninitiated ears: Yet to (younger) me few things that make God more present than Bach. Which ramble is to say that for valuable things some initiation is to be expected

You would I think appreciate Dijkstra: «... and when we recognize the battle against chaos, mess, and unmastered complexity as one of computing science's major callings, we must admit that "Beauty is our Business»
https://www.cs.utexas.edu/users/EWD/transcriptions/EWD06xx/EWD697.html

Rusi

9:55 AM

As u will see @ScottRowe you have the last word 😅(read new closing )

@VictorEijkhout pls check if the new notation notes and philosophy explanations at end help

Scott Rowe

11:37 AM

@Rusi ok, interesting. I also skimmed the Dijkstra reference. One thing I have not yet understood is: if people can write down and solve all this Math stuff with pencil and paper, why do they think they need a computer? I have never seen any necessary connection between programming and math. I think - correct me if I am wrong - that a computer is an electrical machine.

And machines are usually employed to handle jobs that require more strength, speed, endurance or to withstand environments we can't. The word 'math' isn't really in there anywhere, and it seems an odd historical quirk that the development of programming fell in to schools of math or less wrongly, engineering. I haven't as best I recall used the High School or College math I was compelled to study for... basically anything in life or programming.

We use computers now to handle gargantuan amounts of data filtered via relatively simple forms and queries. I personally haven't seen induction used for anything except to solve murder mysteries. I don't understand the math-based view of life, and yes, I have created some complex programs. So I have to admit ignorance.

Guess: maybe the math-heads were the only people smart enough at the time to build the early computers?

Scott Rowe

12:03 PM

@BenI. Yes. I am concerned about how early school CS instruction can be done well, perhaps like how Buffy might be. But I am not inquiring in to that because I expect that people with more experience are working on it. I can relate to a 13 year old, maybe, but probably not a 7 year old. I do know for sure that early life is fast and furious in human development and so we have to start potential experts early, odd cases notwithstanding.

Rusi

Thats a non trivial q @ScottRowe ! If you ask it on main (or meta) I may try my hand at a detailed answer

Scott Rowe

The question "why did it start in Math departments"?

Rusi

For now just try the following: Peer into the innards of a computer:.... ARITHMETIC LOGIC unit; ADDER. What field do these terms come from??

Scott Rowe

Arithmetic is grade-school.

Rusi

So arithmetic not needed for computer creating/programming/understanding???

Scott Rowe

12:08 PM

Arithmetic is necessary. And, as best I can tell, sufficient for the vast amount of programming.

Rusi

Even to get a basic while loop with a multicondition

eg while (ii < N && a[i] != X)...

To get what happens when you're out of othe loop you have to FLIP the and into an or

Thats de Morgan law for you en.wikipedia.org/wiki/De_Morgan%27s_laws

In short math/logic are REALLY embedded into the weft and woof of our field

Scott Rowe

Yes. Also when reading program requirements: "I like coffee and skiing" doesn't mean I want them at the same time, ha ha

Rusi

🤣

Scott Rowe

"We are separated by a common language."

Rusi

But u said soemthng more significant a few hours back

You CRIED seeing the beauty of math

I feel all the "Big Guys' have had that (kind of) emotion and their work is the outpouring of that

Scott Rowe

12:16 PM

Yes, it is overwhelming. Beautiful doesn't mean I understand it. Maybe you could say. "It hurts my mind."

Rusi

See the Turing film "Imitation Game"

Its very touching (on occasion I stat a course by showing it)

Turing fell in love with his friend Christopher Morcom

Christopher died suddenly

Turing was shattered

Scott Rowe

Yes, I liked that film. I also like Bret Victor's "The Future of Programming". What ever happened to him?

Rusi

His whole life that follows is around "What it would take to create Christopher in (different) hardware now that the familiar hardware is gone

From a science/tech pov Godel and Turing are similar
But from philosophical they are quite poles apart

Scott Rowe

Yes. Well, I have had many dogs, cats, a horse... I don't think in terms of reincarnation anymore. Again, the film Arrival is the view I agree with.

Rusi

Yet the same quest for beauty
See http://blog.languager.org/2015/07/cs-history-2.html

Well sure different ppl dont click at the same thing as beautiful

Yet as Plato understood 2.5 millennia ago, the thing called beauty is real nevertheless

As for soul I heard an interesting thing about Dalai Lama

Holding his two arms wide apart DL said: Hindus... soul (atma) left hand
Buddhist no soul (right hand)

Then bringing the arms together he closed "The two are same!!"

Scott Rowe

12:27 PM

Reminds me of the quote: "You have confused the true and the real." Yes. Yes. The Two Real Things. Are One Thing! Yes. Time for breakfast.

Rusi

4:11 PM

@VictorEijkhout I was looking at your book... And the kind of place you work in... And I chuckled
No wonder you get anoyed with ppl like me talking theoretically 😂

For ppl with ur kind of work profile / machine usage I'd give Rust serious consideration (when looking in the FP direction and talking in 2022)
See https://kornel.ski/rust-c-speed

Then of course theres more bleeding edge specialized research
Id heard this guy 30 years ago https://www.cs.colostate.edu/~svr/
His polyhedral model still seems relevant and active

Victor Eijkhout

4:57 PM

@Rusi I'm actually a fairly theoretical person. See my High Perf Sci Computing book. That has for instance asymptotic analyses of why certain data decompositions don't scale. Which brings me to an important point: FP and such is completely stuck in a 1960s world where operations are all that counts. These days, data movement is much more important. (In my world, of course, maybe not in yours.) And I hardly ever see theories that take that into account.

Rusi

Yes I understand.
Do check out the colostate guy

Victor Eijkhout

@Rusi Polyhedral: yes, analysis of loop dependencies. But not many compilers get the idea to tile loops. and that is often needed for high performance. (Not to mention that all that compiler stuff assumes that each core/thread/process can access all the data. What if data can be at different distasnces, and you have to take that into account?)

Mind you, I am regularly very impressed with compilers, but there is a limit to what they can do.

Rusi

Yeah I guess so... Like @BenI. I need to concur This is not remotely my area
Ive never used a machine much more powerful lthan my i7 laptop
And 30 years ago at univ there used to be this joke that we CSists dont know how to grind a machine as the physicists do

Victor Eijkhout

Rust: yeah, on my todo list. Can all my old C/Fortran libraries immediately interface to them? Do Rust compilers obey standard linker conventions?

Rusi

The best we can do is self-compile gcc

Those guys OTOH had some sort of cluster (probly a toy by your standards)

When they got a larger cluster they would one more electron (!!!!) to their simulation and up the 25 hr grind to 50

My knowledge is 2nd/3rd hand.

But its preparing to enter the Linux kernel zdnet.com/article/…

Whilch is BIG step consiering how staunchly Linus T has opposed any talk of replacing C for all these 30 odd years

Which basically means I expect interop with C to be either good or hi priority

This chat is annoying

Somethinhg I wrote is cut off!!

Rusi

5:27 PM

I was going to say something re quora.com/…

Then saw you have answered there 😁

Still... The answer by Tikhon Jelvis distinguishing first order second order and DSL creation&use is worth ur consideration

G Night

Victor Eijkhout

@Rusi I should download my quora answers. There is occasional good stuff there. Sleep tight.

Rusi

@VictorEijkhout Pls do check out my emendations to ur question
Particularly the Notation notes and philosophy at end
After that I'll consistent-ify both answers

GN!

Rusi

6:00 PM

@ScottRowe your question which I said is worth considering on main or meta «One thing I have not yet understood is: if people can write down and solve all this Math stuff with pencil and paper, why do they think they need a computer?»

GN!

Buffy

7:41 PM

@Rusi Note that «One thing I have not yet understood is: if people can write down and solve all this Math stuff with pencil and paper, why do they think they need a computer?» is not a question for meta. Meta is about this site and its operation, not about content on the site. That might be a question on one of the math sites (but not meta there either).

Ben I.

8:15 PM

@ScottRowe But translating arithmetic into exclusively boolean operations is a mathematical exercise, and had to be done before there were electronic computers. Similarly, bases, which have now become part of engineering, were pretty esoteric math. Binary equivalents to booleans, then using boolean algebra to simulate calculation of numbers, sounds like pure math to me.

Especially because these were not "practical" in any sense when this was all developed, which is where engineering tends to step in

But the real mathematical story that I love is much more exciting, and starts with Godel and winds up at Von Neumann.

They didn't set out to create a computer, exactly. They were concerned with fabric-of-the-universe stuff. They wanted to figure out what was knowable and what was definitely unknowable.

user2236

Does the CS mandate require children to learn typing as early as possible? @BenI.

Ben I.

@user2236 In my state, typing is not part of the CS mandate at all.

user2236

hmm

Ben I.

In fact, they took pains to make sure that teaching things like Word, Powerpoint, typing, and internet safety could not be used by schools to dodge the requirement for CS

user2236

sounds like common core

Scott Rowe

8:30 PM

Microsoft products now include very good dictation, for free. Try it if you have 365. So, forget typing when you think of children, they won't bother.

Ben I.

@user2236 I'm not sure what you mean by "common core" since the word got politicized enough that everyone who doesn't use curriculums for a living seemed to suddenly forget what it was. But it is a curriculum, so it is the same mechanism and structure as other curricula, yeah

@ScottRowe Computers were an almost accidental outgrowth of the mathematicians' deeper question.

They made some headway into their question, also, but only a bit

Buffy

@user2236 Hmmm. Typing can mean pressing keys on a keyboard or it can mean type systems in computer languages. Which?

user2236

Pressing keys.

Memorizing the keyboard etc.

Scott Rowe

@BenI. In a practical sense, Yoga and a proper meditation are the best bet for getting this answer. I guess that sounds flippant here.

Buffy

I'm so old I learned that in high school and I had to do it in a room dedicated to teaching secretarial work as a job.

I depended on a mechanical typewriter for years. My doctoral dissertation in math was typed on such a machine with a way to insert special characters. They were on short metal arms, stored in a box. Selected and inserted into the typewriter and then struck with some key.

Victor Eijkhout

8:38 PM

@BenI. Computers were not an outgrowth of anything mathematical. Turing machines were. Actual computers were built to integrate ordinary differential equations so that the miiltary could compute artillery trajectories. What do the "NI" in ENIAC stand for?

Buffy

We paid one of the departmental secretaries to type those.

Scott Rowe

Rusi wanted to ask 'my' question... I was responding to the Dijkstra reference, where he shows all these steps to solve a math question, and I always wonder why Mathematicians don't show actual practical problems that really require a computer.

Buffy

@VictorEijkhout How is "integrate ordinary differential equations" not mathematics???

Ben I.

@VictorEijkhout I was led to understand that Von Neumann's architecture was an attempt to create an electronic Turing Machine.

Buffy

And don't confuse analog computers with digital. Analog computers aren't Turing machines, actually, and produce only approximate answers.

Ben I.

8:40 PM

Is that a false history?

Buffy

The digital computer preceded Von Neumann. His advance was storing the program in digital form along with the data.

But that is different from a TM, where the "program" isn't necessarily on the tape.

Ben I.

Not to mention that von Neumann was, again, a mathematician.

Buffy

For a universal TM it is.

@BenI. As was Turing - a mathematician.

Ben I.

Yeah, they all were. It was the only field that could have created this stuff

Buffy

@ScottRowe Actually, there are a few problems for which a computer was essential in the proof. The four color theorem is an example, and IIRC, Fermat's Last Theorem.

Scott Rowe

8:48 PM

But we had decades of electronics before that, with very smart people going back to Faraday. And we had tabulating machines doing the Census and so on. It was only a matter of time before computers would be built. The Mathematics emphasis is just a big accident.

Yes, so they should demonstrate those problems.

Victor Eijkhout

9:44 PM

@Buffy You might say i't scomputational physics. Computers were built to solve physics problems. Stuff like computability is fairly independent of computing machinery. Von Neuman: I wonder how much he put his name on electronics reports written by engineers working under him. He understood mathematics, he understood numerical methods, I wonder how good he was at vacuum tubes.

Scott Rowe

I found math classes very alienating, so I went towards something I could relate to: electronics. There are these parts and you connect them, they get warm and sound comes out. TV was fascinating, until I read how a kid plowing a farm field solved that problem. Pretty concrete stuff. Computers just let you tell them the steps, like training an animal, and they do it, no math required.

Transcript for