These are really some of the best lectures on TH-cam. I already shared a bunch with my friends. Everyone watching who appreciates these should do the same! (please record more!!)
other people: make a click-bait thumbnail about an epic chili eating video, build it up for 15 minutes, finally eat it and make a fuss about how it hurts, etc Mr. Ackley: describes computing and 'technical debt' as if they were the fundamental building blocks of the universe while exuding a generally zen philosophy, then eats the chili without warning.
I took a CS course taught by Professor Ackley at UNM. I can honestly say that he was my favorite CS professor. He is an incredibly articulate and fair professor and while his course was taught in C++ -- a language I don't use professionally or personally today -- the challenging course material coupled with his teaching ability really pushed me forward and galvanized my interest in problem solving through programming. I really enjoyed this talk and the presentation style, and look forward to seeing the rest of the series. I have to admit, though, that without his emacs editor open in text-mode and projected on a screen, I felt like something was missing here ;) I'm also pretty sure I'm going to have a green chile cheeseburger tonight...
Speaking of CS instruction, Dave once subbed for my 351 professor, Terran Lane, and that was my only encounter with him. In retrospect, his lesson on design patterns in general and MVC specifically was more valuable and formative than anything else I learned during my undergrad. My fellow programmers at Aecom and Intel would tell you how obstinately I insist on programming to patterns now :) P.S. I might have learned some useful stuff from Terran too, but I'm still bitter that he marked one of my programs down a letter grade because I wrote possessive "its" with an apostrophe in some comments.
That whole "never think of optimizing up-front" thing is kind of being replaced by a simple awareness of the whole technical-currency scene. Brilliant.
***** Yup, it works perfectly as 'reduction to absurdity', cause one to ask how we ever thought such a model could scale. And I think Tufte would fully agree with your solution.
Could you quantify what you are losing/sacrificing with these increasing levels of abstraction? Could you say that by reducing size to a 1-bit variable you have sacrificed X amount of...ummm...precision? information? state? You yourself gave the example that the habenero is particularly difficult to classify as red vs. green, so slotting it into a red/green "projection" (if I may be so bold?) of the color dimension incurs a significant loss of something that I can't quite come up with a word for here. I wonder if you could actually measure that loss and say that in your hyperspace one dimension (length) requires a loss of X "bits" (or whatever this property is) while another dimension (color) more severely requires a loss of Y "bits". One way of viewing this is that length was already one-dimensional so projecting it to a one-bit value is less "severe" than doing so with color which is commonly three-dimensional in most human vision...or something like that. Also, I think a 3-light-year chile would "seat four" quite comfortably.
I'd think with sufficient additional assumptions one could quantify the information loss in going to a single bit. But that's kind of looking the wrong way -- in hyperspace we can add additional predicates on additional bits to make whatever finer distinctions we choose to implement. A three-light-year chili would be a really Big Jim.
I see, so you could have a red/green predicate and then a red/orange predicate and a green/orange predicate and a unique combination of those three bits would capture a finer gradation of the over-arching color dimension. Of course, that approach simply adds back all the complexity that was just dispensed via reduction to a single bit in the first place, and more to the point you then have categorically related predicates (all the color predicates) which begs for some sort of predicate grouping. I'm over-thinking it, don't worry about it. Good video, looking forward to others.
I could watch these videos all day... Like a computational Bob Ross in terms of presentation. Wonderful videos.
"Computational Bob Ross" hahahahaha so so true
These are really some of the best lectures on TH-cam. I already shared a bunch with my friends. Everyone watching who appreciates these should do the same! (please record more!!)
Thanks for the comment and the shares! Shall see, but hoping to do a couple more this summer.
other people: make a click-bait thumbnail about an epic chili eating video, build it up for 15 minutes, finally eat it and make a fuss about how it hurts, etc
Mr. Ackley: describes computing and 'technical debt' as if they were the fundamental building blocks of the universe while exuding a generally zen philosophy, then eats the chili without warning.
I took a CS course taught by Professor Ackley at UNM. I can honestly say that he was my favorite CS professor. He is an incredibly articulate and fair professor and while his course was taught in C++ -- a language I don't use professionally or personally today -- the challenging course material coupled with his teaching ability really pushed me forward and galvanized my interest in problem solving through programming. I really enjoyed this talk and the presentation style, and look forward to seeing the rest of the series. I have to admit, though, that without his emacs editor open in text-mode and projected on a screen, I felt like something was missing here ;) I'm also pretty sure I'm going to have a green chile cheeseburger tonight...
Hey! Thanks for the share and the memories! I'm trying to wean myself off of lecturing with emacs, but.. ..I'm still kind of doing it.
Speaking of CS instruction, Dave once subbed for my 351 professor, Terran Lane, and that was my only encounter with him. In retrospect, his lesson on design patterns in general and MVC specifically was more valuable and formative than anything else I learned during my undergrad. My fellow programmers at Aecom and Intel would tell you how obstinately I insist on programming to patterns now :)
P.S. I might have learned some useful stuff from Terran too, but I'm still bitter that he marked one of my programs down a letter grade because I wrote possessive "its" with an apostrophe in some comments.
That whole "never think of optimizing up-front" thing is kind of being replaced by a simple awareness of the whole technical-currency scene.
Brilliant.
I've never heard the term "technical debt" before, but now that I've heard it it just seems to fit so well. Thanks for introducing it to me!
Great class!! Cheers from Venezuela
Great as always! Plus in this one, at 8:58 we see the best Chili Data Visualization ever. No wait not "best", what's the other one? :)
Gary D Wait, that was supposed to be the worst.. :)
***** Yup, it works perfectly as 'reduction to absurdity', cause one to ask how we ever thought such a model could scale. And I think Tufte would fully agree with your solution.
I know this is from some years ago, but was this series ever continued, maybe on another platform? This lecture is so good!
Thanks for the feedback! It hasn't continued yet, but I've been working another part of it just in the last month! We'll see!
@@T2TileProject When I saw this reply, I just finished watching this:
th-cam.com/video/Ag0uTQkJH9A/w-d-xo.html
Very interesting stuff! Thanks!
hrrmm world's slowest series perhaps th-cam.com/video/ScYgBxLupAs/w-d-xo.html
Sometimes you can have things be limitless but with logarithmic curve.
Could you quantify what you are losing/sacrificing with these increasing levels of abstraction? Could you say that by reducing size to a 1-bit variable you have sacrificed X amount of...ummm...precision? information? state? You yourself gave the example that the habenero is particularly difficult to classify as red vs. green, so slotting it into a red/green "projection" (if I may be so bold?) of the color dimension incurs a significant loss of something that I can't quite come up with a word for here. I wonder if you could actually measure that loss and say that in your hyperspace one dimension (length) requires a loss of X "bits" (or whatever this property is) while another dimension (color) more severely requires a loss of Y "bits". One way of viewing this is that length was already one-dimensional so projecting it to a one-bit value is less "severe" than doing so with color which is commonly three-dimensional in most human vision...or something like that.
Also, I think a 3-light-year chile would "seat four" quite comfortably.
I'd think with sufficient additional assumptions one could quantify the information loss in going to a single bit. But that's kind of looking the wrong way -- in hyperspace we can add additional predicates on additional bits to make whatever finer distinctions we choose to implement.
A three-light-year chili would be a really Big Jim.
I see, so you could have a red/green predicate and then a red/orange predicate and a green/orange predicate and a unique combination of those three bits would capture a finer gradation of the over-arching color dimension. Of course, that approach simply adds back all the complexity that was just dispensed via reduction to a single bit in the first place, and more to the point you then have categorically related predicates (all the color predicates) which begs for some sort of predicate grouping.
I'm over-thinking it, don't worry about it.
Good video, looking forward to others.
Thanks for your thoughts. Indeed, we will address 'some sort of predicate grouping' in part 2!
Ah, yes. Finally
:) Thanks for noticing.
"the arrow goes on for ever" the system is scaleable!! :D
Ah, 10 seconds later that point is made
Doesn't Quantum Mechanics use reals
Colour me subscribed!