The next Code Mesh Conference will be on 8 - 9 November 2017 (with Workshops on 7 November) - subscribe to receive exclusive content, updates and benefits.
What an amazing talk! It's nice to see these two titans interact on the stage, each one with its own personality. I don't know the type of kool-aid that Alan is drinking every day, but for his age, this guy is astonishing.
“The problem with object-oriented languages is they’ve got all this implicit environment that they carry around with them. You wanted a banana but what you got was a gorilla holding the banana and the entire jungle.”-Joe Armstrong Alan Kay got one back by just skimming over Erlang ...
There are two versions of OOP. The Simula version which puts emphasis on classes and inheritance, and the Smalltalk version, which is about message passing. Joe was referring to the first version.
51:10 onwards for those who’re looking for making sense and meaning in life. Look at the data (whatever you’re being presented) but pay attention to what the data is not-indicating. One of the most enlightening talk on sense building, of higher order in creation and possibly keeping sane in a crazy world.
Smalltalk was excellent for its time and could still rival many programming languages today. What Erlang got right is isolation. Smalltalk aspired to it but it never really reached that goal. That said Erlang had more than a decade of advances to benefit from so it's hardly apples vs apples. Not to mention Smalltalk was really good at doing user interfaces. Something Erlang really struggles with (unfortunately). I think a next gen assembly might be a chimera between Smalltalk, Erlang, E, Elm, the Internet and this idea of introspection that exists in spirit in Joe Armstrong's UBF and some LISP meta object protocols (dealing with Aliens or dynamic m2m).
thecount25 Don't you think we had enough chimeras? Mr. Kay is the first to say that Smalltalk leaves a lot to be desired. It hasn't really been improved upon in a fundamental or even big way though.
Also erlang had allot of support financially but small talk was too and still is challenging because it forces you to think differently which is really difficult. It's not perfect for all situations but I think the time is now. There's plenty of room for all methods. Unfortunately money does matter, you get better qa and quality controls.
Ive seen a lot of talks and interviews that Alan Kay has given on youtube and Ive always been impressed by his low key and approachable style of explaining things at their simplest level. The only thing that I thought was missing was him opening up a dialogue with people in the tech industry to help guide in applying his design principals and ideas. So I was looking forward to seeing this interview and how Alan would react to Erlang, thinking he would react positively to a programming language that applies a lot of his ideas. Such was not the case, to be honest, that wasnt an interview, or even a conversation, it was a lecture driven by live prompting of topics by Armstrong. Alan was looking at the audience the entire time and giving a talk to them. Behind Alan's timid and soft spoken demeanor we finally see the pompous egotistical side of Alan Kay show itself when he is forced in a situation where he has to share ideas instead of pontificate on them, he simply refused dialogue and made it speech entirely about him. Which is not entirely a bad thing, what Alan Kay does best is knock people out of stale dialogue and force people to look at what they really should be looking at. But he totally ignored Joe and made Erlang a mere footnote in his speech. Still a lot to learn, but kind of sad that Alan cant share the spotlight.
You two really need to work on your comprehension skills and your history knowledge. Basically the stuff included here. What a wonderful talk! Alan should really be working on his memoirs he can tell some fantastic stories. Feynman and Douglas Adams as personal friends. Holy crap!
He certainly has an ego, but who wouldn't with the amount of knowledge and wisdom he has? His world view, even when restricted to computer science and programming languages, is so broad (particularly since he participated in and shaped a lot of it) that for him to refer to Erlang only briefly is understandable. I agree Alan tends to pontificate and seems to intentionally remain vague about his ideas sometimes but at the level he's thinking (and considering how far off modern comp sci has gone from Alan et al's visions) I really can't blame him
Now that you mention... I have long wondered about the exact motivation for Alan's famous dig on Dijkstra ("arrogance in computer science is measured in nano-Dijkstras."). Maybe it's a case of pot & kettle?
why, he says he hates that kind of stuff as too mechanistic, too gear-like -- and eventually too religious (functions/monads). also, he praises Sketchpad for its "imprecise engineering-based (with tolerances) programming that can't fail", "no theorem-provers!", etc. I guess he thinks that whole thing is a dead end, a "subproblem that we substituted for the real problem" perhaps. ----- (what's CLEAR?)
Perhaps he meant clean? en.wikipedia.org/wiki/Clean_(programming_language) though im not too knowledgeable on the history of programming languages, and am wondering why he mentioned it. (usually ISWIM or ML are sufficient symoblic/canonical terms addressing this family of languages)
True! It's not a surprise that he is dismissive of that stuff, given his professed views. I guess what I think is fascinating is the fact that he could be so dismissive of that entire branch of things! From my perspective, it is totally obvious that the half-century of research in that direction has been fruitful, whether or not you think it's an entire solution to the "real problem". It's interesting to me that one can have such an open mind and such a closed one at the same time. He at least took the time to sneer at logic programming :) re: Clear: "The module systems of Ada, ML, C++ and Lotos have all been influenced by the OBJ module system; Lotos also uses the initial algebra semantics that was pioneered by OBJ. The OBJ module system ideas are a further development of ideas pioneered in the Clear language, which was joint work of Joseph Goguen and Rod Burstall in the 1970s." -- cseweb.ucsd.edu/~goguen/sys/obj.html
fruitful, yes, but still "mechanistic". I think the inevitable "self-healing" self-compiling adaptable evolving systems might use this stuff as their back-end, locally (i.e. each node for its lower-level needs), that's where its place is, I think. --- thanks for the link!
organisms are not mechanistic; their smallest building blocks are. Most important distinction is, nobody would program the "live" systems down to minute details, like we are forced to do, still, while using the precise and mechanistic math-based languages. Using a language that talks about every atom's exact position is not useful in dealing with fluids. But of course fluids are made of atoms.
Hmm, while I really agree with most of his sentiments, I think Alan Kay kind of misses the mark on what the "Monad" people are trying to do. In my mind they are trying to do exactly what he's advocating, by thinking of how to express the general patterns, i.e. they advance the theory that the written form of programming languages constrains the thinking and that if you borrow from category theory the language will help you to see deeper patterns, I don't think it's fair to say that it's practiced like a religion, though that's certainly true of some people as in all fields. It's tempting to say that because a typical part of learning the language is a mechanical application of these tools in order to learn the dialect, just like learning any mathematical subject, you need some suspension of disbelief until you fake your way into using it, and then it just becomes another lens to see things through. I do believe that was always the intent of the theorists behind it.
Why call them Monads and not EncapsulatedBinders, FlatMappables, WrappedAppliers, WrappedPipes? Also static typing that can catch 50-80% of errors at compile time like in Haskell requires that you categorize every single interaction with the right type, and the Haskell developers themselves with their 160+ IQ seem to get it wrong some times.
@ Python has plenty of complexity today, for example records can be implemented as tuples, namedtuples, data classes, dictionaries, and classes. A lot of complexity of Python is shared with other Algol languages, while Haskell has new set of complexities as they try to statically type every single pattern of mutable state interaction. Static typing does make coding more complicated, as you need a category for every type of interaction although sometimes Haskell can help you by inferring the types.
"just like learning any mathematical subject, you need some suspension of disbelief until you fake your way into using it" - this concept has a name: Wittgenstein's Ladder. Instead of incessantly entering some variant of an ivory tower, I think that we should all acknowledge pragmatism as a de facto guiding principle: We code, and if the code works, we stick to it as long as we can.
He was a media theorist that coined the term, "The Medium Is The Message/Massage" He's great. Worth looking into his work if you'd like some insight into computers as a dynamic media format.
@@beatresistance6090 From where I see it the joy of this talk is the tension between two programming language creators, with related concepts, implementing and defending their decisions in real-time. Lecturers seldom see their ideas challenged on stage, and here we witness the result
@@beatresistance6090 I think it's hard for the un-initiated to be sympathetic. What this comes off to to me is, someone who is very excited, very sympathetic, & who has a real pulse on Alan & is trying to evince him to offer up more. His enthusiasm is bubbling over, it's Joe's moment, and he's excited, and happy to finally get to make this happen, and it takes pushing, but he does find the ground to push Alan into interesting conversations, into some deeper topics, even though Alan is not readily wanting to go there at first. It's imperfect, that itself is so inline with Joe's best nature: that we are imperfect, and that true resiliency is about coping & soldiering on, about willingness to make the best of things. This interview may not be the best for you or I. But it is a very valiant go at bringing out as much as he can, by bringing 110% to the table, always pushing for more, better, deeper, and that enthusiasm and energy and excitement- I would like to see more of it in the world.
@@rektide well written, but sorry this was interrupting and not really interviewing. period. See this moment out of many at 1:3:26 ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-fhOHn9TClXY.html . you call this good interviewing ?
That is because you are a disrespectful, ageist, twat. If you are lucky you will live long enough that one day you will be an old fart that nobody cares about either. It will come to you faster than you think. Except unlike Alan Kay you will never have done anything interesting during your short stay on Earth.
