Hellow guys, Welcome to my website, and you are watching How do you implement lazy evaluation?. and this vIdeo is uploaded by Eric Normand at 2019-05-02T10:07:54-07:00. We are pramote this video only for entertainment and educational perpose only. So, I hop you like our website.
| Name | How do you implement lazy evaluation? |
| Video Uploader | Video From Eric Normand |
| Upload Date | This Video Uploaded At 02-05-2019 17:07:54 |
| Video Discription | Lazy evaluation is easily implemented in any language that can create first-class computations. That means functions or objects. In this episode, I explain how to implement a Delay, which is a reusable lazy component that is common in functional programming languages. ►► Audio, Video, and Transcript available: https://lispcast.com/how-do-you-implement-lazy-evaluation/ ►► Subscribe on iTunes: https://itunes.apple.com/us/podcast/thoughts-on-functional-programming-podcast-by-eric/id1364166414 Transcript How do you implement lazy evaluation, even if your language doesn't support it out of the box? By the end of this episode, you'll be able to implement lazy evaluation in your favorite language. My name is Eric Normand and I help people thrive with functional programming. Lazy evaluation is important because it's great for separating two concerns. One concern is how you generate a value, so the code for generating the value. The other concern is how much of that value do you need? Are you even going to need it at all? You have one part of your code say, "Here is how you generate the value," and another part of your code that decides, given that value, how much of it to actually use or whether to use it at all. Let's go over this recipe for how you would build this thing. What we're going to be building, in spoken code, is called a delay. A delay is an object that represents one lazily evaluated value. It represents the value but it hasn't been calculated yet, or maybe it has. It doesn't matter. From the outside, there's just one interface method on that object and it will give you the value. It's an object. Whatever you use in your language to build a little object. I'm going to assume something that has a constructor. The constructor is going to take a calculation and that calculation is what generates the value. Let's give a really dumb example. If I need to calculate the value five plus four but I don't want to do it right now, I want to delay it, I would pass the calculation five plus four to the constructor. How do you make a calculation? How do you pass it? That depends on your language. In something like JavaScript, I would wrap that up in a function. I'd make a function that just returns five plus four. The function has zero arguments. What functional programmers call that, a function of zero arguments -- it's really not a function at all -- it's always going to return the same value. It's practical because it's something that hasn't run yet but you could run it. This is called a thunk. A function of zero arguments. If it's a pure function, it's always going to return the same thing because there's no arguments. It's called a thunk. I think it's supposed to be the incorrect past tense of to think, "I thunk about it." Anyway, it's a thunk. The only thing you can do with it is call it and it'll calculate the answer. When I call this thunk, it'll return nine to me. Until I call it, it hasn't done that addition yet. That's the constructor for the delay object. It takes a thunk. If you're in Java and you don't have...maybe Java 8 does have first class functions now. I don't want to argue about that. In other languages it's going to be an object with a single method called invoke or run, something like that. It's not going to take any arguments. It's just going to finally run the thing. That would be your thunk. Now this object, besides the constructor, it has one method. Doesn't matter what you call it, it could be called get or get value. This is what the method does. It's going to first check whether you have already calculated the value. If you have then it's going to return that value. It has to have saved it somewhere, like some field in the object. If it has not already calculated the value, it's going to run the thunk. It's going to call, invoke or whatever method it is or if it's a function it's going to call it. It's going to take the return value from that function. It's going to save it and then it's going to return the value. The first time you run this method, it will have to call the thunk. The second time though, it will already be saved so you won't need to call it again. What we've done with this...that's it, that's done. What we've done is we've made a reusable and generic object that can work for any calculation you might want to delay. This is a standard functional programming tool, a little reusable thing. It takes a calculation and it delays it. It's just one thing. It only works on functions of zero arguments. Because it does not remember. It's not a memoized but it's like memoized but only for functions of zero arguments. Memoize is another thing we can talk about in a future episode. I just wrote that down, so we'll get to that. |
| Category | Science & Technology |
| Tags | functional programming | software engineering | clojure | java | javascript | haskell | functional programming tutorial | functional programming languages | functional programming course | theory of functional programming | functional programming advantages | functional programming benefits |
