Kako 5-godišnjaku objasniti koncepte kodiranja poput strujanja, obećanja, povezivanja i deklarativnog programiranja
Volim razmišljati o konceptima kodiranja uspoređujući ih sa poznatim stvarima koje znamo u životu. Postoji toliko mnogo analogija o konceptima kodiranja. Neki od njih su dobri, dok su drugi zbunjujući, uglavnom zato što se usredotočuju na djelomične aspekte koncepta, a zanemaruju mnoge. Ovaj će članak sažeti neke analogije za koje smatram da najbolje odgovaraju nekoliko koncepata kodiranja na cjelovit način.
Ažuriranje: Ovaj je članak sada dio moje knjige "Profesionalni programer". Pročitajte ažuriranu verziju ovog sadržaja i više savjeta o programiranju na
jscomplete.com/pro-programmer .
Počet ću s jednostavnim konceptima i prijeći na one teže. Krenimo od samog kodiranja. Kodiranje se može usporediti s pisanjem recepata za kuhanje. Recept u ovoj analogiji je program, a kuharica je računalo. Recept je popis uputa koje kuhar treba slijediti, a program popis uputa za izvršenje računala.
Ovo je vrlo jednostavna analogija s obzirom na to da je recept napisan na ljudskom jeziku, a program na računalnom jeziku, a to su vrlo različiti jezici (osim ako vaši recepti imaju zatvaranja i obećanja!). U receptu također nema puno neočekivanih stvari koje trebate planirati, dok će ih računalni program imati mnogo. Bez obzira na njegovu jednostavnost, dobar je način pokazati kako računalo sekvencijalno provodi popis uputa. Također pokazuje gdje jedna linija uputa može koristiti bilo koji rezultat izvršavanja prethodnih linija uputa.
Neki će recepti čak imati i if-izjave: ako se kuha za 2, 4 ili 8! Neki će recepti imati petlje: tucite tu smjesu sve dok ...
Ova analogija mi se sviđa i zbog svih spremnih predmeta i alata koje možete koristiti u svojim receptima - poput smjese za torte od koje možete raditi kolače i one posude posebno oblikovane koja toliko olakšava stvaranje kolača.
Korištenje gotovih predmeta i alata je poput uključivanja i korištenja paketa koda koji su drugi napisali u vaš vlastiti kôd.
// The making of a cupcake// First steps:$ npm install cake-mix$ npm install cupcake-panNPM je upravitelj paketa za Node.js , koji je vrlo popularan okvir za pisanje JavaScript aplikacija. U ovoj analogiji, Node.js je poput same kuhinje. Omogućuje vam izvršavanje linija u vašim receptima pomoću ugrađenih modula poput pećnice i sudopera.
Govoreći o nezdravoj hrani, sljedeća je analogija za učenje kodiranja i uspoređuje se s prehrambenim navikama. Posebno OBOŽAVAM ovu analogiju i ono što ona prenosi jer mi pomaže da ostanem na putu na svom putu učenja koda. Za mene je ovo počelo u srednjoj školi i trajat će sve dok moj mozak ne dođe do posljednjih uputa: die ();
Učenje kodiranja
Učenje kodiranja je poput pokušaja mršavljenja. Ova se analogija odnosi na stvarno učenje bilo čega, ali učenje kodiranja ovdje je posebna stvar.
"Gubitak kilograma" negativan je pojam. Doista bismo ga trebali zvati "Sticanje zdravlja". U tom je smislu vrlo usporediv sa „Sticanjem znanja“. Obrazovni resursi koji su vam na raspolaganju slični su mogućnostima hrane. Neki su sasvim u redu, neki su izvrsni, a neki su potpuno loši za vas. Zdrava prehrana i vježbanje glavne su dvije aktivnosti koje će vam pomoći da steknete zdravlje. Slično tome, trošenje dobrih obrazovnih resursa i ručno vježbanje kodiranja glavne su dvije aktivnosti koje će vam pomoći da steknete dobro znanje o kodiranju.
Pa kako naučiti "zdravo"? Kad se obvežete da ćete se hraniti zdravo, upotrebljavate filtre poput organskih , lokalnih , masnih kiselina , hranjenih travom i ne-gmo. Potpuno je isto sa zdravim obrazovnim resursima, osim što ove oznake još nisu toliko jasne. Nadam se da će obrazovni resursi jednog dana također imati provjerljive i relevantne oznake. Možda etikete poput "nesponzorirano", "ne-marketinško", "odobreno od strane stručnjaka", "čvrsto uređeno" i "zmajevi-naprijed".
Ipak, umjesto filtriranja prema sadržaju, lako možete filtrirati prema dobrim robnim markama. To radim i s hranom. Znam i vjerujem nekoliko marki i uglavnom ih koristim. Lakše je. Uz obrazovne resurse postoje neke marke (publikacije i ljudi) koje biste trebali stalno pratiti.
Nakon filtriranja unosa znanja samo na dobre resurse, samo trebate vježbati! Vježbajte sve što naučite, ali ne samo ponavljajući točno ono što ste naučili. Izazovite se i da napravite nešto malo drugačije oko tema koje ste naučili. Ako imate sreće, zapećete! Tada ćete trajno naučiti nešto drugo kad se odlijepite.
Vježbanje je i za tijelo i za um.
Varijable
Varijable se koriste u računalnim programima za držanje podataka . Ovo je vrlo pojednostavljena izjava i po mnogim je mjerama jednostavno pogrešna.
Varijable ne sadrže podatke. Oni samo ukazuju na to. Podaci se čuvaju u memoriji računala. Varijable možete usporediti s oznakama koje postavljate na poruke e-pošte (ili bilješke ili datoteke).
Svi uzorci koda u ovom članku napisani su na JavaScript-u. JavaScript je vrlo lak za naučiti računalni jezik.
U Gmailu je oznaka pokazivač na e-poštu ili popis e-adresa. Mnogo naljepnica može upućivati na istu e-poštu. Ovo je slično dodjeli druge varijable postojećoj varijabli:
let work = [email1, email2, email3];let important = work;I rad i važno sada su oznake koje upućuju na potpuno isti popis e-adresa.
Neke varijable predstavljaju konstantne reference . Ne mogu se mijenjati. Ovo je poput oznake " poslano " u Gmailu. Iako možemo promijeniti radnu oznaku gore i uputiti je na drugi popis e-adresa, poslanu oznaku ne možemo promijeniti. Poslanu oznaku ne možete usmjeriti na drugi popis e-adresa. Možete samo ukazati na više e-adresa.
const sent = [];// You cannot change the meaning of sent now// But you can add more values to it:sent.push(new Email());Pogreške i iznimke
Stručnost programera uglavnom se odnosi na to kako se nositi s pogreškama. Stručni programeri vole pogreške jer za njih pogreške znače napredak.
Ponekad očekujemo vidjeti ove prekrasne crvene poruke, a ako ne znamo, kôd je jednostavno pogrešan!
Volim frazu " slušajte svoj kod", jer mislim da se kod razvija tako što nam komunicira koristeći pogreške.
To je točno poput odgoja djece.
Najvažniji roditeljski koncept koji sam shvatila, prakticirajući, jest kako djeca komuniciraju lošim ponašanjem. To je zato što još nemaju logičan mozak. Mislim da programi rade potpuno istu stvar. Oni također komuniciraju neprimjereno (proizvodeći pogreške) jer programi nisu potpuno logični. Vaš zadatak kao programera je dodati više logike u kôd za rješavanje slučajeva koji su izvorno stvorili pogreške. To je upravo poput zadaće roditelja naučiti dijete koje se loše ponaša što nije u redu s tim lošim ponašanjem i što sljedeći put učiniti drugačije.
Neke se pogreške ne mogu popraviti i program koji naiđe na njih trebao bi samo izaći (i ponovo ga pokrenuti). Ovo je kao da vam srce stane. Ne može se puno učiniti osim ponovnog pokretanja električnim udarom. Zbog toga nadgledamo svoje programe i ponovno ih pokrećemo kad dođu u to stanje. Srećom, postupak ponovnog pokretanja programa nije tako dramatičan.
Većina pogrešaka koje se dogode tijekom ranog razvoja programa pomažu u poboljšanju tih programa tako da se pogreške nikad ne dogode. Ovako se odgaja dobra djeca. Ne ponavljaju loše ponašanje, jer sada imaju dobru logiku da ih vodi u dobrom smjeru.
Some errors evolve to be exceptions. Exceptions are expected errors. Errors that we can plan for and recover from. The best coding example here is a Network Connection error while we make a program, for example, download some data. This is very much expected because we know network connections could be unreliable so we plan for that error. When that error happens, let’s label the task of downloading that data as incomplete. Queue it somewhere, and re-try it at a later time (see below for an analogy for queuing).
What we did with this planned exception is give the computer a different set of instructions (a different recipe) to do when that error happens. We do exactly that with our kids as well. We give them instructions about what to do in certain future scenarios that we expect (or fear in this case).
// Hey kidsif (stranger.offersYou(chocolate)) { doNotAccept(); doNotTalkTo(stranger); walkAway();}if (stranger.triesToForceYouToDoSomething()) { screamFor(help); runAway(); call(911);}Reactive Programming and Streams
Reactive programming is a popular method for writing code that is based on reacting to changes. It is inspired by our everyday life and how we take actions and communicate with others. When performing everyday life activities, we try to multitask when we can but the brain cannot multitask no matter how hard we try. The only way we humans can multitask is to switch tasks and split them efficiently during their lifetime. This makes more sense when the tasks that we need to do require some amount of waiting, which is almost always the case. We actually always switch-tasks, even when we are not aware of it.
Reactive programming is simply to program using, and relying on, events instead of the order of lines in the code. Usually, this involves more than one event, and those events happen in a sequence over time. We call this sequence of events a “stream”.
Think of events as anything that might happen in the future. For example, you know that Jane (a store owner) is always tweeting interesting things on Twitter. Every time she tweets something we call that an “event”. If you look at Jane’s Twitter feed, you have a sequence of “events” happening over time (a stream of events). Reactive programming is named so because we get to “react” to those events. For example, imagine that you are waiting for Jane to tweet a promotional code about something cool she sells in her store. You want to “react” to that tweet and buy the cool thing using the promotional code. In a simplified picture, that is exactly what Reactive programming is all about.
To be able to react to an event, we have to be monitoring it. If we do not track the event, we will never know when to react to it. On Twitter, to monitor the events of Jane tweeting, we follow Jane and set our phone to notify us every time she tweets. When she does, we look at the tweet and make a decision on whether we need to further react to it or not.
In reactive programming, the process of monitoring an event is known as listening or subscribing to the event. This is, in fact, very similar to subscribing to a newsletter. When you subscribe to a newsletter on the Web, you supply your email address. Every time there is a new issue of the newsletter your email address will be used as the way for you to get a copy of the issue. Similarly, we subscribe to an event stream with a function. Every time there is a new event, the stream will use the function to enable our code to react to the event. In this analogy, the newsletter platform is the event stream. Every issue of the newsletter is an event and your email is the function you use to subscribe to the event stream.
Now imagine a dynamic newsletter that allows you to select topics and send you only the news items that match your topics. You are basically filtering the newsletter issues to your liking and that is something we can do on event streams as well. Also, imagine that you have subscribed to several newsletters using different email addresses. You later decided that you want all issues of the newsletters to be sent to a new single email address. One easy thing you can do is to set an email rule that forwards any issues from any newsletter to the new email address. You are basically merging multiple newsletter issues into one email address, which is another thing we can do with event streams.
Another way to think about event streams is to compare them to regular arrays. They are actually very similar. Arrays are a sequence of values in space while event streams are a sequence of values over time. In reactive programming, all the functional operations that we can do on an array. Filtering, reducing, mapping, combining, piping can all be done on event streams. We can filter an event stream, reduce the values of an event stream, map an event stream to another, combine streams, and make one stream an input to another. These are all options that yield new streams of values over time.
Callbacks and Promises
Imagine you ask someone to give you something that needs some time to be prepared. They take your order and your name and tell you to wait to be called when your order is ready. After a while, they call your name and give you what you asked for.
The name you originally gave them is the callback function here. They called it with the object that was requested.
This is like when you order a latte from Starbucks (in the store, not in the drive-thru). They synchronously record your order and name and then you wait until your name is called. When that happens, you receive your latte:
starbucks.makeMeALatte({ type: 'Vanilla', size: 'Grande' }, Samer);// "Samer" here is the callback function.// When the Latte is ready, the barista will call Samer // with the ready object// We define a function Samer to process the ready objectfunction Samer(readyLatte) { // drink readyLatte}Now imagine you ask someone to give you something, but they give you something else. Let’s call it a mystery object. They promise you that this mystery object might eventually turn into the thing you originally asked for.
This promise mystery object can turn into one of two possible forms. One form is associated with success and the other with failure.
This is like when we ask a chicken for a chick and the chicken gives us an egg. That egg might successfully turn into a chick or it might die and be useless.
const egg = chicken.makeChick(); // It's a promise!egg.then(chick => raiseChick()) // Success outcome .catch(badEgg => throwBadEgg()) // Fail outcomeQueues and Stacks
When we work with elements of data, there are two popular data structures to store and use these elements: A LIFO Stack and a FIFO queue.
LIFO stands for Last In First Out and FIFO stands for First In First Out.
The simplest analogy of a data stack is the stack of dirty dishes in your sink. When you are done using a dish, you stack it on top of the existing dirty dishes until you are ready to wash them.
When you are ready to wash them, you take the last dirty dish that you stacked and you wash that. In computer terminologies, we say you “popped” a dish.
The last dish you stacked is the first dish you washed. This is LIFO.
The simplest analogy of a data queue is the line of people that forms in front of a checkout or order station. When you are ready to pay for your groceries and take them home, you might need to queue yourself in a line until it is your turn.
The first person to arrive at that queue will be the first person to be done with it. This is FIFO.
Pair Programming
You can drive your car on your own when you go to familiar places, but when it is time to go somewhere far for the first time you use a GPS. If you have someone else in the car with you, a better option would be to have them navigate by giving you the instructions on where to turn next. If you do not follow the instructions and end up taking a bad turn, they will let you know immediately and advise you on how to correct it.
Having a navigator next to you when you drive is like having a pair-programmer. You are not driving alone. You are a team with the same goal: to arrive at your destination safely, without any problems, and with the least amount of time and effort.
You can probably do it yourself without a human navigator or a fancy GPS by using the old-school way and checking a map before you leave. If needed, you can check the map again. If you check the map while driving, you might accidentally hit a curb or put a dent in the car. If you stop to check the map, you will be losing time. Without that pair navigator, you are not as safe and/or the journey will take a lot longer.
The experience of your pair navigator might also teach you new things. They might know of a new shortcut that you do not and one that is not on the map. You learn from their relevant experience, and this is beyond valuable.
If you need to go to two destinations and you have two cars. You might be tempted to think that it would be faster to drive solo and do the destinations in parallel. This might be faster in the short term, but all things considered, time might not be the most important factor here. When it comes to computer programs, using one car and making sure it is dent-free at the end of both journeys might be a far more important factor. This why we love pair programming.
Linting and Task Automation
If you have to drive alone on that long trip, you can still make your journey safer by relying on tools. A map is a tool. The GPS is a better tool. Cruise control is another tool.
Tools that automatically warn you if you do something wrong while driving are similar to linting tools for coding. In JavaScript, the best linting tool today is ESLint. It will warn you about so many wrong things you should not be doing while coding. Best of all, it can do that even before you run your program.
Examples of tools that warn you while you are driving are evolving in modern cars. Cars can now warn you when you cross a lane line unexpectedly, or when you try to turn or change a lane while not seeing that hidden car in your blind spot. Additionally, they warn you when you drive over the speed limit, or when you are about to hit something while trying to park in a tight spot.
Linting tools also evolve to provide more accurate and helpful warnings. ESlint always surprises me with very accurate warnings. Additionally, its default recommendations are getting better with each upgrade.
Another analogy that I love in modern cars is automation. Any task that you repeat often should be automated once its purpose and value are clear. Instead of restarting that program every time you save the file, have a monitor process that automates that. Rather than running a format command on your code before you share it with others, have a command that automatically does that every time you commit your code to source control.
Modern cars automate so many things as well. The obvious example here is adaptive cruise control, but other subtle examples include automatic windshield wipers and automatic high beams at night (my favorite!).
Imperative vs Declarative Programming
When you need to do something, there is always the what and the how aspects of it. What exactly needs to be done and how do we do it.
Imperative programming is about the how. Declarative programming is about the what.
What? How? And why should you care?
An imperative approach represents a list of steps. Do this first, then do that, and after that do something else. For example: Go over a list of numbers one by one and for every one add its value to a running sum.
A declarative approach represents what we have and what we need. For example: We have a list of numbers and we need the sum of those numbers. The imperative language is closer to the computers of today because they only know how to execute instructions. The declarative language is closer to how we think and command. Get it done, please. Somehow!
The good news is computer languages have evolved. Computer languages offer declarative ways to do the needed imperative computer instructions. Just as cars have evolved from manual stick shift into automatic and self-driving ones!
Imperative programming is like driving a stick shift car. You need to do manual steps (press the clutch, depress it slowly, change gears incrementally, etc). Declarative programming is like driving an automatic car — you just specify the “what”: Park or Drive.
You cannot program declaratively unless you have the tools that enable you to do so. While you can imperatively drive an automatic car (by switching to manual mode) you cannot declaratively drive a stick shift car. If all you have is a stick shift car, imperative programming is your only obvious choice. This is unless you take the time to install an automatic gear shifter, which might be worth it in the long term. If you can afford a new car, you will probably go with an automatic one unless you are that true nerd who still likes to program with Assembly!
Assemblyis the original true imperative low-level computer language with pure instructions that directly translate into machine code.
Note that imperative programming might produce faster programs. Additionally, declarative programming requires less effort from you. In general, it will also require less effort to be maintained. Coding does not have to be one way or the other. Any non-trivial computer program will most likely have a little bit of both approaches. Also, knowing how to code declaratively is great, but it does not mean that you do not need to learn the imperative ways as well. You should simply be confident using both.
Tools that enable you to program declaratively evolve into better and faster ways to get you where you are heading. The ultimate declarative experience with modern cars is the self-driving ones. The “what” becomes the destination and the car will do the rest. This is somehow, probably, the future of programming as well. We will have programs that understand all goals and they can just work their magic to generate logic to get us to those goals.
What is your favorite analogy? Let me know in the responses section below.
Thanks for reading!