The 2 Biggest Mistakes I Made When Learning to Code | Hacker News Reader

The 2 Biggest Mistakes I Made When Learning to Code

suneelius.com

147 pointssuneel010186 comments

If you have been starting to write code and study code for a year, you are still learning to code. You have not learned to code, you are not even a journeyman yet.

I've been in the industry about 9 years, and coding since I was a child. My parents caught me staying up late with a flashlight under the covers as a teenager reading books on C instead of nudie mags. My first program with an event-driven UI ran on a NeXT Color Station. This is not to brag, it's to say I've been doing this basically since I could decide what to do, and I still have a lot to learn about coding, computer science and software engineering.

So you've made your third biggest mistake; you've vastly underestimated what you are in for. You have never "learned" to program, you are "learning." Every year will ask you to acquire new skills and use new models and learn new domains.

> So you've made your third biggest mistake; you've vastly underestimated what you are in for. You have never "learned" to program, you are "learning."

I am afraid your comment is even more mistaken. Even more it is discouraging in a really bad way. Read carefully and I'll explain. Even if your HN score is 10 times mine.

In my opinion you are creating a false dichotomy between having learned and learning. Someone who has just started to walk/bike etc has still learned to walk/bike. They are by no means near entering Olympics for the next few years but they have made their own life very much simpler and more enjoyable. (Father of small children here.)

Comments like yours are discouraging at least one specific subset of people from doing the one thing that can possibly help them increase their skills, namely use them. I even think that for most people even if they can learn a lot about coding by just reading, reading and immideately applying it is by far a quicker route.

Anecdata: I once told a 16 y.o. intern how to use basic perl and Visual Basic for Applications. After 14 days where he would run his scripts, check, check with his supervisor, fix the code, learn more perl and vba and so on he came back with a report that would have taken weeks to finish by hand. He is now on his last year on a Bachelor with Honors study and this was his first useful program. Still makes me happy.

(By the way, I also started coding at age 12. Having programmed for both $100+ million companies as well as startups I am not one of those struggeling to learn the basics. Anymore, that is : )

> In my opinion you are creating a false dichotomy between having learned and learning. Someone who has just started to walk/bike etc has still learned to walk/bike. They are by no means near entering Olympics for the next few years but they have made their own life very much simpler and more enjoyable. (Father of small children here.)

The fundamental difference here is that bikes are ridden in a way that's fundamentally unchanged for decades. Contrary to this "settled" field of knowledge, programming is constantly invalidating itself. Even if you achieve competency in a limited field (perhaps operating system design or web development), that field will almost certainly overturn itself within 5 years for any reasonably broad definition of field. Consider what modern linux looks like compared to the original version; a lot of new techniques have emerged to address requirements.

To enter that field, this new knowledge is not "optional." It's required. The goalposts for competency aren't just shifting; they've got a nuclear power supply and tank treads and they're out of control.

In this, it is not unlike being a doctor. Their knowledge base is similarly in constant flux (although not quite so violently as ours). A doctor who does not constantly improve and update their knowledge will be a substantially worse doctor than one who does. A perfect example of this is pre-scientific doctors like homeopaths, who basically do nothing with an almost frightening level of dedication and fervor.

> Comments like yours are discouraging at least one specific subset of people from doing the one thing that can possibly help them increase their skills

I did not direct this at some 16 year old kid or a non-programmer, nor did I put it in a venue where non-tech people read. Even if I did, citing final consequence is hardly a good argument. In any case, the industry does a pretty good job of discouraging people from joining as it is. When I was in college, there was a 60% dropout rate between lower and upper division classes.

Checked your profile and you seem to be smart, yet I think you are wrong again : ) (sure you have got enough sleep lately? not been talking to much to marketers?)

> "The fundamental difference here is that bikes are ridden in a way that's fundamentally unchanged for decades. Contrary to this "settled" field of knowledge, programming is constantly invalidating itself."

The most well known styles of programming (imperative, functional, object oriented etc) have been unchanged for two or more decades as well.

And a 10 year old c or html tutorial still works and still lets people start solving real world problems.

I guess I'm leaving this discussion here. You know way to many cool expressions for me: '

>"programming is constantly invalidating itself" - this holds true for Java, C and Lisp right? Or are we talking about how for loops aren't valid anymore?

> "Consider what modern linux looks like compared to the original version;" - We are talking about that "unix clone" thing from early ninetys? Has it become a Windows clone without me noticing?

> "A doctor who does not constantly improve and update their knowledge will be a substantially worse doctor than one who does. A perfect example of this is pre-scientific doctors like homeopaths, who basically do nothing with an almost frightening level of dedication and fervor." - Comparing software developers to doctors. Go tell some doctors. Try car mecanic. They also have to stay up to date.

> "I did not direct this at some 16 year old kid or a non-programmer, nor did I put it in a venue where non-tech people read."

No, but you directed it at a new programmer who shared something that a lot of people in that situation seems to struggle with. Hey, not just new programmers, even Kent Beck has blogged about a similar situation: http://www.threeriversinstitute.org/JustShip.html

> Checked your profile and you seem to be smart, yet I think you are wrong again : ) (sure you have got enough sleep lately? not been talking to much to marketers?)

I am not sure why you decided to include this toxic, thorny little jab in your post. Please know that it's not appreciated.

> The most well known styles of programming (imperative, functional, object oriented etc) have been unchanged for two or more decades as well.

This is not true. The state of the art in all of these has advanced considerably, and many things that were once considered good practice have been discarded over time.

> And a 10 year old c or html tutorial still works and still lets people start solving real world problems.

I'm not sure what definition of "real world problem" you're talking about. Certainly nothing I learned 10 years ago directly pertains to my work today... Not pure math; its quite rare to use C for that sort of problem anymore. Not networking problems, 10 years ago you couldn't write servers the way you can today (nor should you save for perhaps embedded systems like microcontrollers?) Not UI either. Perhaps simple text munging? Doing that in C is pure futility compared to what we have today.

As for HTML, a 10 year old HTML tutorial would give you almost no insight into modern webpages, and a very large amount of it would be "N-hancements".

> No, but you directed it at a new programmer who shared something that a lot of people in that situation seems to struggle with.

I'm not sure what your point is. Nor am I a huge fan of Beck's philosophy in this.

Learning to walk/bike is in no way comparable to learning to code. Learning to walk/bike is primarily a physical muscular action to train your body to keep it's balance etc. Once you have learned to walk or bike, you have learned it. Learning to code is an activity of the brain where there is no clear defining line as to when you have learned. You might feel that you have learned a few languages or frameworks but there is still more or still something else that you can learn. So, it's amore gradual and continuous process. Hope this makes sense.

> Once you have learned to walk or bike, you have learned it.

Go tell that to someone who competes in running or biking.

> You might feel that you have learned a few languages or frameworks but there is still more or still something else that you can learn.

Again, same goes for biking: There are all kinds of tricks you can do. Yet once people manage to ride a bicycle for a few hundred meters without falling they have picked up the concept and we say they can. Even if there are still lots to learn (traffic rules, all kinds of neat tricks etc)

I argue that the same goes for coding: Once people understand the basic building blocks and start writing their first useful or amusing program from their own imagination we say they can program.

At this point they can start making their own and possibly also other peoples lives more comfortable. Start making their own prototypes which is what I read from the original post.

Would you hire them as programmers? I would not. Do they need to read more books or attend more courses to start prototyping their own ideas like the original post was talking about? Thankfully not.

Absolutely right, there's no end to it. What I hope the reader will achieve is the basic ability to prototype their ideas. That's where I am. In no way do I think I've learned all there is to code. Every day I'm learning precisely how little I know. Heck, I just started learning Lisp and truly enjoying how unfamiliar and powerful functional programming can be.

"Heck, I just started learning Lisp and truly enjoying how unfamiliar and powerful functional programming can be."

Try Haskell before you have too many preconceptions of what "should" be to burn away. :)

Sounds good :)

A recovering non-coder?

I think the problem is a lot of people jump in, start coding and then never go back and learn about the nuances of a language.

Case in point was myself when I was learning Javascript. Sure, I could get the JS to do what I wanted, but it was clunky, used a lot of memory and was slow. Fast forward a few years and I've gone back and read Crockford's books several times, and with more studying, I've gotten much better at writing JS. I now try and make my JS as lean and as fast as possible. It's a totally different approach then when I started.

The problem is thinking once you know how to do something, you're done. Like a friend once told me, "You don't learn to be a programmer, you learn to be a student of the language you choose."

This goes hand-in-hand with people writing lots of code (especially when first starting out) and not reading much.

There seems to be a huge lack of people simply reading code to learn from it, rather than poring through dozens of recommended "Learn to Program in X in 7 Days!". A novelist probably reads more than they write; the same is rarely the case for programers (at least in my experience, especially in the case of people who are less pedantic than others).

You make some valid points.

However, had you not taken the road you did, the material you read much later on probably wouldn't have made much sense and probably could have made you turn away from it all right at the start, right?

You get excited when you see results right away. Once you do enough things that begin to make you wonder, "Gee, there's gotta be a way to do this faster or more efficiently." When your curiosity drives you to search for answers is when you delve much deeper into the nuances and the grit of a language that becomes important for scale.

You're right, there's always some retrospective bias, but I'll never get back those days of reading about PHP, since reading Python accomplished the same and actually was much nicer for a beginner. The truth is, your MVP will evolve and will require you to learn things, new languages and new technologies. The fun is never over. For me it's continuing and im so glad of that.

Well put. I've been programming professionally for 8 years (and as a student for another 8 years prior--HP48G RPN represent!). And to this day, I find myself regularly attracted to articles, books, and blog posts about "basic" lower-division CS topics: principles of object-oriented design, language design, patterns, compilers, etc. It's the same old stuff, but reading it is more enriching and insightful the more real-world programming experience you have.

yep. When I just started I struggled through GoF's Design Patterns and even sorta tricked myself into believing I understood it. Well, I didn't. At all. 5 years later though it was one of the most insightful and useful reads ever.

I agree that learning the nuances of languages is important, but to demand this aspect of learning to someone who is just beginning to code is actually harmful in my opinion.

As students of computer science, we have strong tendency to learn things bottom-up. We take care in learning the fundamentals before applying theories in real work. This is not a bad thing. Better quality codes are shipped and become more maintanable.

But for someone who is just starting out, the bottom-up approach becomes an insurmountable obstacle. As a result, many give up too soon and never end up enjoying the joy of creating real working applications.

Look back to when we first started to code. I wrote some bad BASIC code in high school abusing GOTO and GOSUB everywhere I could. I knew nothing about programming language paradigms or memeory management, but I had a lot of fun coding without reading any book on CS.

I don't know that #1 was a mistake. When getting into any new field, it is very hard to tell what is important and what isn't, what is a solution to the problem and what isn't. If you are in school, sure they'll guide you, but if you are self-learning, there is no really good filter. So learn about everything. Take a breadth-first approach. Some of the specifics you learn will be wrong, useless, or otherwise inappropriate. However, getting an understanding as to why they aren't relevant is in fact learning the field.

The more information about a topic you have in your head, the more likely you are to see the connections between subfields, how various bits are important and unique or just ho-hum standard implementations.

I regularly see people in some niche area say "look we have found this amazing new way to handle the data-flood", only to tell them they have reinvented NoSQL or a message queue, or a technique for dealing with matrices that the image processing or simulation folks have been doing for years (and all of them poorly, making recognizable mistakes). A breadth-first overview has much advantage.

Agreed. It's easy to look back and know what were needed in retrospect but it's hard to even know which is for what until some time is spent to learn its capabilities

Step 1 in learning to code:

10 PRINT "COCKS"

20 GOTO 10

Step 2 in learning to code:

10 INPUT "How many cocks do you want",C

20 FOR I = 1 TO C

30 PRINT "8===========D"

40 NEXT I

I'm not recommending that your first steps into programming be in BASIC, or involve cocks. (On both accounts I would recommend the exact inverse.) But this is how a million programmers got started: useless programs that amused us. Never, ever, ever underestimate a useless program, especially for beginners. The author of this article seemed to want to focus on building a prototype for his Web startup. And that's fine, as a long-term goal. But if you're brand new to the world of programming it's probably a good idea to start with the basics, like printing or drawing stuff on the screen, doing some simple (or, later mor complicated) mathematics, or munging or transforming data in interesting ways. This helps keep the budding programmer focused on the patterns and "shape" of their language, which knowledge they can then later direct to a specific purpose like Web or smartphone apps.

I totally disagree. I can't stand tutorials or classes that present useless stuff as exercises. It's so much easier to work on something that you can at least imagine would have some value. And it is much, much easier to work on something that you think is either cool or have always wanted to build, etc.

Also, 1/2 a downgrade for the lame, childish example.

coding lame, childish "examples" are great to get people interested in learning to code.

they provide immediate gratification in terms that non-developers find humorous. which is a great way to spark interest.

Your standard for the importance of things being useful might be slightly out of perspective when you consider the main goal here is to learn: when you first start programming it doesn’t matter what the program does just as long as it did something, you understood it, and you can build on that knowledge to the next thing. That’s why we start with Hello World.

Still I think his point is relevant. For a lot of people the fact that something is useful contributes a whole lot to the motivation for learning it.

He said "...useless programs that amused us" I think that everyone missed the "amused us" part. If your program is useless but still does something that you think is amusing, then that is a great feeling

If I'm learning a new language, the very first thing I want to see is Hello World. It lets me see the basic syntax. Introducing a simple variable and if statement comes next. You pretty much have to start here before you can start creating.

Congrats on learning to code but the biggest mistakes new hackers are making is diving right into learning a specific lang.

Learning programming fundamentals is a better way to go. That way, you can jump from lang to lang.

Starting is hard so congrats again.

I disagree. You didn't learn to talk by studying rhetoric. You imitated sounds and words that people made around you. Then you started being able to express your own ideas. Then, maybe, you learned how to say things well.

It's the same with computers. While some people may enjoy learning deductively about abstractions like algorithms and paradigms, studies have shown that most people learn inductively, by having something they want to do and figuring out what to do.

Human speech is a bad analogy. A better analogy would be grammar and spelling, which would make OP's argument stronger. The very argument you disagree with.

Truth is, you first learn the basics of speech communication. You understand sounds come from your mouth and given a specific cadence and tone you can form words to get what you want. This is no different than understanding the basic principals of computer programming. Principals all computer languages use. Specifics of the language (syntax, grammar, spelling) is something that can be mastered later, but without the principals, you're not going to master anything. This is a major reason why there are so many bad PHP examples out there. People don't learn to program, they learn to write PHP code poorly.

I find both of these analogies inadequate.

Spend more time around young children. They most definitely do not learn grammar before they can communicate. It is not a linear thing at all. People pick up parts of grammar like the Subject Verb Object syntax as they learn to mimmic diction and build vocabulary. They're constantly "wrong" and need to be corrected. With persistence and practice and, most importantly, interest, we acquire taste and style.

Communication is very much a trial and error kind of learning that is holistic rather than linear. "Learn x, then do y" versus "do y, then learn x" is a false dichotomy.

I disagree (at least that this is universal). If you start with just fundamentals but no specific language, you can't whip up anything to play with. The amount I was able to learn easily without doing anything was quite low, so I'd say start with a language with a responsive interactive shell so that you can quickly see what things do for yourself, and (crucially) see what happens when you make small changes to whatever examples you might be looking at.

But there's a bit of this I agree with: don't get focused on just one language. Once you can get basic stuff done in one language, it's worth learning how to do stuff in a different language. For me, the second language was almost as hard as the first, but after that it got a lot easier.

(Which is not to say that I don't think fundamentals are important: I'd suggest eventually hitting the core algorithms, data structures, computer architecture stuff—even if it doesn't seem immediately relevant—as well as playing around with more out-there stuff (Prolog was always a favorite of mine). But I don't think all that's going to be interesting at first to someone who's just getting started out of a practical, "I want to build something", not a "I want to learn all of this!", motivation.)

Agree.

Learning the fundamentals doesn't mean not doing actual programming at the same time. You have to code to have a better understanding on the fundamentals.

My personal experience is that after many years of programming, the fundamentals I learned in school still help me get better. There were many such moments that you were digging into some interesting technical topic and you suddenly realized that some fundamentals had another layer of meaning or application you didn't know or I didn't fully understand before.

Also, understanding different types of programming languages, like OOP, FP, logic programming, declarative programming, etc. help improving one's programming skills, no matter what programming you primarily use at work.

Agreed. I am largely self taught as well, and have found a lot of value in reading texts on CS / programming fundamentals :

  * Computer Architecture
  * Operating Systems
  * Algorithms
  * Data Structures
  * Database Fundamentals
  * etc.

In addition, more seminal-yet-not-directly-practical books such as SICP are quite valuable. I've been meaning to read TECS (a copy of which is staring at me as we speak), however I have not had the time to devote to its projects yet.

Edited to add :

In addition, reading up on web application architecture was quite helpful. This is increasingly important as modern frameworks may otherwise abstract away concepts to the point of making them very black-boxy.

That said, my goal was not to rush towards an MVP or prototype. I intend to be a competent developer first, and a founder second. If these were reversed, I might well have adopted the OP's approach instead.

Are these things that were helpful or got you started?

I don't think anyone is challenging the value of studying those topics. It just seems to be general consensus you don't start with them.

I suppose it depends on how started is defined, but yes - having an understanding of the topics I mentioned preceded working as a junior developer.

What was the order of learning for loop, function and class? Did those proceed architecture, algorithms, data structures, etc...?

Personally, the challenge of learning the more base knowledge and gaining a deeper understanding of computer science is that until you are really good at making things, you honestly cannot tell the difference.

We can certainly agree to disagree, but I don't think that being "really good at making things" is a pre-requisite to understanding some of the principles underlying said "things". In many ways they are related-yet-distinct disciplines (Computer Science vs. Software Engineering).

I readily admit that I may suffer from some bias in this regard, since I find more abstract topics to be a lot of fun. There are many roads to Rome, however this is one that I personally found to be useful.

I have some friends/colleagues like that, and they are a certainly a minority. All very deep thinkers and great to work with. I could see how learning all of the abstract material would help with that learning type, so I don't think we disagree. Everyone should use whatever approach works best for themselves. This is probably a better point than anything else.

I was focused on my own experience and the experience I've had with the majority. Generalization and such.

I agree to an extent, but this shouldn't be construed to mean what I think is a bigger mistake that new programmers[1] make: bounce around from language to language when you're still trying to learn how to program. I made the same mistake when I was learning. The language is (largely) irrelevant, so by trying to learn Ruby when you've barely learned C (vice-versa, whatever), all you're doing is overloading yourself with syntax, when you should be focusing on nailing down the syntax of some language, (essentially) any language, so that you can begin to focus on learning how to program and not what the asterisk means in that context.

[1] Tagent: am I the only one who hates "code" as a verb and "coder" as a noun? It makes it sound so rote, and reminds me of "code monkey." I'll take programmer/developer/hacker/engineer.

I agree and disagree...

Learning the fundamentals is important, but one must put the theories in to practice by learning a language, writing some code, making mistakes, correcting them, and learning how the fundamentals work in practice.

thats why cs courses has projects, to test the theories in to practice, the good thing about fundamentals is that it teaches you that the language its not an end for it self instead its just a tool to express yourself

True. I think there may be two camps though. It depends on if you want to learn and be able to master the fundamental theories that ground everything or have a more pragmatic approach where you accomplish things using tools you doesn't fully understand.

My example would be: when was the last time you implemented your own merge sort?

True that my knowledge of algorithms benefits me, but maybe all I need to know is that it sorts an unsorted array.

In the end, the best approach most likely depends on the person. Some people need to understand things fully, at the lowest level before they can build on them. Others need to see something in use and be able to play with it to wrap their minds around it. I know I'm personally a combination of the two.

Regardless of the type of person you are though, I think everyone is advocating that you do something to get the ball rolling.

Speaking only for myself, I "learned to code" six years before I got to college to study CS. CS taught me a bunch about what I was doing right and what I was doing wrong. My first paid programming job, as a senior/first-year master student taught me what all that meant when applied to what a business operation needs from its coders.

Learning to code, means being able to write a o(n^2) sort, not necessarily being able to tell why that algorithm is bad, or what O(n^2) means.

We all have different ways of learning. When it comes to programming, learning the fundamentals first can save a lot of time and frustration. Not everyone can pick a language and start learning right away.

I think a better way to start is by learning the fundamentals like: - variable declarations - if statements - loops - functions

Those fundamentals are the same whether you're doing backend coding or frontend javascript.

Remember, it's all syntax.

That might or might not be ideal theoretically - but it requires an enormous amount of self discipline and persistence if your're on your own.

But I think all self-taught programmers (of which I am one) should set aside some time for the fundamentals once productive in one language.

I would talk a more organic approach. Jump right in and when you come upon some bigger subject matter (sorting algorithms for instance), take a break and go study the subject a bit..

I don't know why you were downvoted but there is definitely something to be said for what you argue.

Best quote in the article:

    There are so many benefits to jumping right in. 
    You’ll quickly get over any fear you may have of programming. 
    You’ll start seeing the fruits of your labor right away.

This did it for me. I've been having this exact same thought for the past two weeks. After hitting development bottlenecks on my last project, I committed to learning to code to the point where I could hack together functional prototypes. I've been training by manual, and was just speaking with a friend about how dry it is and the fact that I'd be more fulfilled working on a project and couldn't wait for that.

We have an extremely small, simple project that we've been talking about. Screw it. I'm dumping the book and starting on that as soon as I get home.

Exactly what I needed to read.

Awesome! I have been trying the same sort of thing and documenting it here: http://mksht.crisnoble.com

I was inspired by: http://mroosendaal.prosite.com/628/612/work/msced and http://designitcodeit.com/

Plus Mig Reyes's quote "Just make a bunch of shit" or the more eloquent way that Ira Glass put it "the most important thing you can do is do a lot of work. Put yourself on a deadline so that every week you will finish one story. It is only by going through a volume of work that you will close that gap, and your work will be as good as your ambitions."

It is funny. I was talking to a kid that was learning to play guitar because he wanted to play cover songs, and was frustrated by the lessons that seemed to have nothing to do with what he wanted. I told him that he didn't have to wait even one more day, since playing covers songs is one of the best ways in which you learn guitar. His ultimate goal could be accomplished almost immediately by choosing one song he liked that only had 3 chords, and just strumming those three cords over and over again in succession until it didn't sound half bad. Would take less than a week, -tops.

It's the same principle, and one I value highly. I think tackling a new discipline, and the wave of information that rushes toward you as you put your toe in the water can sometimes distract us from simple grounded approaches that we know have worked well for us in the past.

Thank you for the links!

I learned piano in a similar way. It was a lot of fun, and allows me to play some songs passably, but I still feel hobbled by not knowing certain basic techniques that I would have learned sooner in a more conventional set of lessons. Maybe it comes after the "project" type learning, but at some point, to move up, you have to learn the boring basics.

Agreed! I'm also a pianist, but because I started young, I was lucky that my parents and teacher forced to me practice technique all the time. You should check out the Liszt exercises, they will make you strong and fast: http://books.google.com/books/about/Liszt_Technical_Exercise...

With coding, there was no such outside pressure, so until I joined Yipit, I had tons of bad, hacky habits. In order to get to the next level, I had to rebuild from the ground up and learn the fundamentals really well.

I agree with this. I think the hardest part of learning a new skill is starting out. Working on a concrete problem provides a positive feedback loop that keeps you engaged, learning, and moving foreword. But that doesn't mean you should skip fundamentals.

For coding you could hammer out a project, get familiar with a language and the technologies you used to implement it, then download a Stanford lecture on proper Programming Methodology.

This is such a nice comment and I'm glad my post could help! Good luck hacking on that project. Let us know how it goes. I'm sure you will learn a ton :)

I disagree slightly, as I don't think that following that path of reading instead of jumping in is inherently worse.

Just jumping into writing code is great for motivational reasons because you'll have something tangible, something to be proud about and to help you get excited and keep learning. But you'll still lack insight in a lot of the surrounding environment, things that any programmer should know.

Attempting to do this the other way, reading as much as possible and as in depth as possible, will be good to give you some insight, but you'll have nothing to show for it until you start coding and it'll still take time for you to learn some stuff that can only really be understood with practice.

What I'm trying to get at is that it doesn't really matter how you start, as you are going to have to dedicate 10,000 hours to learn, polish and excel at any craft.

Well, you need to know what is out there, and what is possible. So at least skim over some of the technologies. You could waste a lot of time if you didn't know say, Sass, existed.

There's also a timing problem. If you invest now, and learn technology T, then when you are in the middle of developing a project, the marginal cost of using T is very low. But if you are up to your neck in a project, and then need to take a week off to learn T, you are not likely to do so.

Personally, when re-starting my tech career I wish I had learned Java (or any JVM language) over PHP, PostgreSQL over MySQL, and Dojo over JQuery.

Great insight. Every time someone says, "I want to learn to code." They should have a project/task in mind and just go. The stack that Suneel lays out is a good foundation for web dev, but as he admits, he like many others invested more time than was necessary in books/documentation/tutorials.

More people need to just jump in. If you get to a point that you want a feature from jQuery, look it up and implement it, but don't feel like you need to wrap your mind around every feature before you can begin. Great work and keep at it.

>Every time someone says, "I want to learn to code." They should have a project/task in mind and just go.

If there was a list of software projects somewhere, ordered by difficulty, with tips on how to get started and how to extend it later, I would be very interested.

I can code (somewhat), but I find that any time I have an itch to scratch, it's much easier to google it and solve it through existing software and configuration changes. Every time I've had a project in mind, I'm quickly in over my head. I'd like to move beyond scripting and actually build self contained software I can post on Github and give to my friends, but I guess I don't know where is a reasonable place to start.

What do you wish existed and build that?

Oh, it's too big?

Build a piece of it.

Oh, that's too big?

Build a piece of it.

The problem you're probably having is that you're saying, "I want THIS" and THIS is too big for you to imagine. So, just write a small subset, maybe 3% of that. Then you'll realize you can easily build this next little bit over here or over here and eventually get there.

http://mindprod.com/project/projects.html

That's a nice list to help people when brainstorming. I wasn't necessarily implying that the project had to be something new or different, or even specific to the user. A lot of people learn rails or django by making some microblogging twitter clone. On a recent relevant topic, sometimes the best ideas come from emulation and slight modification, so maybe you just challenge yourself to replicate a site or web app you use.

That point reminds me of Kathy Sierra's distinction between just-in-time learning and just-in-case learning: http://headrush.typepad.com/creating_passionate_users/2005/0...

I've had a completely different experience. The first few times I tried to write code, I learned only what I needed to know for the specific app. I hated it. It seemed like I was just learning a bunch of random syntax and beating them into shape.

Later, I actually tried the textbook first approach, and this worked much better for me. By learning the framework and being able to observe the patterns, learning new commands was much easier and I could often guess at exactly what I'd need to do. I also remembered commands for much longer than I did with the first approach.

YMMV, but some people have more theory-centric learning styles and some have more application-centric. Test and figure out what works for you.

I wonder if the advent of web programming makes this process all the more confusing?

When I took basic Java and VB programming courses in college 10 years ago, we focused only on Java and VB. There was no HTML, CSS, JS, frameworks, etc. And you could have a complete, functioning program with GUI from the start.

Now, to create a web app, you have to learn all those other things in addition to Ruby, or Python, or PHP. As you advance in skills you'll likely move to either front-end or back-end, but when you start out creating everything from scratch, it feels like there's so much more.

(This coming from someone who did IT work for the 10 years following college, and is not trying to break into the web developing world).

Ask a developer who is working in the area that you are interested in to give you some guidance. My girlfriend decided to become a web designer. Do she asked an experienced developer, me. I said HTML, CSS, JavaScript and JQuery. She added Photoshop and some image creation tools. She learned the parts I suggested and decided to go to a community college to get some focus. (Actually, she originally signed up to get the student discounts on the software.) In a few months she will get a job to get some practical experience. Then she will tackle her own projects.

This pretty much applies for any endeavor ; not just learning to code. Thanks for the post.

2 Biggest mistakes when learning to [TASK X]

Mistake #1: I spent too much time learning things that I didn’t actually need.

Mistake #2: I didn’t start [TASK X] right away.

YMMV But I typically do #1 when learning / starting anything I don't know about. That is I some proportional amount of time googling/wiki reading/pinging friends for all the info/standards/tools/libs around subject. Review* them for worth, spend more time with ones that pass.

It is very helpful to have a idea of the breadth of and resource for a subject.

Review typically means "experimental coding". The 2nd point "start coding now" is a Truism. Just don't start coding "production". Learn some, think, first.

As much it pains me to admit it, Nikes' "Just Do It" applies here (and in life) as the author has learned. The procrastination and anxiety that comes with tackling something new can be overwhelming. The task oriented approach narrows focus and makes you think about what you need to learn. Fear is also a great motivator especially when your lead is breathing down your neck with a 'JGID - just get it done' attitude. :)

suneel, many people make this mistake. I went almost EXACTLY the same route you did. Although, I had already known HTML, CSS, and some Javascript, I was wildly unprepared for what I would encounter next just to learn how to get a working prototype. It basically took me 8 months before I got something running on Heroku. I went through four different Django books with some success here and there, including the Django Project's simple Poll tutorial three times without grasping it. Little to my understanding, all of it made sense once I used the book written by one of the creators of Django himself. I also used a lot of StackOverflow and the guys hanging out in #Python and #Django Freenode IRC channels were awesome!!!

The trial and error got me to learn a lot of things. In addition to the six things that you've listed, I forced myself to learn VIM as well. You probably use VirtualEnv, which you've forgotten to list.

Anyway, it's always good to see others going down the same road and sharing their experiences. Do you have some links to some of your early projects?

Totally agree. Prototyping is all you need to test out a startup idea. And if you learn by doing, everything you learn is useful!

It's easy to worry about every details and learn every piece of tech yourself all at once, but it's a recipe for unstable foundations and disasters of distraction.

Creating small achievements by prototyping simple ideas in unfamiliar technologies gives me small, concrete rewards to celebrate the journey.

Knuth hath said: "A person who is more than casually interested in computers should be well schooled in machine language, since it is a fundamental part of a computer."

I agree with that sentiment. The whole "focus on what works" is not the same thing as "understanding what the hell you're doing."

I think you can reduce the list a lot more--and should. I'd choose a single language and learn it well and go from there. And I'd only work the command line features. So Ruby on the command line or JS or Java would be my choices (but only one).

This is actually just 1 mistake.

Start writing code, make it do what you need it to do today. This will focus your efforts to what you need to do at the present moment and you won't end up trying to figure out how to deploy an app you haven't built yet.

They say you learn best through experience.

I learned CakePHP by coding my thesis with it. Now, I wanted to learn python + postgres by planning to port my thesis to it.

Why would you want to switch to Postgres out of curiosity? Also, what MVC framework are you switching to?

I used MySQL and I'm trying Postgres out of curiosity too. Django is the MVC framework that I will use. :)

Everyone out there who wants to learn to code, stop reading these comments and use the time to just go learn!

Yeah, away with all that crap, just use Django already! Good decision.

I'd be wary of recommending Django to a beginner, unless their project definitely required it.

Flask is lightweight, simple, and you can understand most of it in a very short period of time. I'd say it taught you to code more, whereas Django teaches you to use Django.

Better yet, use Express and focus on mastering one single language for the client-side and the server-side: either JavaScript or CoffeeScript.

keep blogging suneel....maybe this will finally get me to start learning!

To Generalize (for web):

HTML/CSS/JS (View)

jQuery (View Framework, couldhave been YUI, for example)

Programming language

Framework in said language

I still think actual SQL / DB knowledge is requisite to do all but the most minor of apps, but he is specifically trying to get people from being overwhelmed.

I'm struggling with Play!, because I need to know HTML/CSS/Java/Scala subset (templates)/EBean ORM/the massive library that is Play.

Just a simple app touches on so many libraries and helper functions, it's hard to get a grasp of what's going on. I don't even build a form anymore, I build a form-ish object that I pass to the controller that fills in data to the ORM that inserts stuff into the database. o_O

You put the cart before the horse (which was actually your first mistake). When you get out of college - get a job, work a while - then start a side business. Your startup success odds will more than quadruple following this advice.

Reply on news.ycombinator.com