The wrong question: “I want to learn to code, what should I do?”

If you ask someone you know for advice, you aren't going to get the same thing you get if you ask Google, and you aren't even necessarily looking for the same thing. I was in exactly this situation earlier in the summer, when my brother asked me, after taking MIT's intro to programming class using Python, "What language should I learn next?". When he asked me that, he wasn't looking for the best possible answer from the best engineers who are posting to Stack Overflow, he was asking his brother, who might be able to help him when he stumbles, since I would likely recommend languages I know. I also know that he's not a CS student, he's studying urban planning, so I could reply with that understanding, and say that he might not need to learn another language at all, but instead could branch out with Python to hack on stuff he didn't get to work on in his CS and GIS classes, like web programming or games.

So no, don't just point them at Google. Tell them YOUR story about how you learned to program, and tell them what YOU would do differently. Tell them to learn a language that you know, so you can help them. Or a language you want to learn, so you can learn with them. I know that teaching yourself via Google can be extremely gratifying, and I think it works great for some people, myself included. However it doesn't work for everyone, and sometimes it's nice to have a personal connection like this: "I learned to program when X helped me to do Y."

edit

as I was writing this I was itching to reach for something I'd read before, and I remembered now that it was probably Norvig's "Teach Yourself Programming in Ten Years". There he suggests a way to choose a language: "Use your friends."

Suppose the person hasn't already asked this question to google:

Take the opportunity to provide them some good resources. Teach them some basics. Teach them how to use google well to find their own resources. Here's the kicker - when people ask questions like this, maybe they aren't autodidacts, maybe they just need some support getting over the fear of the unknown, maybe they just want feedback from someone who actually knows (something JFGI doesn't provide too well...)

Now suppose they have:

The course of action is surprisingly similar, but ask them what they've already learned and what the confusion is before showing them stuff. Perhaps the person asking has done this but doesn't know where to start -- there is a lot of conflicting and bad advice on the web. Perhaps they don't have the framework to filter things that are and aren't related. Maybe they don't even know what they actually want to do -- there have been plenty of times I've asked someone how to $X only to discover that I really wanted to do $Y that I didn't even know about (even though Google will happily provide 100000 how to $X guides).

Half of understanding anything is understanding the vocabulary surrounding it. If a person doesn't know the vocabulary, they can't just google it. (Also, those of us who spend all day programming and basically live in front of the computer at places like HN have a surprisingly deep understanding of who to trust and who not to trust and apply these filters without even realizing it when scouring google results, as well as a good social network of chatrooms and twitter etc to ask for pointers at, which a noob may well be trying to establish when asking the question).

Get started and build something. Stumbling and not knowing already but figuring it out are part of the process and are valuable skills.

Also the whole buy a mac thing. It never ceases to amaze me how many people completely fail to realize how much privilege it takes to have the money to purchase a mac. Or the time it can take to save the money. When someone can get started on things that are a fraction of the cost.

You can get a cheap office depot or walmart laptop and an entire year of linode or whatever vps you like for the cost of a used mac

The idea of cultivating a self starter, motivated to research mindset is good, for people halfway through the beginner stage. The idea that a complete beginner even knows how to "bounce around between [different search results] to find the best ones" is silly, even starting with generally high quality content like quora and stack overflow. It leads people to learn the easiest things rather than walking the line between motivation through fast results and pushing themselves to learn good techniques and challenging subjects. There is a reason that good educators, who design curricula walking this line, are respected.

This type of "ignore the so called 'experts' and absorb the lowest common denominator crap soup of online programming advice" learning plan is doomed. The best case is that it wastes the persons time and saps their motivation. The probable case is that they end up learning a large number of bad habits, do things like reimplement a bubble sort based on advice from the comments on the php documentation site and forever end up as hack intermediate coders.

People asking for this kind of advice want some guidelines from people who've done this before and have perspective on good paths to learning and how to avoid mistakes that no one needs to repeat.

I am a proponent of branching out early in your programming career to learn different languages, toolsets and problem domains because it makes you a better programmer. But only once you've got a base level of knowledge as to how coding works and even then they will need advice on what other paths to take to help the most.

This article promotes introversion more than its own point of "don't ask, do", asking Mr. Computer isn't much different than asking a person. If I ask someone how to get started doing X; I'm not fucking asking permission. I'm looking for some resources, a knowledgeable person is a much better at steering a beginner than a search engine.

For some reason, you're equating drive with choice of resource. You're answering "RTFM" when the question is "Wheres the manual?" I get that your point centers on leaning on others vs figuring it out on your own, but you thoroughly overshot.

Bad answers can be ad-ridden, no-content crap, bikeshedding pundits, obsolete documentation, overkill solutions, corporate labyrinth websites selling overkill solutions, incomplete fragmented mailing list archives, or material way too dense and advanced to be consumed without guidance. But they lead to frustration and often to more bad questions.

Sometimes the answers aren't bad, they're just decidedly suboptimal. If your friend wanted to make a website, and you left it up to them to go google the answer, there's a good chance they'll end up using GoDaddy.

Chances are if you see this question, it means someone thinks you, either individually or as part of some community, will give them better answers than a machine will.

You can give them a huge leg up by giving them a programming vocabulary to help them on their way.

There's that famous saying about once you have a name for something the problem is pretty much solved...

You can also help them heaps by telling them what's important and what is not important. Most books on programming languages etc start with a whole bunch of stuff they should probably put into the basket of "computer trivia I don't need to know yet"...

1) Dreamers. They have something that nobody will do for them, so they're going to see how easy it is to do for themselves. They won't make it.

2) Doers. They don't have a single project. They have multiple projects. The projects are swimming in their head and they need to get them out. They just need a nudge in the right direction, but more help is almost always better.

Notice that neither of them are asking the wrong question. It's just a matter of follow-through. Given even a tiny bit of direction, the doers will make it there... But you can't start walking until you have some idea where the destination is. That's all they want.

The aptitude of a programmer is also the same that allows him to self-learn much of the art of programming by himself.

If they have an answer to the first question, I can work backwards with them to narrow it down to tools and languages that possibly make sense. Going with Python is a generic and common choice because there aren't many things you can't do in Python.

If they don't have an answer, it might not be a case lost. Some people truly need just about anything to work on, in order to impose a reason to start tinkering with programming. However, usually it's about overcoming fear for programming but in that case they probably won't make it, either.

For the rest 99% of people I just mumble something and point vaguely to teaching yourself to program in ten years.

As a person who has just recently undertaken the skill of programming, I (personally) feel that Google isn't much help in the beginning. As someone mentioned below, you can't ask what you don't know. Nor can you trust every bit of advice dished out in Google search results. If you rely on Google, you'll likely get 100 competing views on everything under the sun, which often leads to analysis paralysis or just sheer overwhelm.

As much as some people might tire of hearing "I want to learn to code, what should I do?", it's a question that I'd hope people aren't ashamed to ask, because It's an honest question. Pointing them to Google and advising them to figure it out just doesn't help them in any concrete way. I'm new to programming, but I'll never get sick of helping people learn the things that I already know- in fact I LOVE doing just that.

As a beginner, it's comforting to know that a real person is helping you along and not just an anonymous page picked out of the SERPs.

Often, we would have issues with client servers, and i would spend hours trying to find the root cause of the problem, with not too much luck. When i would pass the job onto him, he'd figure out the cause of the problem in a much shorter amount of time, and therefore come to conclusion for a possible solution very quickly and with some conviction.

I always thought he was a complete genius, and wish that i was as clever and intelligent as him, and felt a little cheated by the universe that i hadnt been blessed with such great IT skills, especially since i had great passion for IT.

Its only after much frustration, that i learned that the only thing that differed between he and i was his approach. Whereas i would check the virus state, and processor usage and look for anything suspect on the desktop, he would check the event logs and then look up the error codes at the microsoft knowledgebase.

One i started taking the same approach as he, i found myself fixing problems even faster than him and coming up with even better solutions to the problem (all the while experiencing much lesser stress).

When i decided to learn programming , i observed the same thing. His approach to attacking programming problems was different to mine, and now i am learning to assess the problems in the same way as him.

I've also seen this in my cousin. He starts a project, and takes one step at a time, and breaks the larger problem done, all the while telling himself that there is an answer for each component problem he faces.

I am trying to foster the same mindset now, and this blog post is an inspiration for that. Bottom line is, the approach is more important than existing skill set.

What I took away from the experience is to jump right in to the most advanced stage possible and work your way back to understanding. Baby steps only serve to hide the big picture and ultimately hinder your ability to understand. This is something I have applied to all of my learning since that day and I feel it has served me well. I always struggled in a formal education setting and I found out why: They teach backwards.

When you are learning to cook you don't mix flour and water together and look at the scientific properties that cause them to interact the way they do. No, you find a delicious recipe and prepare the entire dish. Each time you subsequently make it you add a twist to see how it affects the overall taste. Eventually you start to increase your understanding how the pieces fit together and can begin to come up with new combinations of your own. You will probably want to know why flour and water behave the way they do, but that comes later.

In my opinion, the way to learn how to program is to start with a recipe (i.e. source code). The rise of open source makes that easy. Make some small changes of your own. Break it down to understand why the original author did what he did. Once you are armed with the basics, you can start to apply that knowledge to meaningful Google searches.

If someone asks for help and this is the first thing you say, you're already thinking about it the wrong way.

What is their background? Their aptitude? Why do they want to learn? How do they learn best? Are their expectations reasonable? In short, what would be the smartest way to help this individual? One suboptimal way is to post essentially a multi-paragraph lmgtfy and assume that everybody will hound down the knowledge they need with the same geeky obsessiveness you would.

Edit: I meant to mention that remark about getting a Mac. "Embarrassed" to have considered anything else? I'm a big Mac fan (Cocoa developer) but that is ridiculous enough that I'm tempted to see the whole post as a joke.

I wrote about this last month, in fact:

http://blog.codelesson.com/why-learn-by-google-is-a-mistake/

Also as a beginner it is often very difficult to separate good from bad advice, so it can make a lot of sense to ask even simple questions in e.g. an IRC channel even if you could find an answer with Google.

If you can't teach yourself and 'self start' - you're not going to be able to get very far.

The author presumes that everyone interested in coding should have a high level of confidence already. I disagree, and feel that if someone wants to learn about a subject, asking questions about that subject is exactly what they should be doing.

There are so many different paths you could take when you set out on this quest that without a little guidance, you could waste a lot of time. Save learning to be an independent hacker for lesson two.

I've learnt to program long before Google existed and I never had this question. I was just learning, whenever I had a chance: reading articles, buying books.

Programming is one of the few subjects where it is possible to learn so much of it simply by searching online and practicing by yourself.

Anyone used http://twitter.github.com/bootstrap/? Seems like a good starting point for someone wanting to get going without creating your own CSS classes.

Nice, clean design and code. Think it will be perfect for mockups and user flow design. Going to give it bash, will report back if anyone is interested.