Ask HN: What's the most difficult part of learning to program?
23 comments
Getting stuck even though there's a better way (established pattern or architecture) that you don't even know exists.
If you work in a large company like Google, then there's lots of code and design documents you can read and, moreover, many helpful colleagues you can ask for advice (and who will provide detailed comments on your design documents).
If you're working on your own, or in a smaller company, then you might do things in a suboptimal way for a very long time just because you didn't see someone do it a better way. You might re-invent the optimal solution, but that's not the usual case.
Another way to look at this problem: "it's difficult to avoid having 1 year of experience 10 times"
If you work in a large company like Google, then there's lots of code and design documents you can read and, moreover, many helpful colleagues you can ask for advice (and who will provide detailed comments on your design documents).
If you're working on your own, or in a smaller company, then you might do things in a suboptimal way for a very long time just because you didn't see someone do it a better way. You might re-invent the optimal solution, but that's not the usual case.
Another way to look at this problem: "it's difficult to avoid having 1 year of experience 10 times"
Can you please explain what would be the best option to break out of the rut when one is mostly working on his/her own and doing a lot of suboptimal learning?
I frequently find myself in the second situation and only a handful of times have I been in the first situation.
I frequently find myself in the second situation and only a handful of times have I been in the first situation.
A few ideas based on things that have helped me:
1. Go to a local meetup group for your favourite language, take your laptop along, and show some folks what you've been working on. Show them what it does, and then walk them through the code. You can do this over beers. You will end up learning something either by explaining things to others, or by receiving questions and feedback from more experienced developers.
2. Ask a more experienced friend to meet you one weekend, for a couple of hours, to look at some of your code. Getting a code review from someone outside your normal environment will likely result in a few valuable tips.
3. Read books about architecture, like 'Web scalability for startup engineers' or books targeted at intermediate developers, like 'Powerful Python', and immediately go back to some past work to figure out where those patterns might be of use.
4. Ask any and all of your friends who are engineers about what they've been working on, what difficult problems they've had, and how they are working to solve them.
5. Read open source code. (This is a bit hard, as not all open source code is of high quality, and it's hard to know if it's good unless you're already a great developer.)
1. Go to a local meetup group for your favourite language, take your laptop along, and show some folks what you've been working on. Show them what it does, and then walk them through the code. You can do this over beers. You will end up learning something either by explaining things to others, or by receiving questions and feedback from more experienced developers.
2. Ask a more experienced friend to meet you one weekend, for a couple of hours, to look at some of your code. Getting a code review from someone outside your normal environment will likely result in a few valuable tips.
3. Read books about architecture, like 'Web scalability for startup engineers' or books targeted at intermediate developers, like 'Powerful Python', and immediately go back to some past work to figure out where those patterns might be of use.
4. Ask any and all of your friends who are engineers about what they've been working on, what difficult problems they've had, and how they are working to solve them.
5. Read open source code. (This is a bit hard, as not all open source code is of high quality, and it's hard to know if it's good unless you're already a great developer.)
Thanks! I really appreciate your insightful comments.
Personally, I think the most growth I had in software engineering was when I had my code reviewed by others, and I participated in reviews of other's code.
You learn a lot from other people by looking through their code, and you learn ways to improve your code by having people with more experience review your code.
Maybe find someone who is familiar with the language/application you are interested in and is willing to review your code.
You learn a lot from other people by looking through their code, and you learn ways to improve your code by having people with more experience review your code.
Maybe find someone who is familiar with the language/application you are interested in and is willing to review your code.
There were many different things, at different stages of my education and career.
But I think the hardest was the very first threshold I had to cross, when I first started taking legitimate CS courses. It had more to do with just the sheer ability to hold onto complex thoughts, data structures, algorithms, in my head.
I think a lot of engineers take this for granted, including myself. But I remember, just starting off, and realizing I felt like I was running out of working memory in my mind. It was almost like hitting a threshold where I really felt like I was stretching my mental ability to go through processes, run through algorithms, visualize concepts, etc. Eventually, as with anything, you get better and do these things faster and thinking about CS becomes second nature. But i specifically remember feeling like I reached my mental limits till I and enough practice to get good at thinking about CS and holding many things in my mind as I thought through algorithms.
But I think the hardest was the very first threshold I had to cross, when I first started taking legitimate CS courses. It had more to do with just the sheer ability to hold onto complex thoughts, data structures, algorithms, in my head.
I think a lot of engineers take this for granted, including myself. But I remember, just starting off, and realizing I felt like I was running out of working memory in my mind. It was almost like hitting a threshold where I really felt like I was stretching my mental ability to go through processes, run through algorithms, visualize concepts, etc. Eventually, as with anything, you get better and do these things faster and thinking about CS becomes second nature. But i specifically remember feeling like I reached my mental limits till I and enough practice to get good at thinking about CS and holding many things in my mind as I thought through algorithms.
Programming itself I think is fairly easy if you have the mindset for it. What isnt easy is making your program work with existing technology and programs.
Same goes for writing vs reading code. Maybe we teach and learn programming all wrong. For instance, when learning a natural language you dont read a manual on the grammar and then go off and try to write a novel looking up vocabulary along the way... There is absolutely no way you would be succesful with that approach, you would invent your own messed up version of the language that seems absolutely strange to everyone else. But that is sort of what we expect people to do when they learn to program.
It might be better to spend a year reading and dissecting good code of varying complexity in a debugger before you ever even write a line of code.
Same goes for writing vs reading code. Maybe we teach and learn programming all wrong. For instance, when learning a natural language you dont read a manual on the grammar and then go off and try to write a novel looking up vocabulary along the way... There is absolutely no way you would be succesful with that approach, you would invent your own messed up version of the language that seems absolutely strange to everyone else. But that is sort of what we expect people to do when they learn to program.
It might be better to spend a year reading and dissecting good code of varying complexity in a debugger before you ever even write a line of code.
Realising that maintenance is the job and when you get to looking at legacy code, that it's not crap, but just represents a person/team's point in time understanding of a domain. Your understanding is different or maybe the domain has changed (no business stands still) and so you would code it differently here and now. But you might not have back then with the understanding and knowledge the team had when they wrote it.
Realising that in x days/months/years your own code will be the legacy and viewed by other developers and yourself. Your code will represent your understanding at the point in time you write it. It's natural if it looks horrid or incomplete or bad or that you missed something obvious. Don't beat yourself up over it.
Look on any code with the understanding and compassion with which you would have the next developer look on your code.
That and making sure to get up, stretch regularly and look out the window and get out the building. Smell the roses. That sort of thing.
Realising that in x days/months/years your own code will be the legacy and viewed by other developers and yourself. Your code will represent your understanding at the point in time you write it. It's natural if it looks horrid or incomplete or bad or that you missed something obvious. Don't beat yourself up over it.
Look on any code with the understanding and compassion with which you would have the next developer look on your code.
That and making sure to get up, stretch regularly and look out the window and get out the building. Smell the roses. That sort of thing.
As a self taught programmer who has mostly utilized project based learning, the most challenging parts for me have been:
* Knowing my own knowledge gaps. I know there's a lot I don't know, but it can be difficult to pinpoint _what that is_ exactly because if I don't encounter it in a project I may not even see the gap until I _do_ encounter it one day. Supplementing with short structured courses like Pluralsight, Safari Books Online, etc has helped in this regard.
* Being confident that you're solving things in a good way. "This approach works, but is this really the best approach or is there some existing pattern or data structure I should be using for this that I just don't know about yet?"
Both of these things have become easier to address with time and practice.
* Knowing my own knowledge gaps. I know there's a lot I don't know, but it can be difficult to pinpoint _what that is_ exactly because if I don't encounter it in a project I may not even see the gap until I _do_ encounter it one day. Supplementing with short structured courses like Pluralsight, Safari Books Online, etc has helped in this regard.
* Being confident that you're solving things in a good way. "This approach works, but is this really the best approach or is there some existing pattern or data structure I should be using for this that I just don't know about yet?"
Both of these things have become easier to address with time and practice.
I think it's hard to believe in yourself. If you don't believe in yourself, it's harder to be persistent enough. The lack of persistence fuels the lack of belief in yourself and the loop continues until you give up for good.
Writing code is hard, but confidence and persistence will get you a lot closer.
Writing code is hard, but confidence and persistence will get you a lot closer.
Learning to program is actually pretty easy. You can get up and running in a few minutes with your first Hello World program and that's how most programming languages are designed.
For me personally, the hardest part was having the patience to actually manage to build something useful other than the sample toy programs that came with the textbook.
You need to have the patience to read tons of documentation, google for example code, ask questions on Stackoverflow when stuck, so on and so forth. Once you do this for a single project and get something useful built, it gets much easier :)
For me personally, the hardest part was having the patience to actually manage to build something useful other than the sample toy programs that came with the textbook.
You need to have the patience to read tons of documentation, google for example code, ask questions on Stackoverflow when stuck, so on and so forth. Once you do this for a single project and get something useful built, it gets much easier :)
Becoming fluent in the language syntax is potentially the hardest but it is also the most important because of all the imprecise language that people use to write about programming.
If you are language fluent then you can read the source code and know exactly what is going on, otherwise you are stuck trying to make sense of someone's hyped up abstraction of a basic programming design pattern.
If you are language fluent then you can read the source code and know exactly what is going on, otherwise you are stuck trying to make sense of someone's hyped up abstraction of a basic programming design pattern.
For me it's going from knowing the syntax of a language to actually designing and building something useful. No one really gets this right as far as I can tell. I can solve lots of stupid little puzzles in several languages but when it comes to building something without a tutorial giving me step by step direction I'm completely lost.
Then break that 'building something' into 'lots of stupid little puzzles'. That is the trick!
Learning what to learn.
I eventually got frustrated with books and courses and instead just visualized what I wanted the page/site to look like and do.
I then googled a lot of problems I was having ("How do I sort columns in an HTML table using JS") and then made code work.
For especially tricky problems, I would ask my friends how they would approach a problem.
I eventually got frustrated with books and courses and instead just visualized what I wanted the page/site to look like and do.
I then googled a lot of problems I was having ("How do I sort columns in an HTML table using JS") and then made code work.
For especially tricky problems, I would ask my friends how they would approach a problem.
I think understanding how to use abstractions, naming, and simplicity are things you learn from experience. If you could start out with some good rules of thumb covering these, it would go a long way towards making you a better programmer.
I think it is the fact that most of the time
a) things don't work b) it is hard to say why they don't work
Annecdotaly, from looking at my mom, a 50+ year old economist working on insurance backend platform, any sort of error would be hugely discouraging.
a) things don't work b) it is hard to say why they don't work
Annecdotaly, from looking at my mom, a 50+ year old economist working on insurance backend platform, any sort of error would be hugely discouraging.
Learning how to balance short-term benefits (with long-term costs) against long-term benefits (with short-term costs).
That's why people talk about it being "easy" to write a program that compiles and runs, but hard to write good code.
That's why people talk about it being "easy" to write a program that compiles and runs, but hard to write good code.
getting stuff to work initially (just break tasks down, test each bit, eventually you will have things going)
worrying you are doing something wrong - you probably are but remind yourself you just have to start somewhere. As you get it more you will learn of better ways to do stuff. There will be times when you paint yourself into a corner and have to redo just about everything, common even for experienced programmers.
Then I would say getting your brain trained to understand what you can do with a language, how the computer reacts to different program tasks, and how to write stuff so the user gets what they need.
worrying you are doing something wrong - you probably are but remind yourself you just have to start somewhere. As you get it more you will learn of better ways to do stuff. There will be times when you paint yourself into a corner and have to redo just about everything, common even for experienced programmers.
Then I would say getting your brain trained to understand what you can do with a language, how the computer reacts to different program tasks, and how to write stuff so the user gets what they need.
Getting dev environments set up and debugging obscure error messages. Persistence/stubbornness to fight through these obstacles is one of the most important traits for programming.
learning and setting up an IDE. ;)
probably consistency and dedication. there is so much to learn. so finding that area you love is the key.
probably consistency and dedication. there is so much to learn. so finding that area you love is the key.
1. It's much harder to read other people's code than to code yourself, so it's very easy to get put off by the apparent complexity of other people's work.
But you don't know why their work is like that. You don't know if their coding is actually good, you don't know what odd OS/browser/language version bugs are being fixed by the weirdness there, you can't see all the work that went into the program before... It can be misleading, and make you feel like you 'don't get it'.
2. Tutorials being very heavily slanted towards the beginner and expert ends of the spectrum. It's harder to go from intermediate to expert than from beginner to intermediate or learn more as an expert, because few tutorials are written for that level of expertise.
3. Similarly, figuring out why something may work the way it does on a fundamental level can be tricky. Tutorials often have a 'just trust us on this' attitude towards explaining tihngs, and don't really go into why different types of variables work, why loops work, why say Redux works the way it does, etc.
4. And I guess for some people, just the whole mindset can be difficult to get into. Programming is a very precision driven exercise, and for people used to the 'good enough' way that most of society works, that can feel jarring and awkward. It can feel like being the world's strictest dog trainer, with the computer or language or compiler or what not being a pooch that can't figure out what you want from it unless you teach it in the exact right way.