What code should I read?
-
There's a lot of the .NET world that is now open source (Roslyn code compilers, Entity Framework)...I'd say reading through how they organise their code would be somewhat informative.
Great suggestion. In fact, at the time your post came in, I just happened to be reading this blog post, which pointed me in the same direction. I also just happened to come across this tool, which I should think would come in very handy when spelunking some of the more cavernous codebases out there (like Roslyn). Thanks!
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
I started my development career in unit testing. One of the things we had to do is code review on every unit (function). I also had to test it of course which forced me to look at the code / design requirements, and to write tests for it. But... starting with testing teaches you a whole lot of bad habits. The quality of my tests were never looked at, so I didn't really care about how/what I coded at the time. But when I started coding, I could use the things I learned while testing others code and the transition went smoothly.
"Program testing can be used to show the presence of bugs, but never to show their absence." << please vote!! >>
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
Study the design (code may not be necessary) of the C++ STL API. (good API design - hard to create inefficient code) Followup: Compare it with the Java collections API design. (especially LinkedList methods that use or return indexes/offsets!) (okay API design - too easy to create inefficient code)
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
I advice you to write more code, not read. Write everything you wanna learn. In the reality programming is separated as follows: First half is the "art" people. They do write code. They usually do it for personal pleasure, earning money is just a pleasant side effect. They lead the innovation. As you learn more APIs, platform and languages, as you usually heard, they have a single inventor, not a company. It is not useful skill to find a work with, neither appreciated in most companies. Second half is the "science" people. They do not need to read and write code, because they know the theory. The theory is really complex explanation of something really simple, which once you learned the complex stuffs seems you may thought of it without learning, but you didn't. You can explain the innovations of the "art" people and can handle anything new you are thrown into. To the company (especially management) you are like fast-learner, so you are valuable. This type of skill is really useful to find a job. Third half is the "business" people. You read a lot of other people code, trying to understand their ideas. You do not need your own ideas, you follow the ideas of other people, which are usually so many that they will overwhelm you. As you get older (and more experienced) you will begin to prefer legacy code you know, than the new APIs you'll have to learn. To the management you are the one that do the all the work. This skill is not useful to find a job, but really useful to keep it.
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
As always, It Depends. I recommend this a lot to new coders. People learning a new language. I also recommend learning/reading code that solves problems you understand, or you truly want to understand. LZW Compression was eye opening. Also, are you reading for a Novel experience or to emulate the Greats? Learning a new language or fine tuning your skills? It all matters. Are you going to skim and then focus or focus on every line of code, executing it in your head? Reading many of your replies, it sounds like you are pretty far along in programming. You are more interested in learning from the masters. To me, that means delving down into the more complicated code like Roslyn. I enjoy reading well written libraries the most. Even the elegance of x = passed || default; to initialize something conditionally is cool. A line of code is like a single thought. Developing such thoughts to harness a little more value than 3 lines of code is cool. Of course, always read the old C standard library code!
-
I dunno... I've seen code from MS that leaves me scratching my head and wondering :omg: :omg: :omg: :confused::confused::confused: WTF was this guy thinkin?
If it's not broken, fix it until it is
-
There's a lot of the .NET world that is now open source (Roslyn code compilers, Entity Framework)...I'd say reading through how they organise their code would be somewhat informative.
If you're the person who wants to drive a humvee or a unimog, then microsoft code is for you. It's ugly. It's functional. It's uninspired. It's very durable. I have never seen microsoft code that was in any way elegant or subtle. It's ok code to learn on, but will never inspire you with beauty.
-
Come on, OG. I expected this from some of the others, but not from _you_! Would you advise an auto-designer not to study a Lamborghini? Would a growing architect gain nothing from a study of the Burj Al Arab, or the One World Trade Center? Can an aspiring composer learn nothing from analyzing the techniques Beethoven, Bach, and Bublé? The same argument about "the final product" could be made about each of those fields, but it wouldn't hold up. Of course I would love to study the personal notes of Tolkien where he divulges all his inner grapplings with plot twist connundrums, but in lieu of that I am still a much better author having merely _read_ LOTR three times. All I am saying, is that when it comes to code, it is much more difficult (for me, anyways) to find the open source code that is worthy of being studied. So I am simply asking for recommendations. Have you ever read a program -- perhaps in a completely different field than your own -- which made you say, "Wow, that was put together well. It's intuitive, clean, elegant, and robust." I think somebody needs to start compiling a list of such masterpieces for the rest of us to study and admire.
kdmote wrote:
Would you advise an auto-designer not to study a Lamborghini?
It's easy to spot a lamborghini by its smooth lines and the sound its motor makes. With code, you have to get it into your head before discovering if it's a lamborghini or a rusty Yugo with 100,000 miles on it. For every epic software poem, there are 10,000 drab tomes of uninspired code.
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
Anything with 101 samples... Official Visual Studio 2010 Samples for C# 4.0 in C#, VB.NET, C++[^] 101 LINQ Samples in C#[^]
-
kdmote wrote:
Would you advise an auto-designer not to study a Lamborghini?
It's easy to spot a lamborghini by its smooth lines and the sound its motor makes. With code, you have to get it into your head before discovering if it's a lamborghini or a rusty Yugo with 100,000 miles on it. For every epic software poem, there are 10,000 drab tomes of uninspired code.
-
Anything with 101 samples... Official Visual Studio 2010 Samples for C# 4.0 in C#, VB.NET, C++[^] 101 LINQ Samples in C#[^]
-
I dunno... I've seen code from MS that leaves me scratching my head and wondering :omg: :omg: :omg: :confused::confused::confused: WTF was this guy thinkin?
If it's not broken, fix it until it is
You mean those "example code" sections that presume you have reached the page after having read the preceding twenty, when really all you did was search straight to this one? I may not last forever but the mess I leave behind certainly will.
-
EXACTLY. Which is why I'm asking for recommendations. I'd like somebody to start compiling a list of the epics, so we can all benefit from studying them.
I've never looked at it in total to be sure, but Stepanov's original Standard Template Library is probably brilliant. Unfortunately, the version that comes with Visual Studio has been messed up with non-core stuff and compiler dependencies, so it's quite hard to understand.
-
[I hope this doesn't cross the line of what I _shouldn't_ post about here...] I was just reminded of the oft-given advice that to improve your coding skills you should "read lots of code". This made me wonder: _What_ code should I read? If I were a painting instructor, I would advise my students to study the masters: Michelangelo, VanGogh, Norman Rockwell, Bill Watterson. But what about coders? What are some of the classic code-bases that every aspiring software developer should read at least once in his/her life? Alas, some of the best are doubtless locked up behind proprietary firewalls, but of the code that is publicly available, which programs would you suggest are worthy of studying or even emulating?
kdmote wrote:
If I were a painting instructor, I would advise my students to study the masters: ... Bill Watterson
I agree!