To expert developers: help to evolve in development

Hi for all

First of all thanks to read this. I really need a lot of opinions.
I'm posting this topic to get a help of you, expert developers. I know the answer of this question is generic. I'm a C++ developer( like you are ) with a problem that affects you or have affected some time. It's about learning.
I explain. I know C++, some subtleties of the language, good knowledge of STL and moderate knowledge of templates.
My torment lies in all of this. I think I have a considerable knowledge of this stuff. But when I get code of others and write my code I never know to say if it's a good solution or not. That means, I can't evaluate properly the code.
When I show the design of my solutions to an experienced developer, their solution of the same problem generally is better than mine. I realized what I think it's a good solution, isn't not many times.
This explains my post. Now I would like to know of you, expert developers and solution designers, what you consider a expert C++ solution developer. I mean, what you consider a developer must know to effectively use C++ and when to apply good solution.
I need good practices and design, I think.

For you all, thanks for your replies

Get this book and read it.
http://www.amazon.com/Effective-Spec.../dp/0201924889

It's a starting point!

Take an algorithms class, maybe. Until your development process naturally includes consideration of big-Oh times, you're liable to miss obvious algorithmic improvements.

Also, be weary of the "grass is always greener" phenomenon. When you start working heavily with other coders, you'll find that some aspects of their style seem superior to yours; others you may not like. Just adapt what strikes your fancy to your own future work and ignore the rest.

Also, always remember: The best code is working code. At the end of the day that's what is most important.

First of all, read all the good, recommended books, or at least the good sites. Also, take some formal classes if you haven't, and hang around and learn from those more experienced than you. Watch what they do and always ask why they did it that way. Of course, like many things, being a good programmer takes practice. You can take all the classes in the world and read all the books in the world, but the main thing you need is experience. Try starting your own open source project or something.

Thanks for all replies dudes. This will help me a lot. If you can post any more comment, fell free.

Quote:

---

Originally Posted by Lindley

Take an algorithms class, maybe. Until your development process naturally includes consideration of big-Oh times, you're liable to miss obvious algorithmic improvements.

Also, be weary of the "grass is always greener" phenomenon. When you start working heavily with other coders, you'll find that some aspects of their style seem superior to yours; others you may not like. Just adapt what strikes your fancy to your own future work and ignore the rest.

Also, always remember: The best code is working code. At the end of the day that's what is most important.

---

Big-O is the most important one, but you do also want to consider Big-Omega, and Big-Sigma.
For instance, Q-sort is Big-O n log n, and Big-Sigma of n, but Big-Omega of n squared. And it's predictable which one you will get depending on your data. That will dictate which sorting algorithm you should use.