4 ms·

This is probably true for going from novice to mediocre in many things, and there might be some skills where you can become great by constantly starting again.

by blowski 10y ago

This is probably true for going from novice to mediocre in many things, and there might be some skills where you can become great by constantly starting again. But in software development, the most valuable skills you learn are not those you pick up when throwing a quick CRUD app together, but those you learn when debugging, refactoring and optimising legacy code. I see too many junior devs get frustrated with something, throw it away, only to rebuild it with very similar problems because they didn't really understand the root cause in the first place.

flanbiscuit 10y ago
Having both worked in startup/tech companies and ad agencies, one of the most valuable skills I've learned is being able to decipher and debug someone else's spaghetti code. Clients shift agencies all the time and sometimes you're just given the entire codebase for site, with no documentation and no help offered from the previous agency, and you have to make an update to it or learn how some vital part of it works so you can recreate it in the new site. That's happened to me quite a few times and it made me a better debugger and coder.
shuntress 10y ago
There is a difference between writing broken code many times while never learning, and writing broken code once learning why it was broken then starting over and writing good code. If you are saying you can (and should) fix code but cannot fix, for example, a joke. I disagree. Only once you know what the problem is can you decide if it makes more sense to start over.
- blowski 10y ago
  I have no idea whether you can fix a joke. I'm not very funny.
  - michaelgrafl 10y ago
    I disagree.
lettergram 10y ago
I somewhat disagree. I'm in the process of building my 5th production rails app, and 8th web app (which will be used in production and make a profit). I've gotten ridiculously better at figuring out how to structure the underlying models. This alone has made me way better at structuring databases for queries, planning properly, etc. My day job is more of a data science role and it's helped a ton to know how to build the whole stack.
jimbokun 10y ago
I think a closer analogy is fast iterations vs waterfall design.

by blowski 10y ago

flanbiscuit 10y ago

Having both worked in startup/tech companies and ad agencies, one of the most valuable skills I've learned is being able to decipher and debug someone else's spaghetti code. Clients shift agencies all the time and sometimes you're just given the entire codebase for site, with no documentation and no help offered from the previous agency, and you have to make an update to it or learn how some vital part of it works so you can recreate it in the new site. That's happened to me quite a few times and it made me a better debugger and coder.

shuntress 10y ago

There is a difference between writing broken code many times while never learning, and writing broken code once learning why it was broken then starting over and writing good code. If you are saying you can (and should) fix code but cannot fix, for example, a joke. I disagree. Only once you know what the problem is can you decide if it makes more sense to start over.

blowski 10y ago
I have no idea whether you can fix a joke. I'm not very funny.
- michaelgrafl 10y ago
  I disagree.

lettergram 10y ago

I somewhat disagree. I'm in the process of building my 5th production rails app, and 8th web app (which will be used in production and make a profit). I've gotten ridiculously better at figuring out how to structure the underlying models. This alone has made me way better at structuring databases for queries, planning properly, etc. My day job is more of a data science role and it's helped a ton to know how to build the whole stack.

jimbokun 10y ago

I think a closer analogy is fast iterations vs waterfall design.