Hacker News Reader

Comment by tom_rath | Hacker News Reader

How could he have neglected to mention Test?!

Test everything. Test it early. Test it extensively.

If testing 'isn't needed' and is a 'waste of time' for the simple thing you're doing, it'll only take a few moments away from your busy day. If it was needed, you'll be glad you caught the error when you did.

If in doubt, ask any veteran developer about the horror stories they ran into in production due to a minor bug that could have been caught at the point of development with a simple 'complete waste of time' test.

Edit: I cannot understand why this is being down-modded to negative territory. This is classic 'advice to a young engineer'! Your development career is headed to a dead end if you cannot create reliable software.

View on HN

I suspect people are (edit: were) downvoting you because of backlash to TDD (and other testing ideology) zealots.

Knowing how to test software, how to write software that can be easily tested, how to systematically reproduce bugs to can find their underlying cause, etc. is excellent advice. Just don't focus on specific tools for testing so much that you miss cultivating good analytical skills (or forgive sloppiness).

I know plenty of people who love to test everything and for some things it can be useful however I think the problem is more basic; Know The Code.

If you write something that you think might work test it. If you know the code works, let other people test. Testing crap does not get rid of the smell even if it works but sometimes you do what you need to do.

> Test everything. Test it early. Test it extensively.

While in some contexts appropriate and many times helpful, testing is not the way to create reliable software.

Willingness to learn from mistakes, knowing your limits and biases, practicing complete concentration, disciplined reasoning and careful thinking is.

Yes, testing is the way to create reliable software. How else can you find the 'mistakes, limits, biases', etc. which come from the 'concentration, reasoning and careful thinking' you THINK you have?

Learning how to break your software is one of the best ways of ensuring you stop writing software that breaks!

If you don't reason properly, no amount of testing will save you. It is only part (an important one, surely) of what you need to create reliable software.

I saw too many developers in a cycle like this:

test->bang keyboard->test again (not working)->bang keyboard (now a hack) ->test again(still not working - let's try one more thing)->bang keyboard->test again (it works - Don't touch it!)

which is why early focus on testing doesn't seem very sound advice. The original post is about "Career Advice For Young Developers" - and for a young developer, focusing on careful thinking is better than focusing on testing.

Reply on news.ycombinator.com