Waiting for genus

I hate when I am stuck on some problem.  I am very stubborn when it comes to fixing some problem.  I’d often just power though problems.  Just keep trying ideas even if I know most of them will not work hoping I would just stumble on to the solution (or often stumble on to the real problem).

However this powering though often does not work. Sometimes you are just waiting for genus.

For example on FIFA I often have to fix some of our hardest bugs.  Some of these bugs take days if not weeks or months to fix.  Most of the bugs are hard to fix as they are rare and hard to find.  Reproduction steps for these bugs are hard to figure out.  So when you get one of these bugs you really want to go deep.  Release builds… no problem I debug the ASM.  Full symbols are a luxury items often.

This type of work takes time… some time you get luckily and you can see how you can shake the system to cause bugs like this and you can make your own easy to reproduce version of the problem.  But other times I feel I am just waiting for genus.  I have got all of the information I can get from this crash or hang and I am just waiting for some idea that will lead me to understand the problem and the eventual solution.

In this case if I don’t understand the problem yet I just try and keep the problem moving forward.  Can I find out the last few things that happened before this problem?  Can I find a way to see the problem 2 frames before the crash or hang? If you can write this code checking code and you can at least move the problem a step forward.

It is often best not to work on a few of these bugs full-time. You need to keep a few of them up in the air.  Just write enough code to make sure the next time you see it you can move it forward.  While you hope QA gets lucky and finds them again for you.

The worse time for me however is not on these bugs where mostly the problem might be hard to figure out why it is happening but it is when you want to do something like making games more fun.  These are the worst when the problem it’s self is just gooey and you can’t just keep breaking it down. You have to tackle the whole problem head on. If I don’t know what to do here I think you go to gym, or walk, get some tea… all you can do is wait for genus to hit you with some idea.

haskell

Didn’t have too much time tonight but started looking at haskell.   Reminds me a lot of lisp so far.  With First Last Rest and the like.  Haskell’s list comprehension is cool.  Maybe I can finally remember how to use list comprehension  in python.

Starting a website

Today I created a website. Welcome to the 1990s! Mind you with the new tools it is much easier than it was when I was going to school.

Setting up a web site is pretty easy.

  1. Registers a domain name (www.godaddy.com in my case)
  2. Found a hosting service (I used www.bluehost.com)
  3. Start playing with the host configuration (blue host uses cPanel)