anonymous133247 | | | | | login

Vying Games : Blog

Comments

Anonymous says,

Frankly, anyone who really "gets" your point will already have developed the cognitive skills necessary to recognize when they are or should in fact be just hill climbing.

So..... if your target audience will only have to members, those that can't get it, and those that already do; why bother writing/explaining it in the first place?

And it's not even 9am here. ;-)

Eric says,

I actually think I ended this article poorly, because I'm not sure I really made the point I wanted to. There's nothing profound about the hill climbing comparison. For me personally, the value is in changing my frame of mind.

I often want to hit a home run on the first version of some code. Inevitably, I'll only hit a single, or worse a fly out. It's discouraging when code doesn't turn out as perfect as envisioned.

Remembering that the whole process is like hill climbing takes some of the sting out of not being perfect.

So, I was actually writing to remind those who, like myself, get it but are prone to forget it. :)

Anonymous says,

Just to nitpick: you probably mean np-hard (and not np-complete), and exhaustive checks have little to do with both...

Perry The Cynic says,

Note one cautionary corollary: hill-climbing tends to approach the local optimum, particularly when it's done well. You may be climbing up a nice hill, away from the towering mountain across the valley behind you. Hill-climbing is great for keeping in motion and making progress (and thus combat analysis and management paralysis), but poor for measuring that progress against objectifiable standards. For that, it's necessary to take an aerial snapshot once in a while. And do watch the other climbers, particularly the ones going in other directions. :-)

Cheers -- perry

Anonymous says,

Once you are up the local hill, you can see better where the mountain is.

Eric says,

Anonymous, I think you're correct to nitpick the use of np-complete. I probably shouldn't be so casual with such a well defined academic term.

Anonymous says,

> Once you are up the local hill, you can see better where the mountain is.

Don't be silly - he's not talkking about hills. In a hill-climbing exercise you always think that the hill you're climbing is the global uptimum, otherwise you wouldn't climb it.

Adam says,

The problem with hill-climbing leading to local maxima is probably more a problem with the analogy than with the point. Rather than talking hill-climbing, how about simulated annealing (http://en.wikipedia.org/wiki/Simulated_annealing)?

Alright, so most people don't know what annealing is... but once you read the wiki entry, it's even more appropriate to software development:

1) start with any solution.

2) re-implement with a probably-very-different solution based on lessons learned.

3) repeat... if you are learning your lessons, successive implementations will differ less and less.

4) stop when the improvement from another iteration isn't worth the delay.


Commenting on this article has been disabled.

Articles by Tag

Search the Vying Games Blog