Re: Data oriented programming

From: [email protected]
[mailto:[email protected]] On Behalf Of Tony
Albrecht
Sent: Wednesday, October 06, 2010 3:47 PM
To: [email protected]
Subject: Re: [Sweng-Gamedev] Data oriented programming


> Your terse comment lead me to think that you didn't like the 
> idea of considering performance before implementation as it 
> could lead to analysis paralysis. If you were just making a 
> statement about analysis paralysis then, I agree, yes it can. 
> But the fear of analysis paralysis (analysis paralysis paralysis?) 
> is more likely to have a detrimental long term effect on your code 
> base than the former. 

> -Tony

My point is that at some point you have to start writing code. Too often I
see engineers afraid to write a single line because they have not thought
through (and therefore designed for) every single possibility in advance.
Nothing we talk about on this list rises to the same level of coding for
space shuttle or Mars missions, so while considering algorithmic complexity
or not making the obvious performance mistakes during implementation are
Good Things (as has been implied previously), a product needs to ship, and
IMO it really shouldn't go much beyond that before you start writing code. 

The fact is that pretty much everyone here has already done before, whatever
it is they are embarking upon, so experience goes a long way towards
encouraging things that improve performance, even without a ton of thought
up front.

I suppose that wasn't exactly conveyed by a one-liner... 

Greg

_______________________________________________
Sweng-Gamedev mailing list
[email protected]
http://lists.midnightryder.com/listinfo.cgi/sweng-gamedev-midnightryder.com