tag:blogger.com,1999:blog-6950833531562942289.post2199185885379371343..comments2024-03-25T03:36:48.099-07:00Comments on C0DE517E: You CAN'T afford doing things in the wrong way.DEADC0DEhttp://www.blogger.com/profile/01477408942876127202noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-6950833531562942289.post-49834115080230099332008-02-25T14:28:00.000-08:002008-02-25T14:28:00.000-08:00This post maybe should e printed and posted on eve...This post maybe should e printed and posted on every wall in our company :DAlessandro Monopolihttps://www.blogger.com/profile/01449768726687490782noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-60736840158445645422008-02-25T13:47:00.000-08:002008-02-25T13:47:00.000-08:00Of course changes will always happen. So you have ...Of course changes will always happen. So you have to design code for changes and flexibility. But it's also very useful, expecially in for us rendering guys, to at least TRY to minimize those. You really can't develop features blindly. You have to gather as exact as possible information on the requirements and gather experience and confidence on the methods you're going to implement. That does not mean that stuff can't change later on, but at least for the big features you should gain confidence in what you're going to do. And the easiest way to do that, for big and risky tasks, is to go through prototypes. Art guys can prototype the requirements by gathering visual references and do test renderers (do they need 1 or 10 lights? they can prototype that requirement directly in 3dsMax or Maya or whatever). Code guys should prototype the most risky things (100000 instances at 60fps? let's try if it's something that we can reasonably come near to in 2 weeks. It goes at 10fps? Then maybe the feature should be investigated more...)<BR/><BR/>Shouldn't this be OBVIOUS?<BR/>In my experience, it is not.DEADC0DEhttps://www.blogger.com/profile/01477408942876127202noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-70076616102238469252008-02-25T11:51:00.000-08:002008-02-25T11:51:00.000-08:00A very good set of guidelines, indeed. I just want...A very good set of guidelines, indeed. I just want to stress out one thing (clearly stated in the post, but whatever): you'll HAVE to make changes during the development. Design for later changes and implemet taking into account that your code must not only be interpreted by the compiler, but also be easy to understand by HUMANS. This topic might be worth of its own post, though.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-59798675265828283772008-02-23T00:38:00.000-08:002008-02-23T00:38:00.000-08:00Great post, thanks.Great post, thanks.Alehttps://www.blogger.com/profile/13117328634013352655noreply@blogger.com