tag:blogger.com,1999:blog-6950833531562942289.post6171455661294795931..comments2024-03-25T03:36:48.099-07:00Comments on C0DE517E: Performance does not matterDEADC0DEhttp://www.blogger.com/profile/01477408942876127202noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-6950833531562942289.post-54759791356130975222008-06-13T22:11:00.000-07:002008-06-13T22:11:00.000-07:00Yes, I think we are in agreement if I read your co...Yes, I think we are in agreement if I read your comment properly.<BR/><BR/>On your "Framerate matters" post, you say "always have a programs that compiles... always not only something that compiles, but that actually works too."<BR/><BR/>My least favourite thing in programming is to be chasing down a bug, and find it in someone else's code with a comment beside it like /// @TODO Do this properly<BR/><BR/>Actually, there's one worse thing, which is when I find a TODO that I personally forgot about.<BR/><BR/>To paraphrase Yoda, "There is only do, or don't do, there is no TODO".Nickhttps://www.blogger.com/profile/07554585746004123227noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-80880692327114170182008-06-13T00:22:00.000-07:002008-06-13T00:22:00.000-07:00Mhm right even if I think that the best thing is t...Mhm right even if I think that the best thing is to think that any code could be the final one, or even better, that every design could be.<BR/><BR/>In other words, we have to avoid the "prototry" design pattern.<BR/><BR/>Thinking about shipping is not enough, as if a feature is fine, you'll be generally more than happy to ship it even if it's implementation is really poor (and the concept that every feature should be finalled early is already captured by the scrum philosophy, see my later post, "Framerate does matter")DEADC0DEhttps://www.blogger.com/profile/01477408942876127202noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-42679651840910123132008-06-12T23:55:00.000-07:002008-06-12T23:55:00.000-07:00I totally agree with your sentiments, and nice lin...I totally agree with your sentiments, and nice link from the 90's - good philosophy there.<BR/><BR/>Another rule of thumb I like to keep in mind when building new code is "this code might ship." Although it is true that premature optimization can cause you to waste time on something that is ultimately insignificant, on a big project, the deadline, and dozens of show-stopping bugs might arrive before you can go back and fix something that you wrote knowing that it was the wrong thing while you did it. If you remember "this code might ship", it will help you put the "right" amount of effort into every line you write - you don't need to make a polished diamond, but it'll help you not ship crap, too :)Nickhttps://www.blogger.com/profile/07554585746004123227noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-49660594410768492012008-06-07T04:11:00.000-07:002008-06-07T04:11:00.000-07:00Forgot to mention, the thesis itself is in english...Forgot to mention, the thesis itself is in english. The PDF is linked on the bottom of the page.Joneshttps://www.blogger.com/profile/01673391937701268035noreply@blogger.comtag:blogger.com,1999:blog-6950833531562942289.post-90299923909798916792008-06-06T11:53:00.000-07:002008-06-06T11:53:00.000-07:00Nice read. The renderer design your describing is ...Nice read. The renderer design your describing is strinkingly similar to the one I created during my master thesis. Have a look if you're interested: <A HREF="http://www.plm.eecs.uni-kassel.de/plm/index.php?id=713" REL="nofollow">http://www.plm.eecs.uni-kassel.de/plm/index.php?id=713</A><BR/>It will be open-sourced once the code is mature enough.Joneshttps://www.blogger.com/profile/01673391937701268035noreply@blogger.com