-- Posted by sakurag at 9:01 am on Oct. 8, 2008 So, in my day to day life I get to interview really, really smart people.  I ask them some questions (usually to do with fun exciting topics), and they try to answer them for me.  Now you all, being super smart.. I want to see if you can answer one of these question.  Most people don't seem to get them right.. it bums me out.  Maybe someone here will astound me:  If you know the answer, just don't blurt it out.. let people who don't know solve it. Question: I have an image on my computer.  It's got a black background and a white triangle in it.  The image is stored in a bitmap, each pixel contains the color value.  On your own, w/o a paint program, tell me how you would create an outline of this shape from the original image.  Ie, how would your paint program go about doing this. For those of you who don't understand how a computer image looks like up close, think of like zooming in really close to a picture, things get kinda pixelated.  that might help you out. -- Posted by hI jAMES at 3:23 pm on Oct. 8, 2008 You could always use the magic wand.  I didn't break any rules, did I? -- Posted by allsmiles at 3:58 pm on Oct. 8, 2008 How the program, as opposed to the user and the UI? Presumably, it would start from the centre-most pixel and spiral outwards, switching the values of pixel pairs based on orientation from the previous pixel. This would generate a one line black triangle surrounded by a one line white triangle. Then you would just tell it to fill the centremost pixels with black until they encounter a black pixel in their direct vicinity. I dunno, sounds like how it would operate. I'm not all that good with image manipulation, only spent one lesson on it in physics. -- Posted by sadnessness at 11:23 am on Oct. 9, 2008 why does being smart mean we know all about computers and pixels? the actualy intellectual test was nothing to do with it either lol -- Posted by whoami111 at 9:41 am on Oct. 10, 2008 Quote: from sadnessness at 11:23 am on Oct. 9, 2008 why does being smart mean we know all about computers and pixels? the actualy intellectual test was nothing to do with it either lol QFT not all of us are computer geniuses -- Posted by sakurag at 10:26 am on Oct. 10, 2008 Quote: from whoami111 at 9:41 am on Oct. 10, 2008 Quote: from sadnessness at 11:23 am on Oct. 9, 2008 why does being smart mean we know all about computers and pixels?    the actualy intellectual test was nothing to do with it either lol QFT not all of us are computer geniuses It's mathematics, actually... -- Posted by Feminawesome at 2:32 pm on Oct. 11, 2008 Go to the paint program. CTRL + A. CTRL + C. Create a new document. CTRL + V. CTRL + A. Inverse colors. You now have a black triangle~ Done~ -- Posted by allsmiles at 11:14 am on Oct. 12, 2008 That does not create the outline of a triangle. Also, everything but inverse colours was redundant. -- Posted by tell me again at 6:36 pm on Oct. 12, 2008 What the hell, you interview really, really smart people? What do you actually do? I hope your questions are better than this one, because this question isn't clear in what it's asking. But I presume it's to do with contrast - wherever colour values change, there'd be a dot, determining the "edge". -- Posted by medjai at 9:10 pm on Oct. 15, 2008 Easy. Imagine a big ass grid. That's your bitmap potential image space and it's filled with a shitload of squares, each square can hold one color attribute and they ALL have the same shape attribute (a tiny square of space). So obviously a real smooth diagonal is impossible, instead it has to use an equation which determines a means to, using shading, create an optical illusion of a smooth and sharp image. I don't know the equation or algorithm that is used in the market programs I just know it must exist, they may even be proprietary and unique to each program. Either way, you fade out the 'bad' squares by using lighter shades of the color you are representing the triangle to be. Am I right? -- Posted by medjai at 9:12 pm on Oct. 15, 2008 Oh, and my paint program would use vectors, eliminating resolution dependency.