Here I would like to discuss some aspects of modern computer graphics. The article is questionable. Yes it is, even from my own point of view. I don't want state My Not Very Humble Opinion about it. I just want to raise some issues to discuss.
There's a great success in hardware accelerated graphic engines. The game industry really made a progress in it. Any modern video adaptor is a separate computer. Moreover, they are not just computers, they are monsters! In parallel computing, nothing is being developed faster than video accelerators (we don't take into account specialized fields like supercomputers, we discuss computer mainsteream). Modern graphic accelerators can create terrific video effects for no extra cost! The possibility to incorporate parallel vertex and pixel shaders is a real breakthrough.
It is a real breakthrough, but not in computer graphics in general, it's a breakthrough in video effects. Any powerful graphic accelerator (again, a mainstream one) is just a triangle machine. I would say, it's a monster triangle machine, but still, nothing more. Any advanced parallel programming technique won't help us draw a flawless poly-line of arbitrary width. Yes, it also needs to properly display given line joins, line caps, and translucent self-intersecting poly-lines. And of course, with perfect anti-aliasing that would simulate lines of less than one pixel width (and in general, lines of fractional width).
From now on I will tell you about the failure of general computer graphics. Just as the beginning I state the following.
Despite the great success in video effects, the problem of drawing a simple polyline (of any width, any opacity and with given line joins) remains a task. I would say it remains a "mega-task". Nobody could solve it so far. Why? Just because of the market. People just don't need it. Most people want bread and circuses. Right here and right now. It's not my sarcasm, it's the reality. I'm not an exclusion, I want it too as anyboby else.