Any scenario where the SVG's shapes are rendered with anti-aliasing, and the edges of both overlapping shapes are mid-pixel: the edge of the top shape would be rendered with partial alpha, and then beneath that, the edge of the differently-colored lower shape would bleed through as it is also rendered, at that contiguous line, with partial alpha.
EDIT: You can actually see this very effect in action if you zoom in on the diagonal edges of the yellow and blue faces of the logo on the BBC's page for the EU referendum happening right now: http://www.bbc.com/news/politics/eu_referendum/results - on these anti-aliased diagonal edges (enlarged: http://i.imgur.com/GmxUD60.png), since there's no indentation like the one this SE question highlights, the red shape underneath both of the shapes bleeds through.
You can see similar problems even without anti-aliasing. If you have three collinear points A B C, drawing a line from A to B and then B to C may hit slightly different pixels compared to just drawing a single line from A to C. If you try to draw a polygonal mesh which has T-junctions (overlapping edges whose endpoints are not identical) you can end up with slight gaps between adjacent polygons.
It's explained in the answer on the linked page (and in the top comment). Depending on how it is rendered (due perhaps to a buggy renderer) it might show on a screen.
I am @iconjack, poser of the question, ask me anything! Kidding aside, as an avid user of the whole stackexchange suite, it is kind of interesting to me how popular this question became. It was only my second question on graphics design and it's become the all-time 3rd ranked question. Keep it going, folks!
If you're curious, the reason I was looking at the logo svg was that I thinking about making a bumper sticker. When I saw the notches, I was perplexed, and threw the question up on a whim.
You noticed I put the disclaimer at the top about the question not being about politics. I honestly thought there was a chance things would get ugly because of the loose tie-in to the whole election grotesquery. One time I asked a question on english.stackexchange.com, Is "Sent from my iPhone" correctly punctuated and capitalized?" and I practically got hate mail. Not only did I get snarky comments and answers, I believe the question was down to −8 at one point. It's since been upvoted into positive territory and some of the nasty remarks have been deleted, but I never really got what all the hate was about. One friend suggested that Apple fans would think I was dissing the iPhone or something. That's the only theory I've got.
Weird, this link is giving me infinite redirects. I went and found it on the stackexchange site myself under hot for graphicdesign and it did the same thing. The other answers on that subdomain work just fine. Strange.
Do you use HTTPS Everywhere? (I don't know why that would apply in this case, but I recall various people having similar issues with it and Stack Exchange's partial HTTPS support.)
I guess it depends on your definition of "solution" and "MSAA moral equivalent". What you're essentially asking is, "Is there a way to approximate the combined contributions of overlapping polygons to a set of pixels without sampling the exact color at multiple points per pixel?"
One approach would be to avoid sampling at all, by using boolean geometric operations to find the exact shapes of visible polygons within a pixel, then using their color and areas to calculate their combined contributions. I expect the calculation involved will be impractical relative to multisampling, though.
Another approach would be sample exactly once per pixel or, equivalently, render the polygons without antialiasing. Naturally, that would eliminate the show-through of hidden geometry at the expense of the accuracy of pixels around polygon edges. You might then be able to approximate the antialiased values by a post-processing filter, maybe guided by the presence of edges, although doing so will introduce its own set of artifacts.
And if you consider "one sample per pixel" to be a budget over the entire image, you could sample the polygons more densely near polygon edges and more sparsely elsewhere, using nearby samples to approximate the values of pixels with no samples.
When working in finance, you care a lot about public holidays, because interest on wholesale loans between banks is calculated per business day. On global scale, there are more public holiday schemes than countries or timezones. And some countries change their public holidays willy-nilly on short notice.
capote - Don't take it personally. I would read the guidelines; it's just the nature of this community and has been for a long time.
If your comment won't contribute the conversation (in an intellectual sense) don't post it. Some people do post useless comments but generally they are voted down; also, the standard is not the worst comments.
Thanks for the level-headed verbal answer. I guess I just feel like a downvote is a slap, rather than a more appropriate simple response that states the issue, or just being ignored.
Slaps should be for when people are offensive/racist/rude/wrong--I understand my comment contributed nothing, but why would I get the same treatment as someone who contributes negatively?
Shouldn't we be aiming for positivity => upvote, negativity => downvote, neutrality => no vote?
It's not; that's all I can say. I've gotten plenty of them. Note that nobody else knows the score of your comment unless it's < 1.
> why would I get the same treatment as someone who contributes negatively
FWIW, comments that don't contribute do have a negative impact. They encourage more of the same, reducing signal-to-noise. The behavior spreads like a virus.
This link may explain some of the ideas behind how HN works:
I worry that if we expand the definition of acceptable humor on HN, then HN would become something more like reddit. Not that reddit is bad, but I appreciate that the expectation here is to hold comments to a strict standard.
Humor is allowed of course, and sometimes it's highly upvoted, but it tends to be humor that is on topic (and so usually about tech or the tech industry) and also as a part of personal stories. People especially have their guard up here because this thread is tangentially related to politics. And like anywhere else, you have to make the right joke for the room, as in your joke needs to be somewhat harmonious with the mindset of the majority (especially if it's off topic, not related to tech, or inflammatory).
You should aim for substance. Comments going on about downvotes are the antithesis of substance, which is why we have not one but two guidelines asking you not to do this: https://news.ycombinator.com/newsguidelines.html.
Your original problem was that glib comments aren't substantive either, so they're deprecated here, especially when they have an uncivil tinge.
With all the attention they got, roughly 30 downvotes, flagging, your attention, you'd think I posted the most substantive posts this site has ever seen.
Because your posts are boring and aren't contributing anything. Nobody cares that your comments got downvoted, so why waste everybody's time talking about it?
I think your initial reply to spb was downvoted because it came off as sarcastic sour grapes that somebody had answered the question better than you.
I'm seeing a ton of (rightfully) down voted comments here with no clear reason from the down voters. Anyone who gets a down vote for their joke might just think it's because the joke is bad, instead if the more accurate reason that HN comments are more serious.
One of the nicer things I've seen on Reddit was when a mod leaves a reason for removing a comment. It serves to inform the commenter and also remind the community of the rules.
In a thread such as this one with potential for interesting discussion, it'd be a shame if the mods ended up killing it because the comments got out of hand.
A top stickied comment reminding people of the rules and replying to people who had their comments down voted into oblivion with a reminder might go a long way to having a civil discussion about something like this.
