Hacker News new | ask | show | jobs
by chrischen 5948 days ago
If I understood correctly, it seems to resemble a search engine in that it takes all the dots that compose the 3D world and search for the ones that need to be displayed to compose the 2D image on the screen at any given viewing angle?

If that is correct, then I think the search engine example is pretty good at explaining it.

1 comments

Graphics cards already have this functionality, it's nothing new.
Not arbitrarily. Make sure it's nothing new before you say it is. How complicated is the predicate they're searching?
I can say it is nothing new. I have written and seen many patents that describe systems that do that.

As far as the "predicate they are searching" the video mentioned nothing about the predicate other than it is for things that are visible, and that is nothing new.

They could just be being glib. Clipping and LOD aren't the same thing as one another. Consider the 2-D case of sampling a line (uncountably many points) for a raster display. Reimagining our favorite line drawing algorithms as search problems requires a lot of work to figure out how to best search the space. I hope these guys are doing something like that.