Talk:2407: Depth and Breadth

Explain xkcd: It's 'cause you're dumb.
Revision as of 00:32, 5 January 2021 by 141.101.98.68 (talk)
Jump to: navigation, search

where did the quality go 172.69.34.24 19:34, 4 January 2021 (UTC)

I noticed this too. As discussed at User:DgbrtBOT there are two sizes of each comic. The default (smaller) size of 2407 looks much worse than the original, which you can find at https://imgs.xkcd.com/comics/depth_and_breadth_2x.png I suggest we use the larger version for this comic. Alchemistmatt (talk) 20:18, 4 January 2021 (UTC)
I tried to upload the higher quality PNG but I do not have permission; we'll have to wait for an editor to provide their opinion. Alchemistmatt (talk) 20:35, 4 January 2021 (UTC)

It would appear that the first version of the picture of this day's cartoon presents artifacts due to an unusual export method. The image seems to have been exported using the 'nearest neighbor' resampling method, which would explain the jaggy edges. Usually, the images appear to be exported using bilinear downsampling from an white-grey-black original, resulting in a published version with a larger color palette. 162.158.111.161 20:17, 4 January 2021 (UTC)

Randall has uploaded a new image: [1], which I uploaded to explainxkcd. Natg19 (talk) 21:48, 4 January 2021 (UTC)

In the breadth-first, for the second node on the right, the right branch is searched first, while everywhere else, the left branch is. And in deadth-first, the nodes are searched multiple times (e.g. left-most node of layer 3 is search 3 times, assuming a search is at the end of a continuous line). Alternatively, maybe the search goes up first sometimes (it's not actually clear when a node is being looked at), but that doesn't explain the order of the left-most node of layer 2. 172.68.142.201 22:15, 4 January 2021 (UTC)

I noticed that. With perhaps two reasons: 1) aesthetics - drawn to go to nodes .2, .2.1, (.2), .2.2, then .1.1.1 (via .1.1, crossing .2.1, .1.2) would look a bit worse than this (crossing just .1.2); or 2) there's no absolute sorting order vs choice, it's just chance (or aesthetic choice) that .1s take priority over .2s in every other case - 3 out of the four choices is well within explorative chance.
I favour the latter (with maybe an aesthetic bias) as often when I run a tree-searching algorithm I like to randomly splice the next option out of the list of options (rather than run from first to last or last to first) where I am not aware of any advantageous link (maybe in ruling out 'dead' branches early to prune off useless branches early) and thus whatever natural sort-order the structure imposes would create biases.
Alternately, if continued it would definitely prioritise .2s down every .2(-dominant) branch, for a nicely symmetric 'wide-breadth first' pattern (.2.2.2 over .2.2.1, etc) for a pattern only visible once continued beyond the step currently shown. The root choice cannot be anything other than symmetry-breaking, but could as easily be a coinflip. 141.101.98.68 00:32, 5 January 2021 (UTC)

The top two drawings for "depth" and "breadth" are legitimate methods of listing out a tree structure. The next two drawings substitute the "d" and "br" from "depth" and "breadth" to get "brepth" and "deadth". The fifth drawing removes the "th" from "breadth" to get "bread". And the title text substitutes the "p" from "depth" with an "a" to get "death". Rtanenbaum (talk) 22:32, 4 January 2021 (UTC)