Haha ya, it's crazy. I went to Wikipeida.com, clicked on a random link (Ratko_Mladić), and from there it took me 26 clicks of first links to reach the Philosophy page.
When I first read the comic the other day, I went to the page in the comic (Spark Plug) and tried it -- didn't work, I ended up in a cycle between 3 math-related pages. Oops.
This statement shows a gross misunderstanding of the rules.
Meme is quite far down the page. It is well after many links that are neither italicized nor inside parens. The statement whose veracity is being checked says First link (plus some other parameters).
Hah. The hypothesis is false, but the script already has that covered via loop detection... nice :)
panhard
Auverland
Panhard
uh oh... found a loop
panhard -> auverland -> panhard.
So far the longest trail I've found, at 25 steps, was from 'ED209' via pottery, through minerals, states of matter, knowledge, finite sets, mathematics... that's one heck of a wikitrail.
[edit] Scratch that, 'Horst link' is longer via one step, going via transport, commerce, San Juan de Dios Market, Mexico, Romance languages, Precambrian, Chronology, etc...
Genghis Khan -> Borjigin -> Clan -> People -> Human ->
Taxonomy -> Science -> Knowledge -> Fact -> Information -> Sequence -> Mathematics -> Quantity -> Property (philosophy) -> Modern Philosophy -> Philosophy
Mongolian language is in parentheses, so you don't end up in the infinite loop.
So far, the only loop I can see is Panhard, though I'm sure someone will edit the page so that no longer happens. I think the stub pages have a far better chance of defying this law than any other page, though, so it's probably best to start there if you want to find more loops.
Please spend 38.9 seconds trying to understand the very very very simple algorithm that says: Follow the first link that... It doesn't say there are alternate choices. If you went with a different link than the first, it would no longer be the first.
I wonder if we could say everything leads to science. I also reached philosophy via science.
Ratko Mladic (randomly chosen from wikipedia homepage)
Army of Republika Srpska
Military
Use of force
Conflict resolution
Negotiation
Dialogue
Literature
Fiction
Narrative
Latin
Italic languages
Indo-European languages
Language family
Language
Human
Taxonomy
Science
Knowledge
Fact
Information
Sequence
Mathematics
Quantity
Property_(philosophy)
Modern_philosophy
Philosophy
I tried it with Iraq, but the page parsing is slightly wrong. Got this result.
Iraq
Arabic_language
Languages
Human
Precambrian
Eon_(geology)
Chronology
Time
Measurement
Magnitude_(mathematics)
Property_(philosophy)
Modern_philosophy
Philosophy
12 steps to philosophy
But the link to Arabic is in parentheses. The first non-parenthesesed link is to Western Asia.
An article[1] from the Xamuel.com's blog predates the xkcd stripe by a few days and states the sames things but with mathematics instead of philosophy. Also their's some fun with second links.
Two problems. It favors links in the sidebar over article links. I was in "Human" and your script picked out Pre-Cambrian when it should have been Taxonomy.
Second, capitalization matters. It couldn't find "the black keys" but it found "The Black Keys".
It's pretty obvious in hindsight when you think about it. You start with something specific and then get more and more vague until you hit Philosophy.
"Philosophy is the study of general and fundamental problems, such as those connected with existence, knowledge, values, reason, mind, and language." --http://en.wikipedia.org/wiki/Philosophy
Wow, someone just modified Modern Philosophy (now first link is Western Europe), Science and/or Knowledge pages, which makes most previously things tried by me kick into a 25 loop. I'm wondering if it's intentional.
Too easy. The idea was to create an entirely plausible pathway of links that nobody could complain about individually (so they wouldn't get reverted) which nonetheless eventually led to http:en.wikipedia.org/wiki/goatse
The "Quantity" article did not contain the link to Property_(philosophy) before and it was added purely for the sake of this game. See the talk page for more details.
Personally, I think its sad that people are editing wikipedia purely for the sake of making games work out.
Modern Philosophy now goes to Western Europe rather than Philosophy. And since everything seems to go through that, nothing will reach philosophy anymore.
Interesting that the first three i tried took about 17 hops and got funnelled through the "Life" entry.
corndog
sensimilia
halitosis
This took a few less hops and stayed in 'techne':
voip.
Would love to see these searches graphed.
Years ago - one of my favorite sites was Everything2 (still up) - it was run on the slashdot engine. The fun of it was to follow the associated links at the bottom of the entry to see where it would take you.
Slashdot used to link to Everything (back before it was Everything2) as a sort of instant-dictionary for tech terms. Stories would have something like "RSS 2.0(?) and Atom(?) proponents are squaring off..." with the question marks going to the relevant Everything node. Most of the time, the terms wouldn't be defined until the story went live, and then they would be in short order.
It was a neat symbiosis, especially back before Wikipedia existed. However, they were separate sites. Both Slashcode (which runs Slashdot) and the Everything Engine (which runs Everything2) are written in Perl. Everything was never run on Slashcode, though, AFAIK.
Wikipedia:Get_to_Philosophy is not what xkcd is describing, though. That page seems to allow any link to be chosen. xkcd says, "click on the first link in the article text not in parentheses or italics".
For example, Wikipedia:_Get_to_Philosophy provides the example, Optimum "L" filter -> Butterworth filter. But the first link in 'Optimum "L" filter' is "Athanasios Papoulis".
Nice tool :-) It seems to also take words between parentheses though I think. For example on http://en.wikipedia.org/wiki/Iceland I would expect it to follow European not Icelandic.
By extension, you also will always end up at ...
Existence
Sense
Organism
Biology
Natural_science
Science
Knowledge
Fact
Information
Finite_set
Mathematics
Quantity
Property_(philosophy)
Modern_philosophy
Friend of mine was graphing out all of the connections (using Graphviz) and found some interesting patterns. Ultimately you can say the same thing for "Science", "Knowledge", and even "Mathematics".
reiki
spiritual_practice
spirituality
reality
being
eastern_philosophy
islamic_philosophy
islamic_studies
islamic_history
history
george_santayana
madrid
spain
balearic_islands
catalan_language
uh oh... found a loop
music
dance
art_form
senses
perception
eastern_philosophy
islamic_philosophy
islamic_studies
islamic_history
history
george_santayana
madrid
spain
balearic_islands
catalan_language (loop to spain)
war
war
military_history (loop to war)
it seems that the script in the main article isn't using the same pages we are
On-demand page scrape + memoisation is almost certainly a win here. Even if thousands of people are hitting this, a lot will choose some of the same queries (I'm sure Kevin Bacon and xkcd and philosophy are in there a bunch), especially in the tails of the paths (Latin, Mathematics, ...)
I'm quite surprised at how many steps it takes to get from Isaac_Newton to Philosophy (31). I'd expected that would be a quick one. I guess that's just a quirk of the "first link" rule though.
If the draw to "philosophy" is as string at it seems you should end up there without scanning too many articles. You'd need some sort of fixed cap to avoid the few cases where you'd end up scanning thousands of articles, of course.
I find fascinating that if you end up on "Mathematics_education" than you have no chance of getting to "philosophy". (or "physics" is enough, apparently)
the routes should be affected not by selecting the first link on the page but a weighted probability based on the number of clicks each link on the page gets to the other in-subject links in the body.
also sometimes it counts links in sidebars as first, i think that's unintentional.
Why does searching for cannabis just reload the page? I do admit that when it came back and the text entry box was empty I'd already forgotten what I was trying to search for. It wasn't until the next joint when I thought of trying it again that I recalled it had happened previously.
I don't think it works correctly though... it uses "Author" as the first link whereas the first link in the text of the article is "webcomic" ("Author" is in the summary pannel on the side)
The same happens for cities or locations, where there is usually a "Coordinates" in the upper right corner. Still, I did it manually in text bodies, and you eventually end up to Science or Mathematics, which leads to Philosophy.
I posted these stats on a FB status a couple of days ago because of the sudden spike. Now that sudden spike looks small compared to today's sudden spike. Funny how there's different levels of viral, almost akin to different levels of infinite. What would happen if this hit mainstream media?
That was my question, too. There doesn't seem to be a disambiguation that would apply, either. "Greek Language" shouldn't point to "Indo-European languages", it should point to "Greeks". I can't figure out how that went wrong, either, but it is interesting to see a cycle in an acyclic graph. :)
As of 2008, the center of English Wikipedia is the article 2007. From that article, it takes on average 3.45 clicks to get to any of the 2,111,479 articles reachable from it. Disregarding all of the articles that are just lists, years or days of the year, the "real article" closest to the center is United Kingdom, at an average of 3.45 clicks to anywhere else.
>the "real article" closest to the center is United Kingdom
I find this absolutely fascinating. Consider that the reason we're at the centre of the world map is because we had the power to choose to put ourselves there, at the point in history when the maps were being drawn.
Is that imperial influence, long since waned, still echoing across the years, visible in the wikipedia graph?
Actually spell check, has made me a better speller. Of course in the old days when the apple ][ dominated the personal computer space you had to save your word processing file, exit the word processor, switch floppy disks, start the spell checker, switch the disk for the dictionary disk, run against your file, approve/reject suggested changes, save the file, exit the spell checker, switch to the floppy disk with the word processor on it, start the word processor load your file, find the things the spell checker flagged as wrong but couldn't find a correct word for and fix those.
Seriously, it was just easier to get better at spelling.
