slifty / internet_noise Goto Github PK

Surprising to me -- but also something we don't want to happen.

Need to modify the search terms to ensure that files will not be included in the results.

Running the algorithm for an extended period of time (about a day and a half now) results in Google getting suspicious:

One of the flaws with this approach is that the random and algorithmically generated queries look random and algorithmically generated.

In other words, Google can tell that it is a robot, and eventually if the wrong keyword combinations appear it may trigger a forced captcha.

Some possible solutions to explore:

1. Find an alternative URL source
2. Find a way to access the feeling lucky button
3. Change the frequency or general quantity of calls
4. ???

I'm using FireFox with AdNauseam activiated, when I click on the "Make some noise" button it opens two tabs and does one Google search in each tab, but then seems to stop. I waited a few minutes, still stopped searching. Then I would press again on "Make some noise" with left mouse click, but still no more noise. Let me know if I can help.

This project shows real life human web traffic which we could possibly incorporate in addition to the random noise: http://eyebrowse.csail.mit.edu/

I'm trying to replicate the bug. It appears if you leave this running over time far more than 5 tabs will end up being used.

At the very least it would be good to be able to detect when this has happened and close the old tab. Ideally an extra tab would never be opened in the first place.

The top 4,000 nouns is a good start, but it's easily filterable. If you can make a system that tracks certain searches, and searches tangential topics, and then topics tangential to those, you can lose surveillance in a sort of fractal search tree. Also, if you can add a system similar to the tl;dr bots, that then picks out key topics, concepts, and words >12th grade reading level from those tl;drs, and then searches those after some random periods of time after loading the page, I think you will have a system that totally fucks internet surveillance, at least for a while.

Bug

I've noticed a few times now that, if the random page that gets loaded is a specific chrome web store app, the algorithm gets stuck on that page.

Fix

Closing the tab results in the algorithm opening a new tab and continuing to load random pages.

If one of the results has invalid SSL certificates, I'm looking at you, Symantic, the cycle on that page stops.

I have no idea how to deal with this. Checking the validity of the SSL certificate before displaying the page would grow this project immensely.

I don't suggest this gets fixed!

This project is awesome, thanks for it.

After some minutes of internet noise, Google has banned my IP address.

What about using duckduckgo instead?

Currently, Internet Noise loads http and https. However, with HTTPS Everywhere enabled, it eventually stops working. For this reason --and also because it is more secure, I think we should only loads HTTPS-enabled sites.

I was wondering if I can host this on a subdomain on my site in case this goes down.

Could you use an iframe or more to make the requests.

I have been running this for about a week and noticed it has downloaded several files. Any way we could prevent this from happening?

to add the ability to change the number of popups
edit (line 65):
$("#noise").click(function() {
var number_of_popups = document.getElementById('popup_number').value;
for(var i=0;i<number_of_popups;i++){
noisify('noise_' + i);
}
and add (line 84):

For whatever reason the "Feeling Lucky" component of a search doesn't always translate into an actual "feeling lucky" press, resulting in a bunch of google searches instead of an actual random page.

I'd love to be able to configure what kind of noise I create. For example, say I want my ISP to think that I'm really into hats. I could then launch a bunch of pages about hats (buying hats, selling hats, hat fashion, etc.)

I don't believe a browser can read arbitrary files by design, but you should be able to generate the HTML file from a flat file of nouns / weights.

Move nouns array to global scope

Basic rationale for this is to make the nouns array easily accessible to console and manually injected scripts. Very simple change, which I am ready to open a PR for given the appropriate permissions. This will allow more technically inclined users to augment the collection of nouns.