| TheFLY |
02-06-2003 06:36 PM |
ATTN Coders: $48000 to code a good hitbot to cheat anyone.
To be coded in C++
This hitbot will need to do the following to cheat anyone:
- RANDOMLY DISTRIBUTE CLICKS -- use probability (standard deviation and all that) to distribute the clicks -- (Data imported from a click pattern utility). A click based targeting interface will be needed to map out many different click patterns -- the click pattern information will be recorded and stored in a database. Different factors will include -- time between clicks, keyword matching, image loading time, etc. The targeting interface will generate script code that can be edited for more complex spidering of sites based on dynamic variables that can be fed in depending on the campaign...
- MULTI-THREADED TREE ENGINE -- each click from the bot will initiate a hierarchy of clicks. So "deep clicking" would be a formula for two levels deep of clicking would look like this (number of links on the page) * (number of links on the page) * (number of links on page) -- this would be fed into an database of links for each cheating campaign. Then using probability (which would be based on a % based on link distance from the height of the site -- along w/ color contrast, link visibility (based on font for example) decisions. This will allow the bot to fool guys that think they are smart by having "bot traps" that are just little links that the surfer can't see -- where the bot will click by accident.
The engine will need a timing structure. As an original click path is mapped out for each batch of clicks for the main engine, it will need to be stored in a timing structure. The main engine will take commands to fire off clicks based on the hierarchy of sites during the map. A dynamic list will be generated hourly -- each hour, new click maps will be calculated. The result is like a treasure map of times when the engine needs to generate the click's HTTP information.
- RANDOMIZE HTTP/BROWSER info -- country, language, etc. -- cgi information based on industry standards for normal distribution (we bought this info from Hitbox/Websidestory :) -- also factor in a randomness to this ratio -- as LARGE amounts of clicks will be generated. We are going to run this thing on parallel processor machines on T1's all over the country.
- CLICK PRIORITY UTILITY -- Utility to assist in randomizing click patterns. Every site in the database should have a record. Each record will allow different click patterns based on factors like banner placement, category of link (priority levels will be fed into clicking engine)
- EMULATE JAVASCRIPT. We won't be able to support java yet due to the demands on vast # of clicks that will be spidered... (Think if 1 site has 100 links and each site that it has has 100 links, etc. -- we're talking a lot of clicks!) You will need to return cookie information when asked for -- this is easy. To emulate javascript we have the assistance of a Microsoft geek. We just need your interface and engine to integrate by sending hook notices to the javascript emulator.
You have six months to code this -- you get paid weekly. Submit a resume link to me at ICQ#1622420
|
