I have an idea about a program. It will randomize previously written Articles by using a thesaurus and perhaps randomize every 3-4 words. This will trick google into thinking that it is original content. Do you think this is a good idea? Has it already been done?
It has to have been done at some point. Major hangup would be getting a usable thesaurus. Also could/would have the potential to create total gibberish. ethically speaking it would be a gray area. Some against, some for etc. Hey, wouldn't it create a legal twist in asking who owns the content after the transformation.
Nice but it pretty much mangled my sample text. Original: "Flabbergasted" NASA scientists said on Thursday that Martian soil appeared to contain the requirements to support life, although more work would be needed to prove it. Scientists working on the Phoenix Mars Lander mission, which has already found ice on the planet, said preliminary analysis by the lander's instruments on a sample of soil scooped up by the spacecraft's robotic arm had shown it to be much more alkaline than expected. "We basically have found what appears to be the requirements, the nutrients, to support life whether past present or future," Sam Kounaves, the lead investigator for the wet chemistry laboratory on Phoenix, told journalists. Translated: "Flabbergasted" NASA scientists said on weekday that Martian grime appeared to include the requirements to hold life, though more impact would be necessary to establish it. Scientists employed on the constellation Mars town mission, which has already institute cover on the planet, said origin psychotherapy by the lander's instruments on a distribution of grime scooped up by the spacecraft's robotic limb had shown it to be such more alcalescent than expected. "We essentially hit institute what appears to be the requirements, the nutrients, to hold chronicle whether time inform or future," Sam Kounaves, the advance policeman for the dewy alchemy work on Phoenix, told journalists. Why did you make it add words?
How about "Mad Libs"? http://www.amazon.com/Best-Mad-Libs-Roger-Price/dp/0843126981/ref=pd_bbs_sr_3?ie=UTF8&s=books&qid=1214592106&sr=8-3
I think your idea is fundamentally flawed in "trying to trick Google". What you need to make is a program that will create useful article alternatives. That way marketers will be able to test many variations of unique articles to find the one that performs the best. I am guessing that the word substitution should be checked by a human to ensure that the substituted word still has the same meaning otherwise the article will simply be transformed into gibberish!
at this point randompaper.com is useful for rewriting simply to save you time rewriting an article. IMO you should just use the tool then correct everything by hand. It isn't perfect, but its still faster then rewriting the whole article by hand
The problem with this stuff is that IT DOESNT FOOL BIG G! I recently read a very interesting article on shingling algorithms which it is widely believed that Google uses for detecting dupe content and it said that Google can determine the context as well as the content ie simply mixing up words and simple random substitutions will not yield the desired results. Just my .02 cents but it takes a very good (or very brave) programmer to take on the might of Google with their vast resources and outsmart them. Oh, and this has already been done several times ..sorry
I think an ajax thesaurus where you can pick your own words would be better. I tried your auto-spinner and this was the first line: ORIGINAL: "Lenders love to run ads on TV and radio bragging about their super-low rates." SPUN: "Lenders [CENSORED] to separate ads on TV and broadcasting crowing most their super-low rates." Yes, the CENSORED part was the F word...however, it is 55.4% unique according to Dupecop.
Swapping synonyms is not the best approach to content spinning. While the results may either be effective or confusing to a human reader, it is, ironically, the same machine that you are trying to fool that will inevitably call you on it. For instance, compare the following to the previous paragraph: Exchanging similar words is not the best way to go about rewriting. While the results may either be spot-on or fuzzy to a man parser, it is, ironically, the same tool that you are attempting to trick that will have no choice but to message you on it. Both samples still contain the common text: is not the best to . While the results may either be or to a , it is, ironically, the same that you are to that will you on it. To reiterate and expand upon what was mentioned in the previous posts above, even if you change the order of the sentences, the pattern will still be there when inspected at the sentence-level. Even if it is unrecognizable to humans, machines can find the probability of a word or phrase appearing based on its predecessor(s). If you start rearranging words within the sentence without an extensive set of grammar algos, you run the risk of breaking the sentence structure. Once that happens, you are at the point where you may as well run the text through a translator, then back again, because that is essentially what you will be getting. You must also take other factors into consideration, such as newlines, tabs, letter casing and carriage returns. While the average human may not notice these things, it is certain that there are machines out there that do.
I suggest spending half an hour writing your own articles, do a brain storm for each section intro, middle, conclusion and then just write 5 sentences for each section. Most tools like this dont work very well.