The algorithm for comparison is simple, text field for input of checked article. Then the article will be separated by comas to sentences and these sentences will be checked by google for similarity. The similarity factor will be counted in % (number of similar sentences(words) / number of all sentences (words) check).
The design will have three pages:
1. Input page - Input text field for the article and submitt button + choose of searcher, yahoo or google
2. Configuration page - setting of the algorithm configuration
3. Output page - Showing of links with similarities on google or yahoo + counted % of the similarity
Sample of the same code and logic used can be seen on [url removed, login to view] (free account). Also the counted percantage of similarity should be as close as possible to the numbers shown by the [url removed, login to view] (another plagiarism detection site can be seen on [url removed, login to view], but with not good result page)