9CBS reports that Chinese mathematician Wang Xiao Yun, etc., in Crypto 2004, a successful algorithm that succeeds to MD5, GIGIX and Wang Xiong quoted relevant reports in Blog.
MD5 is a summary algorithm, so it is theoretically impossible for the original text from the signature (see below). It is believed that the original text is to be taken from the results of MD5, which is the misunderstanding of the summary algorithm. It is usually applied to digital signatures, used to identify the original origin of the original text - that is, no modification after the signature. If you can use different original texts to produce the same signature, this means that the signature may fail, which can prove that this summary algorithm is insecure.
I looked at Wang Xiaoyun, which is provided by RL, because I am not doing this, so the implementation of the MD5 algorithm itself, and the theory of other people's theories quoted in the article, so I don't understand very well. In this report, the introduction of the MD5 crack is only one and a half. It is not fine to say the specific algorithm, but at the end, two pairs of 1024 original collision examples were attached, and the 512 collisions raised by others in 96 have been attached. A big progress, and the calculation is said to be around an hour.
These are all progress, if this is said to be boiled, it will be a bit self-purple.
Wang Xiaoyun's discovery proves that there is a way to generate collisions, but just as GIGIX