MD5 algorithm not sure

taken from borland.public.delphi.thirdpartytools.general newsgroup (post from Henrick Hellström ).

does it apply to ClrMamePro ?

<<<
Chinese researchers Xiaoyun Wang, Dengguo Feng, Xuejia Lai, Hongbo Yu
have found collisions in several 128-bit digest algorithms including
MD5. The results have been confirmed.

As a result, the MD5 algorithm SHOULD NOT be trusted a component in
digital signatures of any kind. This is in particular relevant for
certificate verification.

The usage of MD5 as an integral part of SSL/TLS is not affected. For
typical setups SSL/TLS mixes SHA-1 and MD5 to form a new digest function
that is as secure as the most secure of the two. Using the MD5 MAC
function in SSL/TLS is however not recommended.

Other digest algorithms have also been broken, but have never been
supported by StreamSec Tools 2.1. These algorithms include SHA-0, MD4,
Haval-128, RipeMD-128.

Read more:
The Chinese paper:
http://eprint.iacr.org/2004/199.pdf
The sci.crypt post with a collision in SHA-0:
cfg007$1h1b$1@io.uvsq.fr
>>>