I could be wrong about this. For instance, http://dl.acm.org/citation.cfm?id=2148250.2148477 http://dl.acm.org/citation.cfm?id=2148250.2148477 https://pdfs.sem

I could be wrong about this. For instance, http://dl.acm.org/citation.cfm?id=2148250.2148477 http://dl.acm.org/citation.cfm?id=2148250.2148477 https://pdfs.semanticscholar.org/868b/443aec1e87951c8a896b1298c1fa9cdd1e1e.pdf https://pdfs.semanticscholar.org/868b/443aec1e87951c8a896b12... "Improved preimage attack on one-block MD4" is a 2012 publication improving the preimage attack on MD4 from 2^107 to 2^95. So 17 years after a practical collision attack on MD4, academics ARE still working on it as a cryptographic primitive, but despite that they have NOT found a practical preimage attack on it.