On Oct 2, 2015, at 1:20 AM, Jorge Timón via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
On Oct 2, 2015 10:03 AM, "Daniele Pinna via bitcoin-dev" <bitcoin-dev@lists.linuxfoundation.org> wrote:
> should an algorithm that guarantees protection from ASIC/FPGA optimization be found.

This is demonstrably impossible: anything that can be done with software can be done with hardware. This is computer science 101.  And specialized hardware can always be more efficient, at least energy-wise.

I encourage Alex and Dmitry to consider submitting their paper to Ledger, where it will be reviewed objectively and with an open mind.  The authors have motivated their work, framed it in its scholarly context, and made explicit the contributions their paper makes.  Their manuscript, "Asymmetric proof-of-work based on the Generalized Birthday problem," clearly represents a great deal of work by the authors and I commend them for their efforts.  

In the link Adam Back provided, Greg Maxwell mentioned that “it is far from clear that 'memory hardness' is actually a useful goal.”  I agree with this statement; however, regardless of whether memory hardness turns out to be a useful goal in regards to cryptocurrency or not, a paper analyzing memory-hard proof-of-work schemes is certainly useful in helping us to figure that out. 

Best regards,
Peter