Re: [Taler] optimal coin spending

[Christian Grothoff]

On 12/20/2015 02:27 PM, Jeff Burdges wrote:
> On Sun, 2015-12-20 at 10:16 +0100, Christian Grothoff wrote:
>> On 12/20/2015 12:00 AM, Jeff Burdges wrote:
>>> As I mention in my first mail, if we use a greedy algorithm, then
>>> we
>>> should run it separately for each mint's denomination set.  It's
>>> easy
>>> to lose the property of being well suited for the greedy algorithm
>>> by
>>> taking the union of two mint's denomination sets. 
>>
>> Sure, I was expecting us to run the algo for each withdraw operation
>> anyway.
> 
> We're talking about purchase algorithms, not withdrawal algorithms, but
> I suppose you just miss typed.  Yes, we'd need to run for each purchase
> operation regardless, but..

Indeed, I did mistype, albeit there is a similar problem for withdrawal.

> Anonymity suffers when spending from multiple mints simultaneously too.

Actually, our current /pay protocol requires all coins to be from the
same mint.

signature.asc
Description: OpenPGP digital signature