namespace bids

timeless time at
Thu Feb 9 18:56:51 UTC 2006

Cahill, Earl wrote:

>Well, my initial message had an open bounty in it, and I have had no
>takers, so good luck with that :)  
>In fact, the reason for my subject line was because I envisioned
>developers submitting bids to get namespaces in memcached.  Then I
>envisioned the changes getting accepted into trunk.  But alas, here we
>are.  No luck on the first vision and no word on whether the second
>vision would have any chance of occurring even if the first vision did.
I suggest pooling your two bounties together, putting out a definite 
number, and then asking someone to implement it for that $ amount. I 
don't think there is a lack of talented coders who could do such a 
thing, but I'm not sure they would be subscribers to this list. They 
might notice an ad on Craig's List or one of those sites that attempt to 
hook up coders with cash.

You might say $500 for a guaranteed O(N) walk-the-keys implementation 
(how hard could that be?) then another $3,000 for an actual O(1) 
clear-a-namespace plus O(NM) do-anything-to-the-namespace 
implementation, where NM=elements in the namespace.

If the code is good and doesn't muck up the memcached internals (most 
obvious is a Makefile option that would enable/disable the feature at 
compile time), I doubt Brad would turn away a patch, but if he did, 
simply fork the project. I could host it (I have a server in a cage for 
my own personal use).


More information about the memcached mailing list