I have been doing that for a couple months for myself with my web server and decided others may want it.  I am a week or two away from getting minergain dot com running (free forever) to do kano plus a few others my volunteers work with.  It will also monitor your miners and alert you if any degrade their performance (bad card, internet, etc).  You'll be able to set email or text alert.  If you sign up, you'll get notified when its ready.

Fortunately I did get to it before you deleted it (found in the Mining Simulator of https://en.bitcoin.it/wiki/Mining).  It is a wonderful sample.  So I see that the date is included in the computation.  If that is accurate, then there will always be an answer.  So that also means that nonce does not make a huge difference and definitely there is no reason to retry skipped ones later on.

So is the Bits something that the pool changes or is that the difficulty level?

Thanks Kano.  Yes, this is all super deep stuff and thanks for your time explaining it. I've only been into mining / crypto for about 2 months, so super green on it all.  So based on your comment, the miner does not even complete the work that is assigned to them?  That must make things insane for the server software.  So does the remaining work get distributed back out to the miners again somehow or is it just ignored?

The basic explanations I have seen of mining on the web are to compute the sha256 of the block plus the number 1, then 2, then 3.  Firstly, I am not certain this is even close to what truly is happening.  If it is somewhat close, then I would assume that each miner is given the block info and a range of numbers to work on sequentially.  I have no idea of the starting and ending range for work (and I am sure it varies by miner), but lets say the first miner's work is 1 through 1000, then my thought is rather than sending 1001 to 2000 to the second miner, the pool would send a random unprocessed starting point.  For example, 1 trillion to 1 trillion + 1000.  The third miner may be 1 billion to 1 billion + 1000, etc.  Beyond that, each pool server would need to have their own range since the servers cannot communicate around the world fast enough.  On each server, a memory cache would be filled in with the next few hundred block ranges which would get mapped into a local database in the background after being assigned and repopulate more random blocks for future requests.  Of course that all gets even more insane since every miner has an individual difficulty but perhaps that could be figured in. 

Eventually someone will be assigned 2001 to 3000 when it randomly comes up.  Obviously if the block answer was found in the 2001 to 3000 work group, then this method dramatically slows what a sequential operation would do.  Anyway, that was my thought - but it is highly possible that is already happening on every pool in the world.  As I said, I'm green on this.  Thanks again and it was a pleasure to chat with you and be a part of your pool for most of my short mining life.

Just curious if it is possible for a block to not be able to generate an answer (go through every possibility)?  I wonder if anyone has analyzed how many possible answers there are for the various difficulties?

Also a tech question / idea for Kano...  Are shares distributed sequentially or randomly to miners around the world?  It seems that randomly picking shares may improve our luck (may make it worse too).  I am a 30+ year web / c++ developer that has combined dedicated servers into a worldwide distributed digital signage solution.  I have a concept on how to code random share distribution to deal with the trillions of shares.  I am tempted to code up my own pool but have no idea where to start and have paying customers to keep happy.  Kano, if this is distributed sequentially, can we chat?