Software Development Amusement

Oct 27, 2011 16:18


I've just written a class called UnreliableChangeMonitor. See, it's for those times when I want to monitor for changes to a resource, but the monitoring process itself can't be relied upon. The things we do.

Leave a comment

Comments 4

ozisim October 27 2011, 08:29:19 UTC
Yes!
That is what the minions are for!!
- actually one thing that this job lacks is a replenishing supply of Minions... :(
...or Interns. You can make do with Interns if ere isn't room in budget for Minions.

Reply

travisjhall October 27 2011, 11:48:42 UTC

I have to say that interns would be completely useless here, but as an analogy, what you say is surprisingly accurate.

Reply


paradigmshifty October 27 2011, 11:04:54 UTC
There is so much wrong with needing to do that that I don't know where to begin...

Reply

travisjhall October 27 2011, 11:46:49 UTC

Actually, the situation isn't that out there. UnreliableChangeMonitor is an abstract class, so could be subclassed for any number of cases, but the reason I have a need for it is just because the monitoring has to be done across a network. If the TCP connection drops, I've got no idea what is going on with the resource until it can be re-established (it's not available, obviously, but availability isn't the change in question) and need to be able to deal with the lack of knowledge appropriately. Even supposedly stable networks have to be assumed to be flaky when writing distributed systems.

Reply


Leave a comment

Up