I am a competitor of Shitbowl, and my product ("Shitdrawer") solves this issue!
However, while it's true that your product is O(1) even in the worst case, when you look closer it's actually O(1) + O(1) even in the best case, as you have to open the Shitdrawer to even get the the most recently used item. Do you have any solution to this?
Shitfloor?
My kids already invented that. If you want theirs, come ask. Includes unlimited eye rolls and muttering.
I would recommend not looking closer!
O(1)+O(1) is just O(11). Not a big problem.
I used to love this product but after a few weeks it started to only make the first 15% of the items available. You then have to use elevated permissions to remove the cached items directly behind these which sometimes results in damage to the items or the operator. Please fix this or I'm going to lower my rating.
Shitdrawer is the Amazon Glacier of Shitbowl. Easy to throw stuff in, generally it stays there very long term.
Shitdrawer also has a more compelling security model, although to be honest shitbowl is more convenient and most people think firewalls are good enough whether or not they have tons of easily opened ports.
It is true that it is theoretically O(n) in the worst case, but in my own production deployments I have tended to find a huge skew towards selecting the most recently inserted couple of items. Here the obvious ocular caching optimization leads to O(1) time retrieval.
That’s right. This system has a high hit rate baked in, based on matching the data structure to the real world usage with a predictable set of content classes. Very elegant.
O(1) implementation is trivial and implemented by most teenagers bedroom floors.
It depends on how good the hash algorithm is. My 2 year old's method tends to have a lot of collisions near the book shelf and the play kitchen, and I'm starting to think it's nondeterministic...
I think the next version, Shitplate, should have this improvement. Space complexity is worse, though.