death in CS for half a decade .
I think you mean century.
The bulk of the sort and search work was pretty well solved by then end of the 1960s.
However due to the massive amounts of data being crunched by the likes of google these algorithms have undergone a bit a renaissance.
In the dawn of computing the data/ram ratio was massive. We didn't have gobs of either but RAM was expensive. In the late 80's through to the early 90's the ratio shifted dramatically. Outside of certain scientific computing domains your typical large data set rarely exceeded ram by an order of magnitude. Heck sorting 30k employees by name while not blinding fast with a poor choice of algorithm is still plenty fast on whatever flavor of qsort shipped with your standard library ... and since the entire problem space is in ram there is no reason to know anything more about the details.
Now when you can only hold 1% or less of your primary keys in ram ... then sorting it becomes interesting again, sort of, because you have to back digging up those old sorting schemes or re-invent them.