[gclist] Off topic messages.

[Charles Fiterman]

At 04:36 PM 12/10/96 -0500, you wrote:
>Apologies to the list for the off-topic message about qsort - 
>it was intended for Charles Fiterman alone.

Some people take a very strict view of topic restrictions I
think a loose view is more usefull. We weren't discussing
evolution v.s. creationism we were discussing a very related
problem and I often find that useful.

There are two common threads between qsort and garbage collection.

First my latest collector uses a sort and I can even see how the
merging decks case can come up easily and for large sets. So 
understanding that case is nessisary to prevent twelve hour 
collector delays.

Second we have a common problem an algorithm that can behave very
differently for different data sets and how to optomize it. Radioactive
decay model in collection is equivlent to random data model for 
sorting. How do we tell what to optomize for? Maybe the success
of my theory don't allow pathological cases will be usefull in
collector strategy.


Charles Fiterman	Geodesic Systems
Phone 312-728-7196	4745 N. Ravenswood Suite 111
Fax   312-728-6096	Chicago Il 60015
cef@geodesic.com

A computer language without garbage collection
  is like a city without garbage collection.