temporal ls

File Browser (Nautilus 2.8.1) always takes a long time to display
/usr/share or any other dir with lots of files. Needs to take advantage
of temporal locatilty. The dir can be monitored for updates.
The temporal locality idea can be extended beyond Nautilus listings to
all results of repeated commands or software code segments. Current
OS's apply it to paging but the results of code segments can be stored
as result pages and then go into the sameor similar temporal locality
algorithms. The results will be idempotent subject to changes to an
input set. E.g. ls on /usr/share is idempotent over time as long as a
new file is not added to /usr/share (or other CRUD - del, update).
The idea may be extended to idle processing time whereas the common
operations are run based on file-change-notification or other
algorithms and the results are cached. Furthermore, the decision of
what to run can be heuristic or based on statistical analysis of what
the user has done on his or her computer (e.g. always runs ll or ls -F
or ls -la for instance). The analysis could go as far as to cluster
commands (and/or code segments) such that clustering implies temporal
locality for the next or following commands (and/or segments of code).
Spatial locality may be leveraged as well and could also be used
loosely with the clustering as small commands such as ls and cp may
cluster together and there for be considered in the same "space" just
as code segments (blocks) withing a program could be spatially local.
TimJowers
P.s> Guess shouldexist.org is down for the count? There were a ton of
great ideas in there.

.