1) Use clock res to set error range, change error to confidence.
2) Allow user interface to change verbosity of output.
3) Fix handling of fragments.
4) Regression with pp?
5) Split off tailgating code
6) Fix tailgating.
7) Document tailgating.
8) Web interface
9) Speedup
10) Sys V scripts to automatically run nettimer as server.
11) Validate new filtering on traces.
12) Examine filtering using one-way delay for RBPP, RT delay for SBPP.
13) Make time-based window.
14) Add test suite.
15) Test PBF w/ 4f/1b connections.
16) Report latency
17) Report loss
18) Run on publicly available traces.
19) Multi-channel effects
20) Available bandwidth
21) Why does this give poor results:
nettimer/src/nettimer -n -r traces/20001204_110701/Ethernet_10Mbs/long/140.247.60.220_1_0_pcap -w nettimer_171.64.69.45_140.247.60.220_0_1_1_node_trace -u "" --maximum_flow_length 10000
22) Using SBPP on assymmetric link=>deduce reverse path?
23) Set up mailing list
24) Send to pittsburgh people
25) Bus-based Ethernet messing up pp and tailgating. Consider using larger bunches.

