Hi all,
I think that some lines in pareto.cc are quite strange: (i) burstlen_ is used for p1_ calculation. In that way ontime_ will be lengthened for tens or hundreds of times and once it's on, it goes ON and ON. After replacing burstlen_ w/ ontime_, everything is more reasonable. (ii) as rem_ gets to 0, an idle period is calculated and set to t. In pareto.cc that idle period was added to t instead of setting to t. I have no idea why t must be (interval_ + idle period). I didn't find old archives discussing this stuff. Any further comment and discussion are welcomed. Merlin Jiang |