Ns中如何设置错误模型的单位？

这个：
set em12 [new ErrorModel]
$em12 unit pkt
$em12 set rate_ 0.0012
$n12 interface-errormodel $em12
这个是不是说em12的错误率是报错误率，大小是0.0012的意思？
我如果要设置错误率是误码率，应该写
$em12 unit 什么呢？

找到了
#
# There are three levels to error generation.
# 1. Single State
#       rate_:  uniform error rate in pkt or byte
# 2. Two State
#       error-free (0) and error (1) states
#       each state has a ranvar determining the length each state
# 3. Multi-State:  extending Two-State in OTcl
#       each state has a ranvar determining the length each state
#       a matrix specifying transition probabilities
#
#    Patched by Jianping Pan (jpan@bbcr.uwaterloo.ca)
#
# Each state is an error model (which could be 1-state or multi-state),
# In addtion, the error model has a matrix of transition probabilities,
# and a start state for the model.  These usually corresond to
# homogeneous Markov chains, but are not restricted to them, because it
# is possible to change the transition probabilities on the fly and
# depending on past history, if you so desire.  Multi-state error models
# reside entirely in Tcl and aren't split between C and Tcl.  One-state
# error models are split objects and ErrorModel is derived from
# Connector.  As an example, a 2-state Markov error model is built-in,
# as ErrorModel/MultiState/TwoStateMarkov Finally, an error *module*
# contains a classifier, a set of dynamically-added ErrorModels, and
# enough plumbing to construct flow-based Errors.
#
在ns-errmodel.tcl文件中

Pan Jianping

这是什么？
潘建平？
潘建平