[R] Data fitting problems...

[nelson -]

Hi all!
  i'm trying to use R to fit a data model... the model describes the
lenght of a process.

I find that it's histogram is of this sort

*
*
*
**
***
***                        ***
******              *********
*******************************

under fisrt "peak" there is 70% of my events and under the second 30%.
Can conclude that my population is bi modal or i'm wrong? The second
peak is good fitted using a normal distribution. The first seems to be
a exponential or weibull distribution. How can i approxximate the
"added" distribution (normal + weibull"? I try nlm or optim but they
don't converge at all...  Another question. how to estimate well
starting points for parameter estimation...

thanks,
  nelson