Re: SLOC of BRAT

From: Konstantin Olchanski (olchansk@sam.triumf.ca)
Date: Thu Jan 03 2002 - 17:33:02 EST

  • Next message: Christian Holm Christensen: "BRAT changes - RPMs possible"

    On Thu, Jan 03, 2002 at 02:17:44PM +0100, Christian Holm Christensen wrote:
    > 
    > I think the price of BRAT is largely overestimated, ...
    
    Thanks for the pointer. I ran it on the BRAHMS DAQ and the E852 code trees,
    as well as the TWIST code (useless- it does not know about F90),
    and I agree with you, the model seems to overestimate the development costs.
    
    I am not an expert on the models used, but I suspect they use
    very pessimistic estimates of programmer productivity.
    
    In other words, we are more productive then average *and* underpaid
    at the same time.
    
    > ... we do other stuff than just writting code for BRAT, like teaching,
    > taking courses, administration, giving talks, and so on, which is not
    > part of the model used, ...
    
    I suspect their models do include the duty factor effects: commercial
    programmers do have to sit on silly meetings, take vacations and go
    home at 5pm. (Here is a ludicrous example- on the  U of BC campus,
    employee/staff parking is only free 9am-5pm. Want to
    work late? Dish out $3.25).
    
    -- 
    Konstantin Olchanski
    Email: olchansk@triumf.ca
    Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
    



    This archive was generated by hypermail 2b30 : Thu Jan 03 2002 - 17:33:41 EST