• Visitors can check out the Forum FAQ by clicking this link. You have to register before you can post: click the REGISTER link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. View our Forum Privacy Policy.
  • Want to receive the latest contracting news and advice straight to your inbox? Sign up to the ContractorUK newsletter here. Every sign up will also be entered into a draw to WIN £100 Amazon vouchers!

For the COVID dimwits on this forum - I'm sure you know who you are

Collapse
X
  •  
  • Filter
  • Time
  • Show
Clear All
new posts

    #21
    Originally posted by sasguru View Post
    Cease your unintelligent brain farting and read this paper. Even though it's not peer reviewed, rightly or wrongly, it's the paper that most govt. policy is now based on.
    The govt having backtracked on its "herd immunity" theory a week or so ago.

    https://www.imperial.ac.uk/media/imp...16-03-2020.pdf

    GitHub - mrc-ide/covid-sim

    CovidSim.cpp

    Comment


      #22
      Originally posted by minestrone View Post
      Apparently the code is (or was, before Microsoft and John Carmack recently worked on it) a 15000 line unfathomable, uncommented mess, from which the model is next to impossible to discern, and which produces wildly different results even when fed the same random number seed.

      We, the undersigned software engineers, call for any papers based on this codebase to be immediately retracted

      The authors, including the prof, maintain that the code is intended to be stochastic in that only the results of many runs using different seeds should be used. But if an individual run with the same seed cannot be replicated (due to internal randomness, presumably using other randomisation techniques such as sampling the microsecond system clock) then how can they distinguish model randomness from random code artifacts or outright bugs? Seems a simple but crucial distinction.
      Last edited by OwlHoot; 10 May 2020, 17:40.
      Work in the public sector? Read the IR35 FAQ here

      Comment


        #23
        Originally posted by OwlHoot View Post
        and which produces wildly different results even when fed the same random number seed

        Comment


          #24
          I’d like to see the unit/integration tests of that. I’m guessing there are none.

          Comment


            #25
            Comments are for wimps.
            When the fun stops, STOP.

            Comment


              #26
              Originally posted by jamesbrown View Post

              I’d like to see the unit/integration tests of that. I’m guessing there are none.
              I very much doubt it. But perhaps there are quantum eigentests, whose results are a superposition of all possible test results.
              Work in the public sector? Read the IR35 FAQ here

              Comment


                #27
                For loop at line 61...

                covid-sim/Bitmap.cpp at master . mrc-ide/covid-sim . GitHub

                I'm kind of in the "coding standards don't improve code" camp. You can get a bad programmer to follow all the received wisdom on brackets, loop depth, testability but it wont make better code than a good programmer not following them.

                But this code is truly fooking terrible. I mean horrific. Storing population data in a bitmap with colour for state...

                /* red for infected */
                /* blue for treated */
                /* green for recovered */
                /* grey for just people */

                What is the colour for infected and untreated?

                There needs to be a push to discredit this crap.

                Comment


                  #28
                  Originally posted by minestrone View Post
                  For loop at line 61...

                  covid-sim/Bitmap.cpp at master . mrc-ide/covid-sim . GitHub

                  I'm kind of in the "coding standards don't improve code" camp. You can get a bad programmer to follow all the received wisdom on brackets, loop depth, testability but it wont make better code than a good programmer not following them.

                  But this code is truly fooking terrible. I mean horrific. Storing population data in a bitmap with colour for state...

                  /* red for infected */
                  /* blue for treated */
                  /* green for recovered */
                  /* grey for just people */

                  What is the colour for infected and untreated?

                  There needs to be a push to discredit this crap.
                  Red

                  HTH
                  I am what I drink, and I'm a bitter man

                  Comment

                  Working...
                  X