Read again....

[Ramine]

Hello.....

We have to be smart please..

If you take a look at the help file of the USL R package here:

https://cran.r-project.org/web/packages/usl/usl.pdf

And if you take a look at this R program from Dr Gunther
the author of USL:

http://www.perfdynamics.com/Classes/Materials/USLcalc.r


You will notice that they are using nonlinear regression
to predict scalability, but when you are using
nonlinear regression it can give different coefficients
from those of the polynomial regression that gives the right
coefficients, so this is a limitation in the nonlinear regression
method, so i have implemented my USL programs using polynomial
regression and this gives the right coefficients Alpha and Beta.

So now i think that my USL programs are stable and more robust.

You can download the new version 2.8 of my programs from:

https://sites.google.com/site/aminer68/universal-scalability-law-for-delphi-and-freepascal



Thank you,
Amine Moulay Ramdane.