This project has moved and is read-only. For the latest updates, please go here.

Breaking Changes

  • Move to github. I'm actually perfectly happy with CodePlex, but the writing is on the wall and github has more eyeballs.
  • Target .NET Core.
  • Add an IntegerPartition class that supports both views of a partition (1+1+ 3 and 2 * 1 + 1 * 3), and useful properties and functions such as Rank and Conjugate.
  • Change default EvaluationSettings so default RelativePrecision and AbsolutePrecision are -1, and individual method change -1 values to their defaults. When first created, EvaluationSettings was only for 1D Integrate, and the defaults are appropriate for that method. If you wanted to change just one property, e.g. EvaluationBudget, you would new up an EvaluationSettings, change that property, pass it to Integrate, and the other properties would still have their default values. But now EvaluationSettings is used for all sorts of other methods that have other defaults. If you want to change just EvaluationBudget for them, you need to new up an EvaluationSettings, change that property, and then... unless you want to use Integration's defaults for the other properties, you need to change them too. And even figuring out what they are isn't trivial. So instead we will set them all to -1 and let the individual methods override -1s.

Additional Features

  • Add specific FitResults for all fit methods. These objects should give (i) direct access to named parameters, so users don't need to remember which parameter has which index, (ii) residuals, if applicable, (iii) ANOVA analysis, if applicable.
  • Add specific EvaluationSettings for all function analysis methods. These objects should expose listeners that give intermediate results and additional settings specific to the kind of analysis.
  • More time series functionality, including full ARIMA fits and population power spectra estimates with uncertainties.
  • Submatrices. For examle: given say a 100 X 200 RectangularMatrix, get a 20 X 10 RectangularMatrix that addresses the same storage as the 10...30 X 50...60 elements of the original matrix.

Last edited Dec 27, 2016 at 4:12 AM by ichbin, version 5


No comments yet.