diff options
author | Eike Rathke <erack@redhat.com> | 2017-06-08 18:13:49 +0200 |
---|---|---|
committer | Eike Rathke <erack@redhat.com> | 2017-06-08 20:40:14 +0200 |
commit | 5ac74f15dcdf85faa2ac1cb7317ea77e31fbcd50 (patch) | |
tree | c837b607c97d44e7e18b9f79aebbd61737c2d2ef /sc | |
parent | 112a82566453fcf85ce3619965ca6e11fa9d29c5 (diff) |
Remove a TODO and explain
Change-Id: Ic648fd2b0a1424aa5ae5d924b29dfd2cb8dddbf9
Diffstat (limited to 'sc')
-rw-r--r-- | sc/inc/scmatrix.hxx | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/sc/inc/scmatrix.hxx b/sc/inc/scmatrix.hxx index 886bbd86840a..553c09f0e416 100644 --- a/sc/inc/scmatrix.hxx +++ b/sc/inc/scmatrix.hxx @@ -133,11 +133,11 @@ public: /** * When adding all numerical matrix elements for a scalar result such as * summation, the interpreter wants to separate the first non-zero value - * with the rest of the summed values. - * - * TODO: Find out if we still need to do this. If not, we can re-write - * ScInterpreter::IterateParameters() to make it simpler and remove this - * struct. + * with the rest of the summed values. This is necessary for better + * numerical stability, unless we sort all by absolute values before + * summing (not really an option) or use another algorithm, e.g. Kahan's + * summation algorithm, + * https://en.wikipedia.org/wiki/Kahan_summation_algorithm */ struct IterateResult { |