HomePhorge

EWS1 works. However, the results slightly differ from the old version. I guess…

Referenced Files
None
Subscribers

Description

EWS1 works. However, the results slightly differ from the old version. I guess it is because of a slightly different abstraction. Int he old version it was was possible to define to which score a boundary belongs. In the new version it is fixed.

Details

Provenance
Maximilian GoetzingerAuthored on May 3 2020, 1:27 PM
goetzingerPushed on May 3 2020, 1:27 PM
Parents
R20:4c7db4e54720: Prepare docs for SA-EWS2 sample data
Branches
Unknown
Tags
Unknown

Event Timeline

Maximilian Goetzinger <maxgot@utu.fi> committed R20:d4e52546dc7b: EWS1 works. However, the results slightly differ from the old version. I guess… (authored by Maximilian Goetzinger <maxgot@utu.fi>).May 3 2020, 1:27 PM

Int he old version it was was possible to define to which score a boundary belongs. In the new version it is fixed.

For the record, the boundaries follow the standard (C++) way of lower and upper bound semantics: the lower bound is inclusive and the upper one is exclusive.

An "exclusive lower bound" can be realized by setting the actual (inclusive) lower bound to the next value.
An "inclusive upper bound" can be realized by setting the actual (exclusive) upper bound to the next value.

The next value is typically

  • Value + 1 for integral types and
  • std::nextafter(Value, INFINITY) (from header cmath) for floating-point types.

Maybe it would be worth taking a note of that in the documentation...