2.4.2001
Pöytäkirja
2.4.2001, 10-12h
Present: Osku, Perttu, Peter, Petteri, Philipp, Toni
1) Testing Plan
Each test case consists of
- Description
(- tested module)
- list of input parameters
- optimal result
- comparison
- auto test/manual test
For the module-tests we have gathered the following ideas, which
should be more specified by the person mentioned for each module:
Config: (Philipp)
- Wrong data
- non-existing file
- wrong numeric params
- float / int
- empty value
- unknown option
- duplicate values
Misc: (Philipp)
- memory allocation
- panic
File: (Perttu)
- Wrong data
- non-existing file
- loading routines
- loading-saving-loading
- loading files from oxxxge (?) sources (couldn't read that)
Gaussian: (Perttu)
- comparison with other programs
- one point test
- border test
- fourier spectrum
- radius > image
- negative radius
- null-radius
Projection: (Petteri)
- clipping
Interpolation: (Petteri)
- comparison with other programs
- negative pixel values
- too big coordinates
Scaling: (Petteri)
- Comparison with other programs
- destination size <= 0
- downscale and then upscale image and compare with original
Rotation: (Petteri)
- 2PI rotation
- alpha-minus-alpha rotation
- trivial images, like lines, circle etc.
- null images
- 3D point
Findmax: (Toni)
- image with known maxima
- box-size <= 0
- unicolor-image
Comparison: (Toni)
- compare same image
- compare with negative
- radius <= 0
- fix-point <= 0, fixpoint > image size
MPI: (Osku)
- sending big data
Osku implements a test-engine.
Ideas to consider (has not to do anything with the testing ?)
- Down scaling
- filtering with Gaussian radius 5 before comparison