[OpenWalnut-Dev] WGridRegular3D

Mathias Goldau math at informatik.uni-leipzig.de
Thu Oct 28 18:13:31 CEST 2010


Am 28.10.2010 17:12, Andre Reichenbach wrote:
> Hi,
> 
> I think our grid has become quite confusing. It seems there are several
> matrices and other data fields that describe the orientation of the
> grid. Also there are some interesting things like getOffset*() returning
> the norm of m_direction* instead of m_offset*, and so on.
> 
> I'm thinking about reworking the grid with the following things in mind:
> 
> - avoiding data duplication (i.e. there is only one orientation matrix)
> - the user should not need to know about the semantics of internal data
> 
> 
>   (this leads to streamlining constructors for example, different
> constructors
>    seem to set different values for several internal data fields, for
> example
>    unit axis vectors for the directions, which in turn leads to
> getOffset*()
>    ignoring the actual offsets as those are represented in the
> transformation matrix... )
> 
> Also, shouldn't grids - as they are parts of datasets - be const? I've
> seen several members that
> suggest transformations on an existing grids are possible.
> 
> I do not want to rush this, as there might be a lot of things that may
> get broken in the process.
> However, we should discuss this at the next meeting, which could be Mon
> 7th, for example.

great, but the 7th of Nov is a Sunday, I guess. What about the 8'th then?

Mathias

-- 
Institut für Informatik
Universität Leipzig
Johannisgasse 26, 04103 Leipzig
Phone: +493419732283


More information about the OpenWalnut-Dev mailing list