[GiNaC-devel] Another bug in clifford.cpp

[Chris Dams]

Dear Vladimir,

On Wed, 4 Jan 2006, Vladimir Kisil wrote:

>   Thus indexed object is chosen as a standard storage for a metric with
>   its first component to hold actual metric (e.g. matrix or
>   tensor). Indices are indeed somehow superficial but having some
>   default values for them seems to be useful for example in
>   same_metric(). 

However, also in same_metric, get_metric is called and that replaces the
indices. So maybe there is an example in which the indices serve a purpose
but it does not seem to be in the same_metric function. The alternative
that would seem to be more logic is to store the thing that should take
the indices. For instance a matrix or a minkmetric. In that case
get_metric would not have to substitute indices but could merely return
indexed(metric, i, j). Is the purpose of having the metric inside an
indexed object to also store the symmetry of the metric?

Best wishes,
Chris