Duplication of common grid in storage

[astamm]

I think the structures devised in this package could become the reference structure for functional data. My only issue is that although it requires that all curves of a sample be evaluated on the same grid (be it one- or multi-dimensional), a copy of that grid is then stored in each dimension of a functional data set with multi-dimensional codomain (e.g. in multiFunData objects).

[ClaraHapp]

Thanks a lot for your comment. As long as the data has a common grid (funData and multiFunData classes), the grid is stored once in the @argvals slot, for each funData object. irregFunData objects are different.

You are right, if a multiFunData object has multiple elements, m[[1]], ..., m[[M]] and the all live on the same domain, the grid is stored M times. However, I don't see it as a big issue, as first of all, the implementation allows for way more complex data with different grids (even different dimensional grids). Secondly, even if your setting is such that the grid is shared, the storage needed for the grid is effectively the storage of one observation, and typically you would have many observations, so the grid shouldn't weight too much.