Constantin Pape

> Would we store the data as 3D with the 3rd dimensions stuck having a length of one in dataspace? Yes, I think this is the only good solution for...

I started to implement this in #139.

Not that this also supports platform independent memory mapping :). http://www.boost.org/doc/libs/1_65_1/libs/iostreams/doc/classes/mapped_file.html

Yes, using `xt::adapt` would probably work as well. The solution would probably include passing a 1D C-array as well as the shape (long array + number of dimensons?) to the...

@jakirkham I think your example is pretty much what we will need for read/write. However, we will also need to pass a pointer to the `Dataset`.

Yep, would make sense. I did not write it that way when working on the python bindings, because the arguments in the glue code need to be properly typed, and...

@weilewei I had a very quick look and it looks promising. I don't have much time this week, I will try to have a closer look in the next week....

@weilewei @halehawk I forgot about your efforts in writing a C wrapper and was just reminded of this. Just wanted to say that this is a very valuable contribution, especially...

> Now we integrated z5 into CESM (an earth system model). Great! > We want to test it on clouds with writing out to object storages directly. What exactly do...

@clbarnes > I guess the easiest MVP would be to factor the C++ end to handle the chunking and compression,, communicating with python with lists of `(block_index, bytes)`. Yes, that...