You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently there is no example showcasing the use of bitwise data. Technically this datatype is applicable to both DPP and DPS.
It is also technically feasible for bitwise DPS to be utilised as group definition, rather than a list of indices. This would only reduce requisite space for groups that select more than 1/32 of streamlines, and unlike a list of indices it would mean that the number of elements cannot be inferred without reading the data, but it's nevertheless worth discussing the prospect, and explicitly rejecting if that's the consensus.
The text was updated successfully, but these errors were encountered:
That's an interesting point that brings back an idea from the original thread (except it was a DPS in uint8 and required a LUT for meaning).
We just have to mention what would be the extension for binary data type? just .bool?
It is a simple idea that allows for quite a lot of liberty. I don't see how it could be misused, I would add it immediately unless anyone has a counter-argument.
We just have to mention what would be the extension for binary data type? just .bool?
I'd probably advocate for .bit rather than .bool. Over and above being consistent with existing MRtrix3 parlance, I think it's a more direct reference to storage, whereas "boolean" to me conveys either an internal interpretation of data or the logic surrounding such.
Currently there is no example showcasing the use of bitwise data. Technically this datatype is applicable to both DPP and DPS.
It is also technically feasible for bitwise DPS to be utilised as group definition, rather than a list of indices. This would only reduce requisite space for groups that select more than 1/32 of streamlines, and unlike a list of indices it would mean that the number of elements cannot be inferred without reading the data, but it's nevertheless worth discussing the prospect, and explicitly rejecting if that's the consensus.
The text was updated successfully, but these errors were encountered: