commonsense team mailing list archive

[Rob Speer <rspeer@xxxxxxx>]

Public bug reported:

Intuitively, when you normalize a vector, you should get back another
vector of unit length, or something that acts like it.

Instead, Divisi tries to normalize it the way it would normalize
something bigger -- chopping it up into slices, finding the norm of each
slice, and then giving you a view that divides those slices by their
norm.

Except these slices are *scalars*, so if this worked it would give you
something like [1.0, 1.0, 1.0, ...]. And it gets an IndexError anyway.

I think that .normalized() on a vector should normalize by the magnitude
of that single vector (whether it uses a view to do so or not).

** Affects: divisi
     Importance: Medium
         Status: Confirmed

** Changed in: divisi
       Status: New => Confirmed

** Changed in: divisi
   Importance: Undecided => Medium

-- 
vector.normalized() doesn't do what you expect
https://bugs.launchpad.net/bugs/412731
You received this bug notification because you are a member of
Commonsense Computing, which is the registrant for Divisi.

Status in Divisi: Confirmed

Bug description:
Intuitively, when you normalize a vector, you should get back another vector of unit length, or something that acts like it.

Instead, Divisi tries to normalize it the way it would normalize something bigger -- chopping it up into slices, finding the norm of each slice, and then giving you a view that divides those slices by their norm.

Except these slices are *scalars*, so if this worked it would give you something like [1.0, 1.0, 1.0, ...]. And it gets an IndexError anyway.

I think that .normalized() on a vector should normalize by the magnitude of that single vector (whether it uses a view to do so or not).