| Ticket UUID: | 2818958 | |||
| Title: | sizing of vectors | |||
| Type: | Bug | Version: | None | |
| Submitter: | dkf | Created on: | 2009-07-09 10:45:13 | |
| Subsystem: | math :: linearalgebra | Assigned To: | arjenmarkus | |
| Priority: | 5 Medium | Severity: | ||
| Status: | Closed | Last Modified: | 2009-07-20 17:47:25 | |
| Resolution: | Fixed | Closed By: | arjenmarkus | |
| Closed on: | 2009-07-20 10:47:25 | |||
| Description: |
I was working (on http://rosettacode.org/wiki/Multiple_Regression#Tcl) with a collection of vectors which were both row-vectors and column-vectors (one being a transposition of the other) and I was trying to multiply them together to get a matrix. This was causing errors; apparently the 'shape' command from this package goes completely to pieces when faced with a list of length 1 (making everything else in the package fail miserably too). A scalar must be a 1x1 matrix. A vector must be a 1xN or Nx1 matrix. If these aren't true, many innocent algorithms die in a fire (and Carl Friedrich Gauss kills a kitten). | |||
| User Comments: |
arjenmarkus added on 2009-07-20 17:47:25:
allow_comments - 1 Fixed this issue - scalars were treated separately so that [dim] could report the correct dimension, but that special casing was used for row vecotrs as well. | |||
Home
Blog
Docs
Downloads
Tickets
Timeline
Branches
Tags
Wiki