[IPOL discuss] possible bug in NL-means (ipol)

[Nicolas Limare]

On Sun, Oct 14, 2012 at 07:04:16PM +0200, Jean-Michel Morel wrote:
> So this is our first bug in a published code in IPOL.

It happened before, in "Algebraic Lens Distortion Model Estimation"
-> http://www.ipol.im/pub/art/2010/ags-alde/

> Since the published code cannot be changed,

A new revised version of the article, including a revised code, can be
published instead of the current version with the approval of the
article editor (Guoshen Yu) or another editor chosen by the editor in
chief. If the changes touch th algorithmic part of the code, a new
reviewer round is needed. The original version will still; be
available from the "History" section of the web page, like we did when
converting from HTML to PDF.

> we will have to publish the erratum and to change the demo code.

One principle of IPOL is that the demo runs exactly the code
published online.

So, we have two solutions:
- publish a revised article, with the updated code and the review and
  approval I explained  in the previous lines, and use this new code
  in the demo, or
- keep the article unchanged, with the same code, and add an editor
  note after the article to inform the readers of the presence of this
  bug in the code and demo.

Moreover, all this should be done with the collaboration of the
authors, because all that is published in their name.

-- 
Nicolas LIMARE - CMLA - ENS Cachan         http://limare.perso.math.cnrs.fr/
IPOL - image processing on line                          http://www.ipol.im/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <https://tools.ipol.im/mailman/archive/discuss/attachments/20121015/4fb14bf0/attachment.pgp>