[IPOL discuss] Corrections of erratas: the MathReviews Policy

[Nicolas Limare]

Hi,

> Errors in an IPOL article or in the IPOL code that alter the
> understanding of the article, or the functionality and result of its
> code must obviously be corrected.
> 
> This dynamic errata correction is a new feature, allowed by
> electronic content. I see no inconvenience if a strict track is kept
> of the corrections and of their dates in the article and code, and
> this is what I have proposed.

I was investigating the possibilities of a better indexation of IPOL
in academic paper databases. One major database for math research is
MathReviews/MathScinet[1] by the Americam Math society. MathReviews
has a very clear policy[2] on revisions, reproduced hereafter. It
says, in brief,
* research depends on reliable content and persistence
* mathematics mathematics crucially depends on the integrity of the
  literature
* once an article is published, itr should only be revised by adding a
  new dated version
* all versions must persist indefinitely
* this applies to any modification of the article: text, title,
  reference, and anciliary information.

[1]http://www.ams.org/mathscinet/
[2]http://www.ams.org/publications/math-reviews/mrelecjour

I think we (IPOL) can not decide to follow another path, unless we
decide not to be considered as a scientific journal. Then, this
requirement for immutable content applies to all the reviewed and
published material, ie article and software. Thus, we can not accept
any correction of the text or code after publication until a system to
handle successive versions of the articles is in place.

This also means that the corrections we already accepted to fix some
software issues do not comply with the MathReview rules, because they
were not implemented in a new distincr version of the article.

8<----------8<----------8<----------8<----------8<----------8<----------

Mathematical Reviews® Policy on Indexing Electronic Journals

Mathematical research depends on a body of research literature that
has reliable content and assured persistence. Mathematicians use the
literature to anchor new research in the old, and mathematics
crucially depends on the integrity of this structure. For many years,
journals have provided the framework for creating this body of
literature. Those journals adhered to standards of scholarship that
were designed to protect their integrity. Recently, however, a few
electronic journals have adopted practices that threaten these past
traditions. This could have profound consequences for future
mathematicians who may not be able to rely on the research literature
in the way we do today.

Articles posted on a journal's website should be considered
"published" unless the journal indicates clearly in the posting itself
that the article is not in final form. Once an article is "published"
it should be revised only in one of two ways -- by adding a link in
the article to a dated revision or by replacing the article with a
dated revision and adding an evident link to the original
article. This practice should apply to every aspect of the published
article, including the text, title, references, and ancillary
information. Published articles and all revisions should persist
indefinitely in the scholarly record.

If a journal currently indexed by Mathematical Reviews® does not adopt
these best practice standards, coverage of that journal will cease and
the editors of the journal will be informed. Coverage will be resumed
only when the journal agrees to these basic standards of scholarship.

This policy was approved by the Mathematical Reviews® Editorial
Committee on October 1, 2005.

-- 
Nicolas LIMARE - CMLA - ENS Cachan    http://www.cmla.ens-cachan.fr/~limare/
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: <http://tools.ipol.im/mailman/archive/discuss/attachments/20111207/f9ced8f3/attachment.pgp>