[IPOL discuss] Corrections of erratas in IPOL articles

[Jean-Michel Morel]

Dear all,

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.

This does not mean that authors will be allowed to change willfully 
their article content or their code.

What kind of corrections will be allowed is a scientific matter. It must 
be decided in each case by a scientific editor and referees.

  For the time being, only two such requests have been done at IPOL:

-one to modify a wrong reference in an article;
-one to change parts of the code managing the demo of an article, but 
without changing the algorithm or code hard core.

In my opinion both requests are manageable. The second will be examined 
by a scientific editor.

Nevertheless in the future, we must be very careful to let authors 
identify clearly what part of their code is submitted. Once accepted, 
this part will not be allowed to vary.  As is clear in the IPOL rules, 
it is the part of the code that matches the algorithmic description 
given in the article that is submitted. The rest of the code, if  not 
refereed, is contingent.

I therefore propose to leave open the option to change the contingent 
part of the code. Nevertheless such a change will request the joint 
agreement of authors and editorial board.

Indeed, we need a dynamic journal, where the online articles are kept 
alive, in particular by updating the online demos to make them 
compatible between them for benchmarks.

I do not think that this contradicts Nicolas' principles, but it 
introduces a distinction between published (refereed) code and 
peripheral (unrefereed) code.

Best,
Jean-Michel



Nicolas Limare a écrit :
> (follow-up on the discuss mailing-list, discuss at list.ipool.im)
> 
> Hi,
> 
> For me, the principles are:
> * IPOL is a journal and the editorial rules of scientific journals
>   must be followed. Being 100% online, having some software content or
>   involving a close community where everyone knows the editors doesn't
>   affect these rules.
> * The journal rules apply to the reviewed content: article and
>   code. This content is published, identified, referenced, and as such
>   should not change over time.
> * As a journal, IPOL is not continuously updated. It will not feature
>   the latest versions of an algorithm or software. These new versions
>   can be distributed on the personal homepages of the authors, or in
>   new articles.
> 
> Here are my comments on Jean-Michel's text. I suggest to reorder the
> items in a logical progession: the article, then the code (implements
> the article), then the demo (uses the code) and archive (of the demo).
> 
>> -IPOL web page settings for articles, demo and archive can be
>> modified by the editorial board;
> 
> What is an "IPOL web page setting"? Is it the content? the
> presentation? the address?
> 
>> -the article itself can also be corrected provided the corrections
>> are approved by editor, or editor+referee (and by all authors).
> 
> I would like to limit that to minor error corrections (orthograph,
> punctuation, bad reference, etc...). If the math of an article is
> wrong, it should either stay wrong, maybe with an Errata "This is
> probably wrong", or restart a full review, with new reviewers (because
> the previous ones did not see the error).
> 
>> -Code corrections can be authorized, if the necessity or utility of
>> the changes is explained, and provided the  code continues doing
>> exactly what the article says it does. Such corrections must
>> therefore be approved by editor, or editor+referee (and by all
>> authors).
> 
> Here again, I would like to limit these modifications to error
> corrections (ie when there is a bug, the code does not work in a case
> not seen during the review) or when we have to change it (an external
> library changed drastically, or there is a copyright, patent or
> license problem). And I suggest to add that the modified code replaces
> the previous one. At any time, one and only one version of the code is
> distributed in an IPOL article.
> 
>> -demo changes by the authors or editors concerning the presentation
>> of results, the parameter ranges, etc. which do not change the code
>> are authorized. Nevertheless, since they may alter the meaning of
>> the article, they should be approved by a competent editor and by
>> all authors of the article.
> 
> I think it is important to repeat that demo must always use the code
> distributed in the IPOL article, in the current version. That is why,
> by design, the algorithm code is always retrieved by the demo system
> from the IPOL web page. Then the "which do not change the code" part
> is not needed.
> 
>> -items of the public archive of each article can be removed by the
>> editorial board (reasons: size, inadequate content, user requests);
> 
> correction: "archive of each demo" instead of "archive of each
> article"
> 
>> For each article or code correction, identifiable "Errata"
>> paragraphs should be inserted or appended in  the article or code
>> and in its "readme",  giving the exact correction (text before text
>> after), its date, and by whom it was approved.
> 
> OK I add this for README.txt to the software guidelines.
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> discuss mailing list
> discuss at list.ipol.im
> http://tools.ipol.im/mailman/listinfo/discuss