[IPOL discuss] [IPOL edit] Suggested changes in Submission Procedure/Author Manual

Miguel Colom colom at cmla.ens-cachan.fr
Wed Jul 26 15:43:54 CEST 2017 

Quoting Pascal Getreuer <getreuer at gmail.com>:

> Dear Miguel,
>
> Thank you for organizing this effort to modernize the IPOL software
> guidelines. It sorely needs to be done.
>
> As a high level comment, I agree with the general theme of relaxing
> constraints to encourage more submissions. I also agree with Pascal
> Monasse's comment advising against removing "all the restrictions
> unilaterally".

Of course.
Actually, I proposed to make some of the changes immediately because  
they were already approved by the Editorial Board (such as allowing  
MATLAB or Python) and I think it's important to change it as soon as  
possible.

About other changes, note that I've send it to the discussion list.

> Let me remind that the point of having the guide is to
> "facilitate the production and review of verifiable and usable software for
> reproducible research." Any rule that doesn't contribute toward that is
> unnecessary, but on the other hand, you don't want a complete willy-nilly
> free-for-all, so please consider
>
>  * portability -- if the code has obscure platform requirements, the
> research isn't really "reproducible"
>  * readability -- if the code works but is a convoluted mess, reviewers and
> users will have a hard time following what it does, and it won't be
> "verifiable"
>  * usabillity -- if the code lacks a clear README or only interoperates
> with an obscure file format, users won't know how to use it

I agree, of course.
My point is that the editor should have a more active role and use  
these points as a checklist. These rules should be in the guidelines  
for the authors, but instead of writing them as "you must..., it's  
mandatory that..., etc" they should explain that their editor will  
help them to have a code of quality by using these guidelines.

This implies changes in the writing of the text and most probably  
rearranging the sections.


>> - We accept not only C/C++ but also Python and MATLAB and we're open
>> to other languages eventually.
>
> Yes, the docs are badly out of date. The IPOL main page has an announcement
> from 2015(!) "From now on, IPOL will accept Matlab source code. Very soon,
> we will update the submission procedure accordingly." without corresponding
> indication in the software guide.
>
>
>>> - All those technical terms give the impression that we're going to
>>> reject anything which does not follow exactly the norm of the compile
>>> (C89, C99, C++98), which is not true.
>>
>> Still, having standard conforming code for C and C++ seems important to
> me for
>> portability. Naturally, as Python and Matlab have no such strict
> standards,
>> this does not apply to them.
>
> Like Pascal said, portability is important. Please nail things down a bit
> to keep that. Do you want authors to make unlimited use of nonstandard C++
> extensions? #pragmas? inline assembly? Probably not.

Of course not.
But the editor should guarantee that the code doesn't contain those.
If we guidelines seem to be really hard to follow, in practice we'll  
have any of these:
1) Please not submitting because they're not technically competent to  
follow these strict guidelines
2) People submitting anyway and Editors who don't really check the code.

Thus, let's keep the guidelines but let's also turn them into a  
checklist for the Editors.

> For C++, I suggest to loosen the guidelines to explicitly allow C++11
> (possibly also C++14 and C++17). It is a huge step forward for the language
> from C++98.
>
> For Matlab, the toolboxes are often individually purchased and very
> expensive. Encourage authors to use only what they need. Encourage
> Octave-compatible Matlab code where possible.

If I'm not wrong, in the last Editorial Board meeting it was decided  
that all the MATLAB code submitted should work in Octave too.

>>> We should remove lines such as:
>>> - max 80 characters per line, max 1000 lines per file
>>
>> Again, why? We do not want messy code, this is not too much to ask from
>> authors.
>
> Whether code fits in 80 chars or uses -Wextra in the makefile is pretty
> superficial, and I am strongly in favor of removing such rules. But we want
> to encourage readable code, and there should be some explicit guidelines.
> It helps the reviewers focus on what we care about and gives something
> authoritative for reviewers to lean on if they need to persuade an author
> to clean up messy code. Based on code I have seen as an IPOL reviewer, the
> main guidelines I would make are (1) "please indent code" (it is super easy
> to do, there are automatic tools for it, and it really does help
> readability), and (2) "please give variables and functions reasonable
> English names, and try to make them correspond to the article".
>
>
>> I'll create a Google Doc with the proposed text and I'll send the link to
> edit it.
>
> I look forward to your Doc!

I have a preliminary Doc with my proposal for the guidelines.
I didn't write the URL publicly because this discussion list is public  
and I don't want that it gets vandalized!

Pascal: I'll send you the URL in a separate email.
If anybody else wants the edit the Doc, please send an email and I'll  
reply with the URL.

Best,
Miguel




> Best,
> Pascal
>
>
>
>
> On Thu, Jul 6, 2017 at 2:57 AM, Miguel Colom <colom at cmla.ens-cachan.fr>
> wrote:
>
>> Quoting Jean-Michel Morel <moreljeanmichel at gmail.com>:
>>
>> Dear all,
>>> Enric is absolutely right  to ask for a formal vote on a final text.
>>> The preparation of the text might be done with an online google  doc?
>>>
>>
>> All right, then.
>> I'll create a Google Doc with the proposed text and I'll send the link to
>> edit it.
>>
>> Best,
>> Miguel
>>
>>
>>
>>
>>
>>
>>>
>>> Jean-Michel Morel
>>> moreljeanmichel at gmail.com
>>>
>>> On Thu, Jul 6, 2017 at 10:47 AM, Loic Simon <loic.simon at ensicaen.fr>
>>> wrote:
>>>
>>> Dear Miguel, Dear Enric,
>>>>
>>>> I think that Miguel should send a draft of his proposed modifications and
>>>> it can be validated/amended by editorial board members without need of a
>>>> meeting. That would allow to at least correct quickly the misleading
>>>> information, and we can start a sound discussion on other points on which
>>>> there is no consensus.
>>>>
>>>> Best,
>>>>
>>>> Loïc
>>>>
>>>> On Thu, Jul 6, 2017 at 10:41 AM, Miguel Colom <colom at cmla.ens-cachan.fr>
>>>> wrote:
>>>>
>>>> Quoting Enric Meinhardt-Llopis <enric.meinhardt at cmla.ens-cachan.fr>:
>>>>>
>>>>> Dear all,
>>>>>
>>>>>>
>>>>>> You are talking about major changes to IPOL here.  Should we summon
>>>>>> the whole editorial board to approve these changes?  (arrange a
>>>>>> meeting, or send a mail to everybody with a poll to explicitly accept
>>>>>> each of the changes)
>>>>>>
>>>>>> Maybe not all the people in the editorial board are reading this
>>>>>> mailing list, or maybe they are not reading it this week.
>>>>>>
>>>>>> I want to have a voice in these changes but I'm not sure that this
>>>>>> mailing list reaches everybody.
>>>>>>
>>>>>>
>>>>> Dear Enric,
>>>>>
>>>>> I think that to remove the false information in the IPOL pages (Software
>>>>> Guidelines, FAQ, help, etc.) is urgente and should be done now. Some of
>>>>> that false claims are that we don't support MATLAB, among others. We
>>>>> wouldn't be deciding anything new, but just correcting the docs.
>>>>>
>>>>> About summoning all the Editorial Board or emailing them, as you wish.
>>>>> My opinion is that we should be way more agile, in the sense that if we
>>>>> detect that our information is false, that we can improve it quickly,
>>>>> and
>>>>> someone has already taken the responsibility to do it, then it should be
>>>>> done immediately.
>>>>>
>>>>> I'm afraid that if we don't be that agile, most of the time things won't
>>>>> be done at all because we'll be waiting for eventual discussions that
>>>>> won't
>>>>> ever happen and in the worst case this wait will block the actions.
>>>>>
>>>>> My proposal is to simply update the docs and to rearrange some parts
>>>>> (move the C/++ to a dedicated part).
>>>>>
>>>>> Given said that, if you really think that it's necessary that all the
>>>>> Editorial Board votes, let's do that then. I'm not opposed at all.
>>>>>
>>>>> Best,
>>>>> Miguel
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Best,
>>>>>> Enric
>>>>>>
>>>>>>
>>>>>>
>>>>>> 2017-07-04 16:10 GMT+02:00, Jose Luis Lisani <joseluis.lisani at uib.es>:
>>>>>>
>>>>>> I agree with Miguel suggestions about the Submission Procedure and the
>>>>>>> Software Guidelines.
>>>>>>> I can update the information appearing in the web page, but first we
>>>>>>> should all agree in the new guidelines.
>>>>>>>
>>>>>>> --
>>>>>> IPOL - Image Processing On Line   - http://ipol.im/
>>>>>>
>>>>>> contact     edit at ipol.im          - http://www.ipol.im/meta/contact/
>>>>>> news+feeds  twitter @IPOL_journal - http://www.ipol.im/meta/feeds/
>>>>>> announces   announce at list.ipol.im - http://tools.ipol.im/mm/announce/
>>>>>> discussions discuss at list.ipol.im  - http://tools.ipol.im/mm/discuss/
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> IPOL - Image Processing On Line   - http://ipol.im/
>>>>>
>>>>> contact     edit at ipol.im          - http://www.ipol.im/meta/contact/
>>>>> news+feeds  twitter @IPOL_journal - http://www.ipol.im/meta/feeds/
>>>>> announces   announce at list.ipol.im - http://tools.ipol.im/mm/announce/
>>>>> discussions discuss at list.ipol.im  - http://tools.ipol.im/mm/discuss/
>>>>>
>>>>>
>>>>
>>>> --
>>>> IPOL - Image Processing On Line   - http://ipol.im/
>>>>
>>>> contact     edit at ipol.im          - http://www.ipol.im/meta/contact/
>>>> news+feeds  twitter @IPOL_journal - http://www.ipol.im/meta/feeds/
>>>> announces   announce at list.ipol.im - http://tools.ipol.im/mm/announce/
>>>> discussions discuss at list.ipol.im  - http://tools.ipol.im/mm/discuss/
>>>>
>>>>
>>>>
>>>
>>
>>
>> --
>> IPOL - Image Processing On Line   - http://ipol.im/
>>
>> contact     edit at ipol.im          - http://www.ipol.im/meta/contact/
>> news+feeds  twitter @IPOL_journal - http://www.ipol.im/meta/feeds/
>> announces   announce at list.ipol.im - http://tools.ipol.im/mm/announce/
>> discussions discuss at list.ipol.im  - http://tools.ipol.im/mm/discuss/
>>
>

More information about the discuss mailing list