search

openjournals / joss-reviews

Reviews for the Journal of Open Source Software
Creative Commons Zero v1.0 Universal
694 stars 36 forks source link

[REVIEW]: The nnlib2 library and nnlib2Rcpp R package for implementing neural networks #2876

Closed whedon closed 3 years ago

whedon commented 3 years ago

Submitting author: @VNNikolaidis (Vasilis Nikolaidis) Repository: https://github.com/VNNikolaidis/nnlib2Rcpp Version: v.0.1.7.a Editor: @kakiac Reviewers: @schnorr, @MohmedSoudy Archive: 10.5281/zenodo.4780958

:warning: JOSS reduced service mode :warning:

Due to the challenges of the COVID-19 pandemic, JOSS is currently operating in a "reduced service mode". You can read more about what that means in our blog post.

Status

status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/ef16eef177a97389ded60b15ef4687aa"><img src="https://joss.theoj.org/papers/ef16eef177a97389ded60b15ef4687aa/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/ef16eef177a97389ded60b15ef4687aa/status.svg)](https://joss.theoj.org/papers/ef16eef177a97389ded60b15ef4687aa)

Reviewers and authors:

Please avoid lengthy details of difficulties in the review thread. Instead, please create a new issue in the target repository and link to those issues (especially acceptance-blockers) by leaving comments in the review thread below. (For completists: if the target issue tracker is also on GitHub, linking the review thread in the issue or vice versa will create corresponding breadcrumb trails in the link target.)

Reviewer instructions & questions

@schnorr, please carry out your review in this issue by updating the checklist below. If you cannot edit the checklist please:

  1. Make sure you're logged in to your GitHub account
  2. Be sure to accept the invite at this URL: https://github.com/openjournals/joss-reviews/invitations

The reviewer guidelines are available here: https://joss.readthedocs.io/en/latest/reviewer_guidelines.html. Any questions/concerns please let @kakiac know.

✨ Please start on your review when you are able, and be sure to complete your review in the next six weeks, at the very latest ✨

Review checklist for @schnorr

Conflict of interest

Code of Conduct

General checks

Functionality

Documentation

Software paper

Review checklist for @MohmedSoudy

Conflict of interest

Code of Conduct

General checks

Functionality

Documentation

Software paper

whedon commented 3 years ago

Hello human, I'm @whedon, a robot that can help you with some common editorial tasks. @schnorr it looks like you're currently assigned to review this paper :tada:.

:warning: JOSS reduced service mode :warning:

Due to the challenges of the COVID-19 pandemic, JOSS is currently operating in a "reduced service mode". You can read more about what that means in our blog post.

:star: Important :star:

If you haven't already, you should seriously consider unsubscribing from GitHub notifications for this (https://github.com/openjournals/joss-reviews) repository. As a reviewer, you're probably currently watching this repository which means for GitHub's default behaviour you will receive notifications (emails) for all reviews 😿

To fix this do the following two things:

  1. Set yourself as 'Not watching' https://github.com/openjournals/joss-reviews:

watching

  1. You may also like to change your default settings for this watching repositories in your GitHub profile here: https://github.com/settings/notifications

notifications

For a list of things I can do to help you, just type:

@whedon commands

For example, to regenerate the paper pdf after making changes in the paper's md or bib files, type:

@whedon generate pdf
whedon commented 3 years ago

PDF failed to compile for issue #2876 with the following error:

Can't find any papers to compile :-(

kakiac commented 3 years ago

@whedon generate pdf from branch JOSS

whedon commented 3 years ago 
Attempting PDF compilation from custom branch JOSS. Reticulating splines etc...
whedon commented 3 years ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

kakiac commented 3 years ago

👋 @agisga @billchenxi @MohmedSoudy @paragkulkarni11 would any of you be available to review this submission for JOSS? We carry out our checklist-driven reviews here in GitHub issues and follow these guidelines: https://joss.readthedocs.io/en/latest/review_criteria.html

MohmedSoudy commented 3 years ago

Am ready to review.

On Wed, Dec 2, 2020 at 2:30 PM kakiac notifications@github.com wrote:

👋 @agisga https://github.com/agisga @billchenxi https://github.com/billchenxi @MohmedSoudy https://github.com/MohmedSoudy @paragkulkarni11 https://github.com/paragkulkarni11 would any of you be available to review this submission for JOSS? We carry out our checklist-driven reviews here in GitHub issues and follow these guidelines: https://joss.readthedocs.io/en/latest/review_criteria.html

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/2876#issuecomment-737200432, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG2XKPZ22TBY5JHBZIN6BELSSYXNZANCNFSM4UJEKQKQ .

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

kakiac commented 3 years ago

Thank you @MohmedSoudy - I will now assign you as a reviewer 😃

kakiac commented 3 years ago

@whedon add @MohmedSoudy as reviewer

whedon commented 3 years ago

OK, @MohmedSoudy is now a reviewer

kakiac commented 3 years ago

👋 Dear Lucas (@schnorr) and Mohmed (@MohmedSoudy),

Many thanks for kindly agreeing to review Vasilis' (@VNNikolaidis) submission for the Journal of Open Source Software (JOSS), we’re delighted to have your help. 🥳

This is the review thread for the paper. All of our communications will happen here from now on.

There are two checklists at the top of this issue for the reviewers, please use these to track your progress of the review. If you could like more details about the JOSS reviewing process (it's slightly more interesting than for other journals our there!), have a look at the JOSS reviewer pages.

The JOSS review is different from most other journals. Our goal is to work with the authors to help them meet our criteria instead of merely passing judgment on the submission. As such, the reviewers are encouraged to submit issues and pull requests on the software repository. When doing so, please mention openjournals/joss-reviews#2876 so that a link is created to this thread (and I can keep an eye on what is happening). Please also feel free to comment and ask questions on this thread. In my experience, it is better to post comments/questions/suggestions as you come across them instead of waiting until you've reviewed the entire package.

We aim for reviews to be completed within about 2-4 weeks. Please let me know if any of you require some more time. We can also use Whedon (our bot) to set automatic reminders if you know you'll be away for a known period of time.

Please feel free to ping me (@kakiac) if you have any questions/concerns.

Many thanks and looking forward to working with you for this submission.

kakiac commented 3 years ago

@whedon re-invite @schnorr as reviewer

whedon commented 3 years ago

The reviewer already has a pending invite.

@schnorr please accept the invite by clicking this link: https://github.com/openjournals/joss-reviews/invitations

kakiac commented 3 years ago

@whedon check references

kakiac commented 3 years ago

@whedon check repository

whedon commented 3 years ago 
Software report (experimental):

github.com/AlDanial/cloc v 1.88  T=0.29 s (150.3 files/s, 28958.2 lines/s)
-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
C++                             18           1039            748           3039
C/C++ Header                    20            712            549           2058
Markdown                         3             43              0             83
R                                2              3              2             10
-------------------------------------------------------------------------------
SUM:                            43           1797           1299           5190
-------------------------------------------------------------------------------

Statistical information for the repository '734b4b02d3f35e7bc12a12e4' was
gathered on 2020/12/03.
The following historical commit information, by author, was found:

Author                     Commits    Insertions      Deletions    % of changes
Vasilis N. Nikolaidi            10          9369           1224          100.00

Below are the number of rows from each author that have survived and are still
intact in the current revision:

Author                     Rows      Stability          Age       % in comments
Vasilis N. Nikolaidi       8145           86.9          0.3               16.22
kakiac commented 3 years ago

@whedon generate pdf from branch JOSS

whedon commented 3 years ago 
Attempting PDF compilation from custom branch JOSS. Reticulating splines etc...
whedon commented 3 years ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

kakiac commented 3 years ago

@whedon check references

VNNikolaidis commented 3 years ago

As the author, I would like to thank everyone for volunteering their time and effort to the review of this paper and software, especially the editor @kakiac. This is my first substitution to JOSS, and I admit I do not know much about the JOSS reviewing process, so any guidance will be appreciated.

As suggested by @kakiac (here), the paper may gain if extended to present further details of the software for the specialist and non-specialist reader. I admit I may have written this brief paper more as an invitation for possible collaborators than a presentation to users.

I have given some thought on how this could be remedied, but hesitate to do changes before the reviewers have taken a look at the material, as I am afraid it may confuse the reviewing process.

This software is actually a combination of two entities (a C++ library and an R package), and also has two target audiences; below, I will explain what I could add for each:

(a) this software may be useful to people only want to use the ‘nnlib2Rcpp’ R package as is. The R package does contains some predefined, ready-to-use (and rather well-known) neural network models. I could easily add some demonstration material for those (and simple examples, similar to those in the package documentation).

(b) The paper also presents the ‘nnlib2’ collection of C++ classes for building neural networks. This is definitely a more advanced subject, as these classes may be used for experimentation with new, unusual or custom models and configurations. Such process also involves creating custom C++ sub-classes and recompiling the package, steps definitely not easy for the non-specialist reader. However this is, in my opinion, an interesting and unusual part of the provided functionality and, with the ‘nnlib2RCpp’ R package having an interface (the “NN” R module) for manipulating such components, this may fit the needs of some more advanced experimenters. I can also add some guidance for doing this (I have already written a brief blog post on this – it may be found here)). Maybe something similar could be helpful to the article readers.

I would appreciate if the editor and/or the reviewers give me their opinions on the above possible changes, and the ‘ok’ to proceed with such changes if they find that are needed.

MohmedSoudy commented 3 years ago

Thank you for inviting me to review this work, regarding the manuscript i believe that authors should revise the grammar in the paper, also comparison with available packages on CRAN such as (nnet, deepnet and h2o) should be mentioned in the statement of need to enrich the manuscript and elaborate the novelty of author's work. regarding the documentation the documentation should be user-friendly and handy so i recommend to visualize the r codes (installation) as r codes. You can find a template in the following link https://github.com/MohmedSoudy/UniprotR/blob/master/README.md .

On Thu, Dec 3, 2020 at 5:13 PM Vasilis N. Nikolaidis < notifications@github.com> wrote:

As the author, I would like to thank everyone for volunteering their time and effort to the review of this paper and software, especially the editor @kakiac https://github.com/kakiac. This is my first substitution to JOSS, and I admit I do not know much about the JOSS reviewing process, so any guidance will be appreciated.

As suggested by @kakiac https://github.com/kakiac (here https://github.com/openjournals/joss-reviews/issues/2775#issuecomment-736551955), the paper may gain if extended to present further details of the software for the specialist and non-specialist reader. I admit I may have written this brief paper more as an invitation for possible collaborators than a presentation to users.

I have given some thought on how this could be remedied, but hesitate to do changes before the reviewers have taken a look at the material, as I am afraid it may confuse the reviewing process.

This software is actually a combination of two entities (a C++ library and an R package), and also has two target audiences; below, I will explain what I could add for each:

(a) this software may be useful to people only want to use the ‘nnlib2Rcpp’ R package as is. The R package does contains some predefined, ready-to-use (and rather well-known) neural network models. I could easily add some demonstration material for those (and simple examples, similar to those in the package documentation).

(b) The paper also presents the ‘nnlib’ collection of C++ classes for building neural networks. This is definitely a more advanced subject, as these classes may be used for experimentation with new, unusual or custom models and configurations. Such process also involves creating custom C++ sub-classes and recompiling the package, steps definitely not easy for the non-specialist reader. However this is, in my opinion, an interesting and unusual part of the provided functionality and, with the ‘nnlib2RCpp’ R package having an interface (the “NN” R module) for manipulating such components, this may fit the needs of some more advanced experimenters. I can also add some guidance for doing this (I have already written a brief blog post on this – it may be found here https://r-posts.com/creating-custom-neural-networks-with-nnlib2rcpp/)). Maybe something similar could be helpful to the article readers.

I would appreciate if the editor and/or the reviewers give me their opinions on the above possible changes, and the ‘ok’ to proceed with such changes if they find that are needed.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/2876#issuecomment-738068535, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG2XKP73H4247NFUWOHTTTDSS6TJZANCNFSM4UJEKQKQ .

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

kakiac commented 3 years ago

Dear @MohmedSoudy, many thanks for these comments. Can I ask you to also update the checklist available at the top of this issue (https://github.com/openjournals/joss-reviews/issues/2876) under your username, as you go through with your review? Reviewing for JOSS is slightly different from other journals in that we ask reviewers to also validate/run the code in the submitted repository - see here: https://joss.readthedocs.io/en/latest/reviewer_guidelines.html, so any specific thoughts on how to improve this are encouraged.

I will be adding your comments issues in the repository to help the authors keep track of what needs to be addressed.

Let me know if you have any questions, or if something is unclear 😄

whedon commented 3 years ago

:wave: @schnorr, please update us on how your review is going.

MohmedSoudy commented 3 years ago

For me i think much efforts could be done as the author didn't update the statement of need that compare his package with available packages in CRAN or Bioconductor and i believe this is a crucial part in the paper to show the novelty or the contribution of this work.

On Tue, Dec 8, 2020 at 4:54 PM whedon notifications@github.com wrote:

👋 @schnorr https://github.com/schnorr, please update us on how your review is going.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/2876#issuecomment-740667727, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG2XKP76LVXYYLWNUYTIMC3STY4ZDANCNFSM4UJEKQKQ .

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

schnorr commented 3 years ago

I completed the General Checks (Repository, License, Contribution and Authorship, Substantial scholarly effort). As the author stated, the R package is a layer to ease the use of the original nnlib2 package - also available in the R package source code as ruled by the CRAN submission procedures. I've left two minor issues regarding the License and the Readme files. The "substantial scholarly effort" has been evaluated using the sloccount which points to about +5K lines of code (including the nnlib2 c++ package that has been previously implemented). The way the commit history is organized made it hard to tell the difference between contributions of the nnlib2 and its R api through Rcpp since there is a single commit that pushes perhaps everything at once. If we consider only Rcpp files with sloccount src/Rcpp* (assuming those are the real contribution of the R package), we have about 800 sloc. So, the package provides an entry door to the more complete nnlib2 package, easing its use through the Rcpp interface. For the R community, this is very welcome.

schnorr commented 3 years ago

Regarding the Functionality perspective (Installation, Functionality, and Performance). Installation. I was able to install the package very easily using both the CRAN sources and the development version available in Github using the installation procedures described in the README.md file. When executing the very first command using the library - from the blog post - p <- new("NN"), I get the message (Note: NN is under development (beta).) using both the CRAN and the development version. Functionality. I followed the instructions of the blog post, as pointed out by the author, to attempt the utilization of the R package. I wasn't able to reproduce the functionality described there (using the development version) because of this:

> p$add_connection_set("perceptron")
Adding (empty) set of perceptron connections to topology.
(once topology is complete, use create_connections() to fill it).
Note: Adding connection set failed.
[1] FALSE
Warning message:
In .CppObject$add_connection_set(...) : Unknown connection set type

I didn't go much further using those recommendations because of this error. Using the documentation available in the PDF file, I was able to rerun all the code snippets with the multiple example usages. I didn't run tests outside of those already provided by the author, but based on these, the functionality is there. Performance. Regarding the performance, I wasn't able to conduct a complete performance evaluation, but the author is aware of the underlying nnlib2 library limitations due to its sequential nature (no parallelism whatsoever) as much of the NN requires strong computational capabilities (accelerated HW for instance).

schnorr commented 3 years ago

Regarding the Documentation (A statement of need, Installation instructions, Example usage, Functionality documentation, Automated tests, Community guidelines). A statement of need, Installation instructions. The documentation has two files (as far as I was able to find). The README.md contains a brief statement of need and short installation instructions. There is no example usage, functionality documentation, etc in the readme. Example usage, Functionality Documentation. The other file is a PDF that serves as documentation for the R package. It is a PDF file in the root of the git repo. It follows the strict guidelines for the CRAN, where every function available in the R package is documented. The PDF files contains an extensive documentation with one example usage to guide the user for every class. Community guidelines . The author do invite people to contribute to the software package (see the README.md file once again) with the list of topics requiring most attention. I wasn't able to find automated tests (continuous integration).

schnorr commented 3 years ago

Regarding Software paper (Summary, A statement of need, State of the field, Quality of writing, References). Summary. The summary provides a high-level functionality and purpose of the software for a diverse, non-specialist audience. Statement of need. The statement of need make it clear that the usage of this library targets students that want to understand the basics of building NN models from scratch, without the aid of more elaborated packages whose simplification is the main target. The text also highlights the R package integration using Rcpp, which is very nice for the target audience to avoid the intrinsic of the C++ language. I believe the author could provide some words on how much functionality of the nnlib2 library is available through the R package. Is there anything missing? State of the field. There is very limited information on how this software package relates to others targeting the same audience. Even if author provides references to Keras and Pytorch, I wonder whether other NN packages tailored for students do exists. Quality of writing. I think the English writing of the 2-page long paper is appropriate. Generally speaking, the paper is very short (only 2 pages) and I have to agree with my reviewer colleague that the software state of the art could be much more explored by listing other software packages with the same goals.

schnorr commented 3 years ago

Hi @kakiac, I've completed a first review. Since this is my first time as reviewer for JOSS, please let me know if I need to improve the procedure.

VNNikolaidis commented 3 years ago

@kakiac , @schnorr , @MohmedSoudy , I thank everyone for their time and useful comments, I will start handling them now. Some of the issues raised at the original repository by @schnorr as part of this review process have already been processed (but left open till I get your feedback). As I am now about to attempt to handle some of the comments stated above, is it appropriate and/or convenient to the editor and reviewers to place my comments here? If not, how should I proceed?

VNNikolaidis commented 3 years ago

If I may I would like to reply to some of the many useful comments by reviewer @schnorr:

When executing the very first command using the library - from the blog post - p <- new("NN"), I get the message (Note: NN is under development (beta).) using both the CRAN and the development version.

Indeed, I have left that message since the NN class is new and very versatile and I wanted to inform its users that the methods it provides may change, while more methods will be (and already have been) added. The message does not imply anything beyond that and was removed as it may be missleading.

I followed the instructions of the blog post, as pointed out by the author, to attempt the utilization of the R package. I wasn't able to reproduce the functionality described there (using the development version) because of this:

> p$add_connection_set("perceptron")
Adding (empty) set of perceptron connections to topology.
(once topology is complete, use create_connections() to fill it).
Note: Adding connection set failed.
[1] FALSE
Warning message:
In .CppObject$add_connection_set(...) : Unknown connection set type

This error is expected, and I can understand why it was confusing. The “perceptron” components are included with version v.0.1.5 which is currently at GitHub, so the example should work if the package is installed from there (not CRAN, this was mentioned in the post). The last version currently submitted to CRAN (v.0.1.4) does not have “perceptron”, but please keep in mind that the blog post is meant to explain how to create the missing “perceptron” component (or any other custom component), from scratch, even using this (v.0.1.4) package source code.

VNNikolaidis commented 3 years ago

Dear Editor, @kakiac , and Reviewers @schnorr , @MohmedSoudy, I wish you a Happy 2021.

As the status of my submission is not very clear, some changes were performed in accordance to your suggestions and comments.

In particular:

(a) a revised version of the paper was uploaded to the repository, which hopefully addresses some of your comments (s.a. the need to specify what differentiates this package from other similar tools, i.e. the State of the field issue). The revised version can be found at branch JOSS of the repository https://github.com/VNNikolaidis/nnlib2Rcpp (same location as before).

(b) following the Reviewer @schnorr suggestions, changes were performed to the repository’s README.md file to clarify installation steps with better markdown and better explain how to access documentation and examples (from package build-in help and CRAN auto-generated reference manual). In addition to the above, a document with instructions and examples (file ‘manual.pdf’) was added to the repository and is now mentioned in the README.md to better guide new users of the package. This file will become the package ‘vignette’ in future versions of ‘nnlib2Rcpp’.

(c) for the Performance issue, no changes were performed, since the software does not make any performance claims (as was correctly remarked by the reviewer @schnorr). Maybe this fact is sufficient to clear the issue.

(d) for the Automated tests issue, the reference manual examples in the build-in R package documentation were examined and verified by Reviewers @schnorr. Since there are no other automated tests, maybe this issue should be cleared.

(e) finally, for the License issue, unfortunately CRAN rules impose this format for the MIT license in submitted R packages. Therefore, the LICENCE file is written this way to be compatible with the strict CRAN submission rules where the name of the copyright holders is listed in a separate file. The LICENSE.md is there for GitHub and all other uses and contains the same licensing information (MIT + the name of copyright holder). As far as I am aware, it is not possible to satisfy CRAN restrictions in a single plain license file that contains both the license and the name of the copyright holder. See for example the discussion in:

https://stackoverflow.com/questions/43550479/how-to-satisfy-both-cran-and-github-license-file-naming-requirements

Since these are restrictions imposed by the target repository (CRAN), no changes were performed. However, this being an R package for CRAN, and LICENSE.md does contains the entire license, maybe this issue should be cleared.

Please let me know if these changes fulfill any shortcomings previously mentioned by the Reviewers, and what further changes are necessary to fulfill the JOSS submission requirements.

VNNikolaidis commented 3 years ago

Dear @kakiac (Editor), I took the liberty to remove the JOSS branch of the nnlib2Rcpp GitHub repository which was created solely for the review of my submission. The (revised) paper is now located at the master branch, which contains both the software and all related materials. Furthermore, I am getting ready to release the new version (v.0.1.6) to CRAN with the changes mentioned in the previous comment - based on the reviewer feedback, as well as other changes - see also CHANGES.md). The package now includes "vignette" documentation, found (in PDF format) as file manual.pdf. Finally I would appreciate any conclusive result on my paper submission, so (in case it is positive) use it in the package's CITATION information.

VNNikolaidis commented 3 years ago

Dear contributors to this review, Editor (@kakiac), reviewers (@schnorr , @MohmedSoudy), and robot (@whedon), I would appreciate any guidance to aid the review process of this submission and complete it towards a final result (positive or negative). It appears to me that the prossess is now halted at an indeterminate status of "open" and "under review", and I truly do now know if this is normal or further actions are required by me (the author). Thank you and regards.

arfon commented 3 years ago

:wave: @VNNikolaidis - apologies for the delay here - it seems like the handling editor (@kakiac) is struggling to find time to edit right now so I will take over from here. Thanks to @MohmedSoudy and @schnorr for all of their work thus far!

My read of this thread is that this submission is close to being accepted.

The one major issue I see that has not been addressed at this point is the "State of the field" requirements. Quoting from @schnorr's review above:

State of the field. There is very limited information on how this software package relates to others targeting the same audience. Even if author provides references to Keras and Pytorch, I wonder whether other NN packages tailored for students do exists.

@VNNikolaidis - could you please let us know how you have (or plan to) address this?

VNNikolaidis commented 3 years ago

@arfon - Thank you for your reply and your involvement. The editor’s ( @kakiac ) situation is more than understood, especially under these strange circumstances. I appreciate her time and involvement so far.

Since the initial submission (which was commented by the reviewers) several changes have been done both to the paper and the software, some of which where to remedy issues mentioned by the reviewers. These changes are described in previous posts above (here and below).

In particular, to address the State of Field, I expanded the related “Statement of Need” section to provide more information on what this software is about. With the abundance of neural network tools available an exhaustive comparison is next-to-impossible, and rarely is any tool completely unique, but I hope these changes better clarify what differentiates this package and may make it fit for some user's (and/or possibly contributor's) purposes. However I do not now if these changes suffice, so I would appreciate the feedback.

schnorr commented 3 years ago

@whedon generate pdf

whedon commented 3 years ago

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

arfon commented 3 years ago

In particular, to address the State of Field, I expanded the related “Statement of Need” section to provide more information on what this software is about. With the abundance of neural network tools available an exhaustive comparison is next-to-impossible, and rarely is any tool completely unique, but I hope these changes better clarify what differentiates this package and may make it fit for some user's (and/or possibly contributor's) purposes. However I do not now if these changes suffice, so I would appreciate the feedback.

Thank you for the update @VNNikolaidis. @schnorr and @MohmedSoudy - could I trouble you both to come and take another look at this submission to decide if you are ready to recommend this submission be accepted?

arfon commented 3 years ago

@schnorr and @MohmedSoudy - could I trouble you both to come and take another look at this submission to decide if you are ready to recommend this submission be accepted?

Friendly reminder to please check on this submission again sometime soon please @schnorr and @MohmedSoudy.

MohmedSoudy commented 3 years ago

Of course. I will check it.

On Sun, May 9, 2021 at 9:01 PM Arfon Smith @.***> wrote:

@schnorr https://github.com/schnorr and @MohmedSoudy https://github.com/MohmedSoudy - could I trouble you both to come and take another look at this submission to decide if you are ready to recommend this submission be accepted?

Friendly reminder to please check on this submission again sometime soon please @schnorr https://github.com/schnorr and @MohmedSoudy https://github.com/MohmedSoudy.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/2876#issuecomment-835864447, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG2XKP6ACKQTZT3UIJ5XVP3TM3LYBANCNFSM4UJEKQKQ .

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

schnorr commented 3 years ago

Sorry @arfon for taking so long to signal. I will check asap and keep you all posted.

MohmedSoudy commented 3 years ago

I think the paper now is ready to be published after applying the required modifications.

On Mon, May 10, 2021 at 9:33 AM Mohmed Soudy @.***> wrote:

Of course. I will check it.

On Sun, May 9, 2021 at 9:01 PM Arfon Smith @.***> wrote:

@schnorr https://github.com/schnorr and @MohmedSoudy https://github.com/MohmedSoudy - could I trouble you both to come and take another look at this submission to decide if you are ready to recommend this submission be accepted?

Friendly reminder to please check on this submission again sometime soon please @schnorr https://github.com/schnorr and @MohmedSoudy https://github.com/MohmedSoudy.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/2876#issuecomment-835864447, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG2XKP6ACKQTZT3UIJ5XVP3TM3LYBANCNFSM4UJEKQKQ .

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

-- Regards, Mohmed Soudy Mohmed

Research Associate

Proteomics & Metabolomics Unit Basic Research 57357 Cancer hospital for children

schnorr commented 3 years ago

So, I've reviewed my notes above against the latest version of the paper and software. The problems of using the software have been solved because now the CRAN version is aligned with the version expected by the blogpost mentioned previously. The concerns of citing related work have been fixed and the differences against the base C++ implementation are now clear. So, yeah, I think we are ready for publication.

arfon commented 3 years ago

@VNNikolaidis – At this point could you make a new release of this software that includes the changes that have resulted from this review. Then, please make an archive of the software in Zenodo/figshare/other service and update this thread with the DOI of the archive? For the Zenodo/figshare archive, please make sure that:

I can then move forward with accepting the submission.

VNNikolaidis commented 3 years ago

@arfon I hope I followed the instructions correctly (used Zenodo) and the related DOI is 10.5281/zenodo.4780958 (i.e. DOI) I thank you and everyone else involved.

arfon commented 3 years ago

@whedon set 10.5281/zenodo.4780958 as archive

whedon commented 3 years ago

OK. 10.5281/zenodo.4780958 is the archive.

arfon commented 3 years ago

@whedon accept