Review Section 5 of Methodology

smiths / AIMSS

Assessing the Impact of MDE (Model Driven Engineering) and code generation on the Sustainability of SCS (Scientific Computing Software)

2 stars 2 forks source link

Review Section 5 of Methodology #49

Closed peter-michalski closed 4 years ago

peter-michalski commented 4 years ago

Please review "Identify Candidate Software" in the Methodology document and provide feedback.

https://github.com/smiths/AIMSS/blob/master/StateOfPractice/Methodology/Methodology.pdf

smiths commented 4 years ago

This issue is assigned to @Ao99, but I'll add a few short comments:

empirical is spelled incorrectly in Step 3
I think we want to add that our search phrases are looking for authoritative lists of software from the domain - you can give some examples of some lists from different domain. We did this in the previous state of the practice exercise
I think we should mention that the domain experts will be asked for their suggestions, and the domain experts will be asked to review the list

Ao99 commented 4 years ago

I put my comments directly in the doc, please check 5ba52ef.

peter-michalski commented 4 years ago

@smiths @Ao99 Thank you for your suggestions, I have made changes and comments in b0fc4c50b6f1f6ba7feaf2128bfed03178501869. We can discuss this at our next meeting.

Regarding "you can give some examples of some lists from different domain", I did not use any what I would call very "authoritative" lists from online searches. I used, among other things, relative keyword searches on GitHub and swMATH. Would this suffice to be authoritative enough? Many of these solvers were also found to be mentioned in papers.

peter-michalski commented 4 years ago

Regarding your comment concerning "The candidate software should have the following properties" @Ao99

[“Ideally have the latest release or source code commit within the last 5 years.” How about this one? Do we care about the “vintage” software? —AD]

I tested some software that has not been updated in about 8 years, as this software was mentioned in multiple places and looked professional. That being said, I agree that age could be noted as something that is considered along with other qualities.

Perhaps a rephrasing of "Ideally the latest release or source code commit should be relatively recent, such as within the last 5 years, unless the candidate software appears to be well recommended and currently in use."

What do you think? I would also appreciate input from @smiths

peter-michalski commented 4 years ago

@smiths @Ao99 All of the above points have been addressed for Section 5. I will now close this issue.