Refactor the Reviews mechanism

[jonquet]

Is your feature request related to a problem? Please describe. OntoPortal used to have a review mechanism to enable any users to write a review (and give a 5-star based note) to an ontology. This was showing up in the Summary page. This features has been certainly de-activated from the UI but there is still the mechanism in the code e.g., http://data.agroportal.lirmm.fr/ontologies/CROPUSAGE/reviews

Describe the solution you'd like My proposition is to re-enable this feature based on a selected choice by the ontology admin. And admin will enable or not user reviews inside the "New/Edit Ontology submission" view and would later have the capacity to see/hide/report reviews that would not be appropriate for an ontology.

Additional context We will implement this in AgroPortal first with the following task list (to be completed and linked to specific issues by @syphax-bouazzouni):

• [ ] Add a metadata property on the Ontology model "enableReviews"
• [ ] Enable the edition (check box) of this property in the "New Ontology" view
• [ ] Enable edition of this property in the "Edit Ontology view" and if checkbox selected then a table will show displaying the reviews as served by the API
• [ ] Modify the Review model to add a "hidden" boolean property that will be available to the admin on Edit Ontology page review table
• [ ] Modify the Review model to add a "reportAbuse" boolean property that will be available on the same page, but for which an activation of the checkbox will fire an email to the portal admins
• [ ] Create a "Review monitoring" tab in the Administration section of OntoPortal that will display all the Reviews for all the ontologies and that would allow a portal admin to delete a review.

[syphax-bouazzouni]

Related https://github.com/ontoportal-lirmm/bioportal_web_ui/pull/69 (the preexistent system of reviews)

I think for this point, we need to also to harmonize the system of user feedback for an ontology administrator In brief, I think we need only one button "Give feedback", that will pop up a modal. In that modal, the user will have the choice between "Proposals", "Comment", "Reviews", "Ontology bridge", "GitHub proposals" and so on, if there is anything else.

That button needs to be visible (not hidden in the notes sections) and well-documented(with a help tooltip).

@Bilelkihal please take note of this, to propose a mockup

[syphax-bouazzouni]

@graybeal Can you please, give us the historical reason why this feature was removed at first (so that we don't redo the same faults)

At some time (8 years ago) there were also "margin notes" (which the code is still in the UI project)

[graybeal]

Can you please, give us the historical reason why this feature was removed at first (so that we don't redo the same faults)

Well, the original implementation and experimentation was before my time, but i will say what I understand and hazard a guess.

Once the original feature was implemented, it's my understanding that an attempt to 'seed' the process was made by recruiting community members (paid?) to commit reviews. I remember seeing about 5 reviews, I'm not sure how many were actually created; I think most or all of those reviews were recruited, not spontaneous.

No other reviews were received in the following years (at least 5 years), to my knowledge.

In the meantime, the value of the reviews become less significant, as (a) the reviews became outdated, (b) the ontologies became outdates, (c) the lack of other reviews made the process less credible, and/or (d) the reviews 'stuck out' in the UI (not in a good way) and raised more questions than they answered.

Because of that slow loss of value, and lack of perceived value even in the early stages, we made the decision to remove the feature from the UI.

I believe that the fundamental issues with reviews, as they were built in BioPortal, were as follows: (A) There were not many people with time or expertise to generate reviews—even back then, many fewer reviewers than assets, while ideally you'd like many reviews per asset. (B) There was no paradigm (other than the 4 factors) to define how the review should be conducted, so the credibility and consistency suffered. (C) If you're going to review something, presumably you are reviewing it against a particular purpose. But ontologies are created for many different purposes, and in many cases can be and was criticized (in a review) for not filling a purpose or goal that the ontology was never trying to fill. Even a goal like 'completeness' is open to interpretation (for what purpose? compared to what gold standard?). (D) At least by the time I started, the reviews were not well-used by BioPortal (e.g., in prioritizing ontology lists or matches) nor well presented in BioPortal (in 'favorites' lists, ontology recommendations, or elsewhere. So the value proposition was never very high.

The document Recommending Terms and Ontologies may be a useful reference.

[graybeal]

Regarding https://github.com/ontoportal/ontoportal-project/issues/10#issuecomment-1354783141, I think harmonizing feedback makes some sense but should be in a different issue. (There's another ticket about choosing/harmonizing an ontology request mechanism which is related, and there are potential conflicts that will result by putting everything in on place. But maybe if it's a tab, to consolidate all those options? Anyway, a different ticket would be good.

[graybeal]

Regarding the second comment in https://github.com/ontoportal/ontoportal-project/issues/10#issuecomment-1354797662 (margin notes) (nice catch! Hadn't known that before): There are two concerns to consider in all of this.

A user looking at an ontology may be confused and frustrated by an accumulation of notes, requests, review, and other meta-content that interferes with their primary use of the ontology (for most users, looking up and referencing or comparing terms, definitions, and other content). The content can be very noisy if almost none of it applies to your immediate task. Giving the user to turn these annotations off and on is essential for preserving a clean and highly usable look and feel.

An ontology author may quickly get very tired of notifications of any of these types. Whether it's best to give them the ability to disable the feature (as suggested above), or just not receive any notifications from the feature (which may lead to a build-up of never-acted-on annotations), or both, it is valuable to consider and handle systematically throughout the system.

[jonquet]

Indeed, the issue was about the Reviews mechanisms. Marginal notes (deprecated) and other feedback mechanisms are very related to Reviews, but in a sense differents as they are more or less all at the level of an object inside an ontology, whereas the Review mechanism is at the level of the ontology globally. However, in a sense the Reviews mechanism is a bit redundant with the global Notes mechanism which allows a user to edit a comment about the ontology globally. This feature is still active in BioPortal (and others):

However I would suggest to focus in this issue on the Reviews.

[jonquet]

Thanks @graybeal for the perspective and REX with BioPortal.

I understand the reasons why the Review feature was "disabled" in BioPortal. I also believe the OntoPortal use cases are more diverse and various than the BioPortal one now and I believe the Reviews might be of interest for certain groups; including I think (maybe I shall turn wrong) in AgroPortal. For instance, I think that a portal with a strongly dedicated/engaged "community manager" could encourage the users to submit reviews and get something out of this. I keep earing from users that they "need ways to know which ontology to select and use" and I think the Reviews mechanism could be part of the solution.

Acknowledging the differences between the OntoPortal installations and their corresponding communities, my proposition is to implement this (back) as an optional component in OntoPortal.

[jonquet]

FYI: I have kicked off a discussion for the generalization of this: https://github.com/ontoportal/ontoportal-project/discussions/13

I suggest we continue here only with the revival/enhancement of the Review mechanism and then later harmonize with the rest.