No federated user access on an individual project

[gueroJeff]

Case: User3 logs into the scorecard to evaluate the translations in project XX_YY. Users 1, 2, and 3 have all been assigned to work on project XX_YY together. User3 enters the project to rate segments and finds that all of the segments have been rated. User3 is seeing User1 and User2's ratings within the project, skewing User3's ability to rate without external influence.

Expected behavior: User3 should be presented with an unrated set of segments to rate. Each user's rating should remain visible to their account only, not to any user account assigned to a project.

[arlelommel]

Thanks for the feedback. At present our working model is that each user would have a separate copy of the project to work on, but I think your approach makes more sense from the user perspective. I will discuss with Jan when he is back and see if we can figure a way to do this. I think the simplest way would be, if a non-admin user is logged in, that we filter the issues by who created them. That should be an easy way to deal with this. Then we would need to figure out how to present the various users distinctly to the admin user.

[gueroJeff]

Thanks Arle. The working model actually is currently broken. Users are not seeing separate copies of the project to work on, but a single copy with each other's annotations. This is visible in the "Bad French" project.

El 8/12/15 a las 2:16 AM, arlelommel escribió:

Thanks for the feedback. At present our working model is that each user would have a separate copy of the project to work on, but I think your approach makes more sense from the user perspective. I will discuss with Jan when he is back and see if we can figure a way to do this. I think the simplest way would be, if a non-admin user is logged in, that we filter the issues by who created them. That should be an easy way to deal with this. Then we would need to figure out how to present the various users distinctly to the admin user.

— Reply to this email directly or view it on GitHub https://github.com/multidimensionalquality/qt21-scorecard/issues/32#issuecomment-130213056.

Jeff Beatty Localization Engineer @mozilla_l10n http://twitter.com/mozilla_l10n 801.367.3763

[arlelommel]

I think I wasn't clear. What we'd need to do is create separate projects that are identical, rather than single instances of a single project. So we would repeat the creation of the project separately for each used rather than assign one project to multiple users.

I'll have to discuss with Jan how to handle this.

-Arle

[gueroJeff]

This hasn't been a problem in the past though. We've been able to set up a project and point various users to it without them seeing each other's work.

El 8/12/15 a las 7:35 AM, arlelommel escribió:

I think I wasn't clear. What we'd need to do is create separate projects that are identical, rather than single instances of a single project. So we would repeat the creation of the project separately for each used rather than assign one project to multiple users.

I'll have to discuss with Jan how to handle this.

-Arle

— Reply to this email directly or view it on GitHub https://github.com/multidimensionalquality/qt21-scorecard/issues/32#issuecomment-130305028.

Jeff Beatty Localization Engineer @mozilla_l10n http://twitter.com/mozilla_l10n 801.367.3763

[jnehring]

The current design of Scorecard is that users work collaboratively on projects so the users see each others ratings. So this is meant as a feature, although it was probably a misunderstanding.

An easy workaround (as suggested above) is to generate separate projects and to assign each user to one project. This solution is easy because it does not require any code changes but I think it is not very elegant to have many almost identical projects.

We can change it in a way that each user sees only his own changes. But it introduces a more complicated workflow for administrators because now administrators have to be able to see the same project with multiple ratings and maybe also an aggregated view that merges the ratings by all users. Administrators do not see one score but many scores and maybe an average score. The administrator view might introduce a work intense change.

[gueroJeff]

El 8/24/15 a las 2:53 AM, Jan Nehring escribió:

The current design of Scorecard is that users work collaboratively on projects so the users see each others ratings. So this is meant as a feature, although it was probably a misunderstanding.

An easy workaround (as suggested above) is to generate separate projects and to assign each user to one project. This solution is easy because it does not require any code changes but I think it is not very elegant to have many almost identical projects.

Agreed.

We can change it in a way that each user sees only his own changes. But it introduces a more complicated workflow for administrators because now administrators have to be able to see the same project with multiple ratings and maybe also an aggregated view that merges the ratings by all users. Administrators do not see one score but many scores and maybe an average score. The administrator view might introduce a work intense change.

True. I don't see another way of being able to maintain IRR integrity across users of a project though when an admin is unable to distinguish between the ratings of two or more individual users.

— Reply to this email directly or view it on GitHub https://github.com/multidimensionalquality/qt21-scorecard/issues/32#issuecomment-134100010.

Jeff Beatty Localization Engineer @mozilla_l10n http://twitter.com/mozilla_l10n 801.367.3763