| |
@@ -0,0 +1,57 @@
|
| |
+ .. _solution_probntec.rst:
|
| |
+
|
| |
+ Conundrum of Tracking of Non-Technical Contributions
|
| |
+ ====
|
| |
+
|
| |
+ As contributions that involve interacting with services are the only ones that
|
| |
+ publish a message on the
|
| |
+ `Fedora Messaging <https://fedora-messaging.readthedocs.io/>`_ bus, they are
|
| |
+ the only ones that get tracked. These services predominantly facilitate for
|
| |
+ technical contributions alone and hence, it becomes increasingly difficult to
|
| |
+ track non-technical contributions. Owing to the wide range of creating non
|
| |
+ technical contributions and a similarly huge number of subjective methods to
|
| |
+ keep track of those - it is comparatively difficult to come up with an
|
| |
+ objective method of describing them. What might look like a contribution to one
|
| |
+ might not look like a contribution to someone else and therefore, the service
|
| |
+ would not be able to reliably track non-technical contributions.
|
| |
+
|
| |
+ One of the workarounds that could help solving this problem is by creating
|
| |
+ badges on the
|
| |
+ `Fedora Badges <https://fedora-arc.readthedocs.io/en/latest/badges/index.html>`_
|
| |
+ platform for each type of non-technical contribution. This cannot be a
|
| |
+ solution as this method scales incredibly poorly as the members go on adding
|
| |
+ new badges of the manually-awarded kind for tracking each type of non-technical
|
| |
+ contributions. This could potentially lead to situations where there are
|
| |
+ thousands of badges pertaining to a type of non-technical contribution but with
|
| |
+ handful of awardees per badge. Going against the norm of having "subject"
|
| |
+ usernames as owners of contribution activities, using Fedora Badges to track
|
| |
+ contributions would lead to treating both "subject" usernames (cause of the
|
| |
+ message publication event) and "object" usernames (involved in the said
|
| |
+ contribution activity) to own a contribution activity.
|
| |
+
|
| |
+ How? Assuming a contribution event where a "Member X" (who has the access to
|
| |
+ award a certain "Badge A" to members who do "Task B") awards a "Badge A" to
|
| |
+ "Member Y". As this was a non-technical contribution and because this could not
|
| |
+ be automatically tracked by the Fedora Badges' Message Consumer entity - the
|
| |
+ event of "Member Y" performing "Task B" could not be reliably and automatically
|
| |
+ tracked. The event of awarding the badge essentially confirms two contribution
|
| |
+ activities - the first one being "Member Y" performing "Task B" and the second
|
| |
+ one being "Member X" awarding the "Member Y" with "Badge A". In the first
|
| |
+ contribution activity, the "Member Y" owns the contribution but they are
|
| |
+ dependent on "Member X" to award them the badge to have their contribution
|
| |
+ tracked. The activity of awarding the badge itself is a contribution so in the
|
| |
+ second activity, the "Member X" owns the contribution.
|
| |
+
|
| |
+ This should be used as an absolute "last ditch effort" to track non-technical
|
| |
+ contributions of the kind that do not automatically publish a message on the
|
| |
+ Fedora Messaging bus when they take place. As the badges cannot be tightly
|
| |
+ associated with the contribution activity itself and the fact that the way of
|
| |
+ tracking those are manual in nature, they are very susceptible to be left
|
| |
+ untracked (due to reasons like the "contribution owner" forgetting to claim
|
| |
+ a badge when they made the "contribution activity" or the "badge felicitator"
|
| |
+ being unavailable when being reached out causing delays in tracking of the
|
| |
+ said activity) or to be tracked unfairly (due to reasons like the "badge
|
| |
+ felicitator" awarding a certain badge to their friends who might not have done
|
| |
+ the said "contribution activity" or random members claiming the badge from the
|
| |
+ associated link in bad faith). In the larger scheme of things - this will end
|
| |
+ greatly skewing the contributor statistics for the worse.
|
| |
Signed-off-by: Akashdeep Dhar akashdeep.dhar@gmail.com