997763 - "By N contributors" feature includes users who made only empty revisions

Status: RESOLVED WONTFIX

(developer.mozilla.org Graveyard :: General, defect)

Description

[Janet Swisher]

What did you do?
================
For example, go to https://developer.mozilla.org/en-US/docs/MDN/Getting_started



What happened?
==============
The "by 5 contributors" line includes a user, "Vovan". If you look in the history, Vovan created 4 revisions, but none of them actually changed the page. 


What should have happened?
==========================
Vovan should not be listed, because Vovan did not contribute to the page. 

Is there anything else we should know?
======================================
The "by N contributors" feature is intended to give recognition to article contributors. Therefore, giving credit to users who didn't actually contribute anything undermines the purpose of the feature, by adding noise.

This is related to bug 812157 but is not the same thing. That bug is about not creating empty revisions. We already have large numbers of empty revisions in the page histories, so we need to ignore the ones that already exist.

Comment 1

[John Karahalis [:openjck]]

Great catch, Janet. I agree that we should not show contributors who saved revisions that included no changes.

If we want to solve two issues at once, and fix this problem without adding code that needs to be maintained, we could go ahead and take care of bug 812157 now. We could even mark this bug a duplicate of bug 812157 if we choose this approach.

Comment 2

[David Walsh :davidwalsh]

This has come up quite a few times now.  Can we add logic in the "Document.save" method which checks for content / title / tags / slug / ... changes, and if none are detected, no create another revision?

Comment 3

[Janet Swisher]

Fixing 8125157 would be great, since that would address the source of the problem. But it would not address how to deal with the existing empty revisions.

The feature also displays users whose revisions have been reverted -- a bit trickier situation to detect.

Comment 4

[Luke Crouch [:groovecoder]]

Hmm ... to be honest, this is a *massive* task, because:

* Every revision is saved with the entire content as-is
* Every revision is saved with a based_on field pointing to the previous revision

To hide non-contributors revisions, we would have to either:

1. On page-save, scan thru each of every contributor's revisions, compare their content to the based_on revision content; only set a flag to display the user in the bar if the contents are different

2. 1-time massive task to perform #1 on every document in the database and remove the empty revisions; resetting all subsequent revisions' based_on values appropriately

#1 is approx. 2 weeks of work; #2 is probably 4-6 weeks of work. :(

Comment 5

[James Bennett [:ubernostrum]]

TBH though, something along the lines of what David suggests in comment #2 is probably going to be good as a long-term plan.

Right now kuma generates new revisions for things that really shouldn't, or at least shouldn't generate the same effect as actually changing the contents. So either coming up with a way to independently track changes to tags review status and other metadata, or to mark them somehow to hide them from things that care about content changes, should be on our roadmap.

And yes, it is a lot of work.

Comment 6

[Māris Fogels [:mars] (please needinfo)]

Empty revisions is a problem that blocks the badges project, too.  Breaking out bug 1018206 to track the empty-revisions problem.

Comment 7

[BMO Automation]

MDN Web Docs' bug reporting has now moved to GitHub. From now on, please file content bugs at https://github.com/mdn/sprints/issues/ and platform bugs at https://github.com/mdn/kuma/issues/.