User talk:Dash1291/GSoC 2012 Application

From mediawiki.org

Passing the edit control in phase 1 might lead to a problem[edit]

I was talking to Ian about this proposal last week, and when I mentioned that a phase 1 system that supports one editor only could trivially be made collaborative-ish by allowing the editor to pass on the editor flag to someone else, Ian immediately remarked that this introduces one of the known problems with collaborative editing: who should the resulting revision(s) be attributed to?

Right now a revision (change to a page) is attributed to a single user, and that's fine because collaboration is not supported. But in a collaborative editing world, and even in a world where concurrent collaboration isn't possible but sequential collaboration is, this is a problem. It's kind of a hard problem that needs to be well thought through, so it might be better to defer this feature until later, or make it easy to disable. --Catrope (talk) 21:06, 19 March 2012 (UTC)Reply

It seems like a pretty simple issue to me. Eventually, the revision structure will have to be modified in a number of ways to support collaborative editing by allowing edits to have multiple authors, as well as including a complete log of all events that happened during the diff in order to make things completely transparent, but until that happens, there are two options that I can think of:
  1. The edit is attributed to either the first or last editor, and other editors are credited exclusively within the edit summary. This poses a number of problems, including causing the edit to not show up on user contributions, as well as allowing "faked" edits by just doing a normal edit but modifying the summary to make it look like someone else did the edit.
  2. Having addition editors, or perhaps just all editors, be attributed in an automatically created log diff (like the "moved this page to xxx" 'diffs') with text something along the lines of "ExampleUser participated in diff 123456 to Example". (If all collaborators are credited this way, than the actual edit would have to have some fake account or something in the space for the editor.) This would fix the usercontribs issue.
Neither of these are ideal, and they couldn't really be used widely, but they'd function well enough as a temporary measure, I think. (I've written more thoughts on collaborative editing at Talk:Future/Real-time collaboration.) --Yair rand (talk) 22:19, 19 March 2012 (UTC)Reply
I had thought of a simple approach of attributing the revision under a single editing session to the editor in control of that session since it will be the only client/user who will be making changes to the page in an editing session. This way contributions will correspond to the actual contributors who committed changes to the page.
Infact, even in the case of multiple editors as in future, similar approach could be used to attribute diffs to the users committed by them during their editing session. I don't know if its feasible and makes sense but I think it could be pulled off that away. -- Ashish Dubey (talk) 15:40, 20 March 2012 (UTC)Reply