There are some technology changes that could make this much easier.
1) make it easy to see *your last version* of an article when you visit it. 2) provide a link to 'diffs since your last edit' 2.1) provide a way to comment directly on that diff, without having to laboriously cut and paste 3) make it easy to snapshot a version of a page and put it in [your userspace] for future work. For instance, there could be a one-click "move this revision to the same name in my userspace" button which would let you work on a set of ideas over time even if you had disagreements with others editing the article in a different direction. Once you were done, of course, you would still have to work out if or how to marge those changes back into the original 4) have an entire tag-set and cleanup section for "articles for merging" to address merge problems. 4.1) under the present process, the # of people who try to reach consensus on merge conflicts a) are few in number; there is nothing like an AfD cycle to bring in new eyes all the time; and b) work with tiny changes, so a 3-person team can simply shut down all incremental changes from someone whose changes they don't understand. Naturally frustrating.
Nothing in particular. Dozens of times every day i edit articles in which i see mistakes. Usually nobody complains, but sometimes the people who wrote most of the article get very upset about the fact that i touched it at all and send me messages saying this. I used to reply and politely explain that that, by definition, is the way wikis work and to cite WP:OWN or its Hebrew counterpart. Sometimes it helps, but sometimes it makes the person even more upset.
In such cases, as an Israeli saying goes, i am right, but i am not clever.
There's a tech/policy change that could make this easier:
Allow revisions to be named. We already allow multiple versions in a fundamental way - past revisions are kept forever. But we make it particularly hard to access them. By allowing revisions to have names or tags, we could make the sort of concern Amir mentions above help improve the project in a positive way, adding additional useful information for readers.
For instance, ARTICLE?u=amir could show the last revision edited by Amir, ARTICLE?t=good could show the last revision expressly marked as 'good' or better quality, ARTICLE?t=eb1911 could be the last revision tagged as 'from the 1911 Britannica' before it started to be significantly modified. Flagged revs could become a feature that chooses a tagset (beyond the most chronologically recent rev) used to decide what most visitors are shown when they visit a page. Users with accounts could choose their own default tagset.
The hard problem would remain deciding what the 99% of visitors who aren't logged in see when they visit a page -- the sort of decision that the flaggedrev feature determines, combined with editorial work of updating the article.
Apostolis writes:
If wikipedia allowed articles to be forked and defined a trust metric that showed which article is more trustworthy, that would solve both previous problems and would also have contradictory ideas together, thus allowing people to have their own opinion about those different opinions and wikipedia wouldnt need to hide the strugle behind curtains.
http://en.wikipedia.org/wiki/Wikipedia:Wikipedia_in_research#Web_Researching...
Of course, this trust metric would have to be personalized, ie give different values depending on who the user trusts.
That's a slightly more radical proposal, but the basic idea of not hiding the struggle of assumptions and opinions from readers is important. Just as we try to recognize significant views within an article as neutral, we should recognize differing trust networks and opinions of reliability. This shoudl both make certain editorial work less burdensome, and provide more information [rather than forcing certain kinds of competing information to fight it out until one side is exhausted or defeated.]
SJ