On 29 September 2010 21:07, Jimmy Wales jwales@wikia-inc.com wrote:
On 9/28/10 7:41 PM, Risker wrote:
Yes it is, and it's an important one. Several of us had already been working on a plan for the second trial, and those of us discussing had widely agreed that it would be much more likely to be successful if more
of
the recommendations on improving the software were incorporated, thus our recommendation that it not proceed so rapidly.
I respect what you are saying here, very much. But I think the right approach is always "release early, release often". There is no need to rush, but there is also no reason not to release fixes as they are available, because there is no particular "ship date" with marketing, etc.
Jimmy, here's where you're wrong. The first version was marketed as the solution that would allow the [[George W. Bush]] article to be publicly edited - it was marketed that way on and off wiki - and instead we had 40 hours of non-stop IP vandalism and browser crashes for almost every reviewer. (The first problem was easily anticipated by just about every administrator on the site, and the second one by anyone who'd already seen what had happened with other very large articles.)
This "product" has to be sold to admins to get them to use it; they saw the first version and all of its significant problems and aren't very interested. And until there is a product that passes their smell test, they still won't be interested. So installing an "upgrade" that hasn't resolved ALL of the significant issues is not going to interest the "consumers".
The advantage of a coordinated effort of a new trial with an upgraded release that has addressed all of the significant issues *and* has been well-tested on the test wiki is that it can be used to market the tool. It doesn't matter whether or not it works well if the people in the position to use the tool cannot be persuaded it is worthy of their attention. Take a look at the stats, Jimmy: Six administrators were responsible for entering 80% of the articles into the first trial, and another 12 responsible for the next 17%. Most administrators were not interested the first time around.
It's pretty hard to maintain motivation, though, when it's clear that the software's going to be a permanent feature regardless of what the project does or thinks, and that any further "trial" is not going to change that fact.
I think that's very very far from true. I think that everything the Foundation has said, and everything that I have said, and everything that (nearly) everyone on all sides has said, indicates nearly 100% universal agreement that in order for the feature to be enabled permanently, it has to achieve consensus.
Consensus is not a "hold one vote and give up if you don't make it" process, but rather an iterative give-and-take.
If I believed that the current version was the best that the Foundation could deliver, I would be adamant about just shutting down PC as soon as is practical, and believe that the right way forward would be to push for major expansion of the use of semi-protection. I would hate to do that, because I think that a well-implemented PC is a better solution than semi-protection, striking a better balance.
My point is this: I think it very far from a foregone conclusion that we will have PC in use in the longterm. It has to improve a lot before that can happen. The early signs, though, are that it was popular.
I'm really curious to know what metric you're using to determine that it was "popular". The *idea* is popular with a significant segment of the community, which is where much of the support in the two polls came from; but the *tool* itself wasn't very popular with many editors. And the concept of administrator-granted "reviewer" permissions went over like a lead balloon with a pretty big segment of the community.
Put the upgrades on the test wiki. Recruit a pile of editors (not just administrators) to really put it through its paces and drive it hard, both those who are technically savvy and those whose strength is content. These editors are your potential change agents; if they're convinced it's working satisfactorily and that major issues have been resolved, they will spread the word on-wiki. Sticking poorly tested software upgrades onto the #7 website, and expecting people to be enthusiastic, is remarkably optimistic.
Risker/Anne