[Foundation-l] Sue Gardner, Erik Möller , William Pietri: Where is FlaggedRevisions?

William Pietri william at scissor.com
Mon Mar 1 04:24:00 UTC 2010

On 02/28/2010 07:32 PM, MZMcBride wrote:
> When it's your biography that reads you once were convicted of murder or
> pedophilia or whatever else, then you can start talking about people being
> wound too tight. When it's only been a delay of a few weeks, then you can
> talk about which forum should be used and so forth.

There's no need to persuade me of the value of Flagged Revisions. I 
already think the project is important, or I wouldn't be working on it.

My point is that drama will slow things down, not speed things up. My 
long experience is that people swearing at programmers impedes progress. 
You should decide which you're after. I figure it's progress, which is 
why I mentioned it.

> What I see is literally zero activity on that site since December 17, 2009.
> All of the tasks appear to have been created on December 16 or 17 and nearly
> all of them are in the "Deliver" phase, which reads to me as though they
> haven't been done.
> I did get the software to output "Found 32 stories (93 points total, 0
> points completed)" for the user JAS and the "Done" button at the top opened
> an empty box.
> Point to me what I'm missing.

Seeing a "Deliver" button means that Aaron, the developer, thinks the 
item is done, but it is not yet visible to others. Once we have a test 
server where people can look at things, then they are delivered. When 
some non-developer (e.g., me, or Howie Fung of the usability team) 
verifies that they are actually done, only then do we mark them as done.

> Production wikis like... the German Wikipedia? What the hell are you talking
> about? Update flaggedrevs.php for the enwiki database, sync it to the
> servers, and let's see what happens. How does that sound?

Like a recipe for breaking one of the world's top ten websites, an 
outcome I would rather avoid.

There have been substantial changes to the code. We don't want to break 
either the English or German Wikipedias, so we test before shipping. 
This is not an unusual approach to running a production web site. 
"Measure twice, cut once," works even better in software than carpentry.

Also, the community doesn't yet believe the software is ready, at least 
judging by the last round of feedback on the labs site. The usability 
team and I agreed with that, as did others, which is what motivated this 
latest round of changes.

As important as it is to get FlaggedRevs out for the community to try, I 
think it's even more important to release it in a form that will yield a 
successful trial. If we release something that's not up to snuff, the 
community may reject it for reasons that have nothing to do with the 
actual idea, an outcome nobody wants.

> When might that be? Is there a specific deadline? If not, why? And if there
> is a deadline and it slips by yet again, what's the consequence to those
> running the project?

There is no specific deadline. The approach I thought best for this 
project was one where we measure actual progress and use that to project 
dates. (That's why I used Pivotal Tracker, a tool designed for tracking 
and measuring real, fine-grained progress.) I explain more here:


It's a pretty standard approach in any of the [[Agile Software 
Development]] processes.

As soon as we can release to labs and check out the new stuff, which I 
ardently hope is soon, we'll have some useful data on productivity. If 
everybody feels the new version is ready to go live, then I am not aware 
of any impediment to public release right after that decision. If, as 
seems likely, there are some further proposed changes, we'll be able to 
estimate development time and project dates.

As to consequences, we all serve at the pleasure of Danese Cooper most 
directly, and to Erik, Sue, and the board from there, so if they think 
we're doing a bad job I'm sure they'll deal with that.

However, in my experience everybody involved is smart, talented, and 
very committed to the mission of the Wikimedia Foundation. Everybody is 
also keenly aware that this is a high-profile, high-priority project. 
Menacing people like that with "consequences" mainly serves to destroy 
motivation, not create it, so if you're truly interested in getting this 
done, I ask you not to do that again.


More information about the foundation-l mailing list