One of the things that I’d been intending to this summer was get back into Wikipedia and the wider Wikimedia arena, after having to go on wikibreak for all of this school year for simply not having the time. A few weeks ago as school wound down to the extent that I was barely going in due to so many subjects having finished I started gradually working myself back in, but found it difficult and not particularly entertaining: it’s hard to get back into the swing of things and I didn’t fancy working through the backlog of things under my wikibreak notice. But I went for it and now it’s great to be working away quietely. As a friend points out however I’m only going to become inactive again from September. But my workload will be a lot less next year because I won’t have my heavy-duty history since I’ve now dropped that subject. So we shall see.

I started off a few days ago by getting back into recent changes patrol. This involves watching the feed of edits to Wikipedia and opening and reverting if necessary problematic, nuisance ones we refer to as vandalism. I started off making good use of my new dual-monitor setup (which I will get round to blogging about) by having CDVF display a flitting list of changes on my smaller screen, with Firefox open on the other. It went reasonably well and I did a fair bit of vandal-whacking; reverting edits and posting warnings to the talk pages of the vandals in question using a pretty little script called Twinkle and if necessary blocking the users when they clearly weren’t going to be constructive. This is the old-school way of patrolling, although my use of scripts to speed some things up would probably be scoffed at by the really nostalgic Wikipedians out there. However, after a bit I found myself being repeatedly out-reverted by users of Huggle, an app that tries to make things even more automated, even faster, getting those vandals down in mere seconds. At first I kept going with my old-fashioned methods (that always used to work) but it was soon clear this wasn’t going to last much longer: I was out-reverted four times on one article by someone using Huggle (it was being vandalised repeatedly). Clearly, I needed to try Huggle if I was going to be a useful patroller.

So I booted over to Windows (the app won’t run in Linux, it uses .net - don’t worry if that doesn’t mean much to you…) and downloaded Huggle, read the instructions and got going. And got very, very scared. The sheer speed of the whole thing makes me worry about how many mistakes can so easily be made. Firstly, you work with two buttons in the main: one to advance to the next edit, one to revert the edit and warn the user generically. Other options are to leave messages (such as encouraging anonymous users to make an account if they are being a decent contributor) and to revert for other reasons, such as the introduction of copyrighted material grabbed from somewhere else on the web. Secondly, all of the above can be done with keyboard shortcuts. So with one hand on space to move on, and one on Q to revert and warn (and I note that such edits are queued up and done in the background for you without interference, so you can shoot ahead assuming you have a decent Internet pipe), I was rattling through hundreds of edits.

As I mentioned earlier, mistakes are very possible. Several times someone else got there before me but I had already hit Q as I could see their revert appear on the screen. So Huggle obiediently reverted their reversion (restoring the vandalism) and warned them for unconstructive edits. Eeek. Frantic scrambling to Firefox to apologise ensued. More simply, one finds oneself getting subconciously into a race to work through the edits and revert ‘before someone else does’. I caught myself thinking like this a few times and took a deep breath.

Another scary thing about this program is how it progresses from warnings to blocks. In general people get four warnings for vandalism, with wording decreasing in friendlyness, before they are eligible for a block, starting at for many of us admins one day, and then repeated violations gaining weeks, months etc. in an attempt to get people bored with the whole practice by not being able to do it for a while. Huggle takes this to the next level. If you hit Q when the user already has four warnings, a form comes up to block the user immeadiately. You get options for length, softness/hardness (if you block an IP address (i.e. an Internet connection rather than a logged in user) you can elect to disable account creation or not, and various other things), and a table displays the warnings, times of said warnings and who they were left by, including the various automated anti-vandal bots we have running. Okay, I thought, this is maybe just a little bit far. I was blocking someone without ever having seen their talk page or contributions beyond being told they had the warnings. But what else was I to do? The evidence was there that a block was appropriate and so I set it. But I was left feeling somewhat uneasy.

In any case the app is very impressive with everything that it can do (I’ve only talked about a small set of its features here) and I congratulate the author, Gurch. If used correctly it’s a great way of relieving the monotony involved in keeping Wikipedia vandalism-free. My only concern is that I would perhaps prefer a few more seconds of vandalism existing than the potential recklessness. But maybe I was just an abusive Huggle user. In any case, I shall now reveal my levels of geek with the fact that I was absorbed in my keyboard shortcutting for roughly two solid hours yesterday afternoon.

I’ve also got back into some admin-specific tasks. Speedy deletion is the process by which articles can be nuked without the usual consensus-building discussion when it is clear we don’t want them; for example blatant advertising or articles containing only nonsensical gibberish can be deleted on-sight by an admin. I’ve always found this to be a nice, useful little time-filler so I got going, and ended up creating controversy. Wow. Straight into the thick of it by spawning several sides of discussion on the administrator’s noticeboard over my deletions. This never normally happens to me! I’m normally very much under the radar on Wikipedia, only doing silly little things that keep me entertaining that aren’t often noticed. But this was a change. I’ve also been doing other such tasks. I spent an hour and a half or so this afternoon working my way through the Articles for Deletion backlog, that is closing discussions and determining whether or not there was a consensus to keep, delete, merge etc. So I’ve been keeping busy and being reasonably useful to the encyclopedia.

I worry, however, about my whole outlook towards things with the project: I’ve never been fantastically good at just getting on with jobs unless they are new (novelty value, which I am getting now since I haven’t been around for so long) or are exciting, such as answering press enquiries coming into Wikimedia. We all know of the right people to call on when things need doing and I’m probably not on anyone’s list (although I most definately have a personal list of *cough*slaves*cough* myself). I need to find useful things to do that I can beaver away at.

All in all, it’s good to be back on what still is a fantastic set of projects, for all its faults.

It seems my recent reminiscence is shared by a friend David Gerard, who pointed me in turn at Clay Shirky, one of Wikimedia’s advisory board members. As David claims he was, I was mentally nodding as I read through, as the patterns that are described for online communities are ones that I think I have been subconciously noting for some time, as I’m sure many have. However, the theme seems to be that things need to be done before a community is set up in order to avoid problems later on. While Jimmy probably envisioned that a community would arise surrounding the encyclopedia project, it seems unlikely that he and his fellow starters would give it much thought as that was not what they were aiming for. If only Shirky’s essay had been written a little earlier, people might start to notice the tell-tale signs of things going wrong before they got to the point of no return. While Wikipedia trundles on and will probably keep doing so in its current form, it could be so much better if we were to sort out a lot of the problems that the community faces. But either people are powerless to do so (and likely don’t know what to do either, like me), or those who do have that power haven’t got a clue what to do about it, however much they want to.

In the comments following his post David seems to think that the best thing to do is to put the developers in charge as an ultimate authority (something we don’t really have at the moment). He claims they already have a great deal of power anyway. While those who administrate our servers will always have the ability to “pull-the-plug”, for some it will cost them their jobs and really, we are not going to have a developer revolution. Yes, we depend on software to keep us going, and yes the developers control the code and the implementation of it, and yes they have a well-established cabal like the rest of us. But I don’t think an automatic shoulder-load of power to them makes sense because a good coder does not necessarily equal a good leader. Don’t get me wrong here, I have great respect for the devs and what they do (you have to be a really good coder or sysadmin to navigate the mess of our servers, apparently) but at the same time you can’t be good at everything at once, generally speaking. A technocracy isn’t fair.

The problem is that I don’t have any solutions to the problem either, and it seems no-one does. So people pull back from the encyclopedia with a “screw this” attitude into their cabals, they work away at either things above at Foundation level, they move to other Wikimedia projects to write dictionary entries or news articles instead (this is fine, but it may not be what they want to do), or they hide away with articles that get little attention and work quietely. People are only going to keep seeing the attraction of this, with closed mailing lists and IRC channels being so much more pleasant. When you are in one or two the outside community starts to look a lot worse than it used to, as you realise the alternative. freenode staff and helpers have had this problem recently as I describe here, because our own private communities are a lot nicer than the outside, public social channels. Unless we do something soon, more and more will retreat up and out - including me.