Jump to content

Task force/Technology/2009-11-02

From Strategic Planning

[2009-11-02 12::54:14] eekim: hi everyone

[2009-11-02 12::54:22] freakolowsky: lo

[2009-11-02 12::54:30] eekim: wikimedia tech task force meeting will start here in about five minutes

[2009-11-02 12::54:38] » AryehGregor2 joined the chat room.

[2009-11-02 12::54:51] eekim: http://strategy.wikimedia.org/wiki/Task_force/Technology_infrastructure,_interface,_and_innovation

[2009-11-02 12::55:33] » mark joined the chat room.

[2009-11-02 12::57:04] eekim: hi mark!

[2009-11-02 12::57:12] mark: hi

[2009-11-02 12::57:27] eekim: sorry about pushing back an hour

[2009-11-02 12::57:33] mark: n/p :)

[2009-11-02 12::58:31] eekim: i'm going to go round up some folks from #wikimedia-tech and #mediawiki, then we can go ahead and get started

[2009-11-02 12::59:09] AryehGregor2: I can only be here for 15 or 20 minutes, unfortunately.

[2009-11-02 12::59:54] » atglenn joined the chat room.

[2009-11-02 13::00:14] » Rygir joined the chat room.

[2009-11-02 13::00:19] eekim: no problem, AryehGregor2

[2009-11-02 13::00:26] » PeterKaminski joined the chat room.

[2009-11-02 13::00:33] eekim: glad to have you here for that long; we'll post the logs, and hopefully you can engage further on the wiki

[2009-11-02 13::00:46] PeterKaminski: hey eugene!

[2009-11-02 13::00:53] eekim: thx for dropping in!

[2009-11-02 13::00:55] » rainman-sr joined the chat room.

[2009-11-02 13::01:03] Rygir: @topic > LiquidThreads is busted > why?

[2009-11-02 13::01:04] PeterKaminski: happy to -- thanks for organizing!

[2009-11-02 13::01:27] eekim: Rygir: we can discuss LiquidThreads and other relevant topics

[2009-11-02 13::02:39] rainman-sr: hello

[2009-11-02 13::02:58] vvv: Isn't it 21 UTC already?

[2009-11-02 13::03:03] eekim: yes it is

[2009-11-02 13::03:07] » flyingparchment joined the chat room.

[2009-11-02 13::03:13] eekim: thanks vvv. let's go ahead and get started :-)

[2009-11-02 13::03:15] AryehGregor2: It's 21:03.

[2009-11-02 13::03:40] eekim: so as background for these conversations, take a look at: http://strategy.wikimedia.org/wiki/Task_force/Technology_infrastructure,_interface,_and_innovation

[2009-11-02 13::03:49] » LauraHale joined the chat room.

[2009-11-02 13::04:01] eekim: hey laura!

[2009-11-02 13::04:09] » rand-away is now known as randmontoya.

[2009-11-02 13::04:10] LauraHale: Hello :)

[2009-11-02 13::04:15] eekim: wikimedia is developing a five-year strategic plan

[2009-11-02 13::04:21] » zrh joined the chat room.

[2009-11-02 13::04:29] eekim: tech obviously plays a really key role

[2009-11-02 13::04:38] » siebrand joined the chat room.

[2009-11-02 13::04:40] LauraHale: Makes sense.

[2009-11-02 13::04:58] eekim: the purpose of this (and future) meetings is to discuss the future of tech in wikimedia

[2009-11-02 13::05:07] rainman-sr: that's great, but what about day-to-day questions like who is going to take the two lead developer positions

[2009-11-02 13::05:12] eekim: and to develop 2-4 solid, well-researched and discussed recommendations

[2009-11-02 13::05:17] AryehGregor2: Two?

[2009-11-02 13::05:24] eekim: yes

[2009-11-02 13::05:37] » Skizzerz joined the chat room.

[2009-11-02 13::05:44] zrh: Do we have an agenda online?

[2009-11-02 13::05:45] eekim: the big picture has important implications on the day-to-day

[2009-11-02 13::05:58] flyingparchment: 4 recommendations to cover 5 years of technology changes?...

[2009-11-02 13::06:03] » Philippe|Wiki joined the chat room.

[2009-11-02 13::06:10] flyingparchment: did you look at how much our tech changed in the last 5 years?

[2009-11-02 13::06:22] » Raymond_ joined the chat room.

[2009-11-02 13::06:24] » Topic changed to "Discussion of Wikimedia Foundation's Strategy Project | http://strategy.wikimedia.org/wiki/Task_force/Technology_infrastructure,_interface,_and_innovation" by rainman-sr.

[2009-11-02 13::06:24] eekim: zrh: the agenda is to hear people's thoughts and to engage in some of the higher-level issues on http://strategy.wikimedia.org/wiki/Task_force/Technology_infrastructure,_interface,_and_innovation

[2009-11-02 13::06:35] zrh: eekim: THX!

[2009-11-02 13::06:49] eekim: flyingparchment: sure

[2009-11-02 13::07:12] AryehGregor2: Is there a live public log whose URL we can put in the /title for people just coming in?

[2009-11-02 13::07:42] eekim: Philippe?

[2009-11-02 13::07:47] Philippe|Wiki: indeed :)

[2009-11-02 13::07:52] eekim: thx :-)

[2009-11-02 13::07:53] » Mike_lifeguard joined the chat room.

[2009-11-02 13::08:04] Philippe|Wiki: workin it;)

[2009-11-02 13::08:07] eekim: on recommendations...

[2009-11-02 13::08:45] » cimon joined the chat room.

[2009-11-02 13::08:47] Philippe|Wiki: bah, where's my bot?! that's gonna make a lot a hassle.

[2009-11-02 13::08:56] mark: the same as we have now, times four

[2009-11-02 13::09:01] flyingparchment: can we convert that to http requests per second? it's not really clear what a 'visit' entails

[2009-11-02 13::09:15] eekim: mark: is that a real answer? :-)

[2009-11-02 13::09:20] mark: absolutely

[2009-11-02 13::09:24] eekim: where does it come from?

[2009-11-02 13::09:39] mark: from... me.

[2009-11-02 13::09:55] atglenn: how many visits a month do we get now?

[2009-11-02 13::09:56] » lc2 joined the chat room.

[2009-11-02 13::09:58] Mike_lifeguard: mark has a very good brain :)

[2009-11-02 13::09:58] Amgine: When you have an expert, listen to xyr.

[2009-11-02 13::10:21] vvv: What happend to LiquidThreads?

[2009-11-02 13::10:25] mark: iirc we handle about 25% of that traffic right now

[2009-11-02 13::10:44] eekim: when you say resources, do you mean servers? ops people? etc.

[2009-11-02 13::10:45] mark: if we were to grow that far, which I doubt to be honest, we wouldn't have problems scaling to that number

[2009-11-02 13::10:50] » dreimark joined the chat room.

[2009-11-02 13::10:52] mark: just servers

[2009-11-02 13::10:59] siebrand: vvv: it's there and it needs work.

[2009-11-02 13::11:05] mark: it doesn't require more people to manage 4x our current infra

[2009-11-02 13::11:06] Amgine: Pipes?

[2009-11-02 13::11:13] eekim: atglenn: about 330 million

[2009-11-02 13::11:23] siebrand: vvv: lqt has many issues, but the concept is nice.

[2009-11-02 13::11:24] mark: oh, only 3 times then ;)

[2009-11-02 13::11:26] Mike_lifeguard: mark: Really? We wouldn't need staff too? Seems to me we're already a bit short, given the current job openings. I'm not sure future expansion will manage itself. :\

[2009-11-02 13::11:27] AryehGregor2: Who makes it up, probably.

[2009-11-02 13::11:28] » dreimark left the chat room.

[2009-11-02 13::11:30] vvv: siebrand: why did the topic say that it was "busted"?

[2009-11-02 13::11:34] » peteforsyth joined the chat room.

[2009-11-02 13::11:39] mark: Mike_lifeguard: yes, but not to handle 3x our current load

[2009-11-02 13::11:43] AryehGregor2: Kind of silly for Wikimedia to rely on a third party to tell it how many visits its own site is getting. :)

[2009-11-02 13::11:53] Mike_lifeguard: AryehGregor2: We're silly people :D

[2009-11-02 13::11:57] eekim: vvv: there are some issues, but werdna is fixing them :-)

[2009-11-02 13::11:57] mark: we can easily scale up to 3x our current amount of requests, with no more resources than money to buy the servers and pay the bandwidth

[2009-11-02 13::11:59] » Marco27 joined the chat room.

[2009-11-02 13::12:00] mark: so that is a problem solved :)

[2009-11-02 13::12:10] flyingparchment: AryehGregor2: just convert those silly numbers into a ratio. 330m now, 1b target, is just x3

[2009-11-02 13::12:13] siebrand: vvv: topic here? Then i entered too late; not in anymore...

[2009-11-02 13::12:25] AryehGregor2: flyingparchment, true, the made-up part just cancels out.

[2009-11-02 13::12:30] flyingparchment: right

[2009-11-02 13::12:34] mark: and then there is the observation that the last 2-3 years, we've not grown a lot at all

[2009-11-02 13::12:39] » payo left the chat room.

[2009-11-02 13::12:54] mark: compared to our explosive growth in the years before

[2009-11-02 13::13:00] Amgine: That's slower than the internet.

[2009-11-02 13::13:05] mark: yeah

[2009-11-02 13::13:06] » payo joined the chat room.

[2009-11-02 13::13:11] AryehGregor2: Probably because everyone is using us as their definitive encyclopedia source these days, and the demand for encyclopedias isn't going to grow by four times anytime soon.

[2009-11-02 13::13:18] Amgine: So in effect, we're getting smaller.

[2009-11-02 13::13:39] AryehGregor2: Well, the fact that the Internet is growing faster than we are doesn't mean we need fewer extra servers. :)

[2009-11-02 13::13:40] atglenn: I wonder how commons usage will grow over time

[2009-11-02 13::13:43] » sarah joined the chat room.

[2009-11-02 13::13:48] AryehGregor2: Of course, if we grow slower than Moore's law . . .

[2009-11-02 13::14:02] sarah: Hi everyone

[2009-11-02 13::14:04] atglenn: compared to the pedias.

[2009-11-02 13::14:05] AryehGregor2: atglenn, bandwidth-wise, hosting lots of big videos would cause a huge amount of growth.

[2009-11-02 13::14:07] » Kunda joined the chat room.

[2009-11-02 13::14:10] » sarah is now known as Guest98456.

[2009-11-02 13::14:11] mark: feature and data growth might be another problem :)

[2009-11-02 13::14:12] Platonides: commons usage increased already more than expected

[2009-11-02 13::14:14] AryehGregor2: Although maybe not such a drastic growth in "visitors".

[2009-11-02 13::14:38] Platonides: to the point that some bots were asked to slow down to give time to get more space

[2009-11-02 13::14:45] eekim: okay, so taking a step back for a second...

[2009-11-02 13::14:51] Platonides: scaling file storage seems important

[2009-11-02 13::14:56] PeterKaminski: what if there's some disruptive change in the way wikipedia works -- like, it gets a lot more usable to edit, so number of editors grows much faster than previously (or some other disruptive change that would affect trends)

[2009-11-02 13::14:59] Amgine: mark: API as a percentage of over-all requests?

[2009-11-02 13::15:00] eekim: 1. how are we measuring hits right now?

[2009-11-02 13::15:15] mark: eekim: http://wiki.wikked.net/wiki/Wikimedia_statistics

[2009-11-02 13::15:21] eekim: 2. what would the needs be to scale linearly (i.e. current usage, just more of it)

[2009-11-02 13::15:27] Natalie: Features, like counting visitors.

[2009-11-02 13::15:35] vvv: Keep in mind that that there may be significant growth of China/India/Russia/etc Internet usage

[2009-11-02 13::15:36] eekim: 3. how do those needs change with changing usage (i.e. multimedia, etc.)?

[2009-11-02 13::15:37] Mike_lifeguard: Platonides: As I understand it, that is more a business-side problem than a tech-side... the budget didn't adequately anticipate the spending that would be needed

[2009-11-02 13::15:37] AryehGregor2: Natalie, we can count visitors now from Squid logs.

[2009-11-02 13::15:44] Natalie: AryehGregor2: Can or do?

[2009-11-02 13::15:50] mark: nr. 3 is the difficult question

[2009-11-02 13::16:00] AryehGregor2: Natalie, can, with no extra performance hit.

[2009-11-02 13::16:13] » Herr_Kriss joined the chat room.

[2009-11-02 13::16:35] vvv: AryehGregor2: do we really keep track of every visit?

[2009-11-02 13::16:51] Platonides: Mike_lifeguard, not anticipating the needs seems in scope :)

[2009-11-02 13::16:53] eekim: okay, so this is a good thing to note immediately. "visits" is more of a marketing metric. for the sake of ops, what we care about are http hits.

[2009-11-02 13::16:57] AryehGregor2: vvv, last I heard we have a full stream available, and store 10% samples or such.

[2009-11-02 13::16:59] Natalie: Mike_lifeguard: The budget doesn't really anticipate. People do. This is human error. :-)

[2009-11-02 13::17:11] » Romaine joined the chat room.

[2009-11-02 13::17:24] mark: eekim: correct :)

[2009-11-02 13::17:36] AryehGregor2: eekim, well, what those hits do is the really important thing. 80k req/s to Squids is okay, but even a small fraction of that that hit the backend would mean trouble.

[2009-11-02 13::17:44] Mike_lifeguard: Platonides: Well, it's not like there is a real technical challenge. It is moreso that the required hardware isn't in the budget. Though I have no idea why the budget would be so inflexible, especially considering that spending is way down

[2009-11-02 13::17:45] eekim: good point

[2009-11-02 13::17:46] Guest98456: Getting an accurate count of visits of real people on the different sites(as opposed to bots) is something that would be very helpful for the entire strategy planning process-

[2009-11-02 13::17:52] » Herr_Kriss left the chat room.

[2009-11-02 13::17:58] eekim: and this speaks to mark's point

[2009-11-02 13::17:59] » zrh left the chat room.

[2009-11-02 13::18:06] lc2: Guest98456: not when it comes to scalability

[2009-11-02 13::18:07] eekim: if usage changes, then caching strategy also probably needs to change

[2009-11-02 13::18:29] lc2: Guest98456: hits are all that count there, not visitors

[2009-11-02 13::18:31] Natalie: eekim: Why would you be worrying about caching strategy?

[2009-11-02 13::18:46] eekim: to figure out the resources we need

[2009-11-02 13::18:49] Natalie: That seems far outside your area of expertise.

[2009-11-02 13::18:50] AryehGregor2: eekim, yes, mostly in the form of "let's not let usage change". :)

[2009-11-02 13::19:00] Amgine: Natalie: strategy, in all its forms.

[2009-11-02 13::19:05] eekim: Natalie, i'm not imposing caching strategy. i'm asking about it. :-)

[2009-11-02 13::19:05] AryehGregor2: (I'm not the best person to answer most of this stuff, but I like to hear myself talk, so.)

[2009-11-02 13::19:21] eekim: but i'm not totally clueless in these matters either

[2009-11-02 13::19:25] AryehGregor2: eekim, fundamentally we're in good shape, because Wikimedia serves almost entirely static content.

[2009-11-02 13::19:26] Natalie: strategy is finding people who do know about it. Focusing on minute details (caching :: tech) seems silly.

[2009-11-02 13::19:29] » peteforsyth_ joined the chat room.

[2009-11-02 13::19:40] AryehGregor2: Serving static content can be made very, very fast.

[2009-11-02 13::19:45] eekim: Natalie, caching has implications outside of tech

[2009-11-02 13::19:53] Natalie: Really?

[2009-11-02 13::19:54] lc2: Natalie: not really a minute detail, is it?

[2009-11-02 13::19:56] eekim: yup

[2009-11-02 13::19:56] AryehGregor2: But it also means it has to actually remain static.

[2009-11-02 13::20:02] PeterKaminski: natalie, the amount of requests that can be served from the cache vs. those served from the dbs has a huge impact on resource planning

[2009-11-02 13::20:03] AryehGregor2: So no anon prefs, and so on.

[2009-11-02 13::20:08] LauraHale: caching can mean less erver load so more money saved.

[2009-11-02 13::20:09] eekim: we don't host databases outside of U.S. partially because of jurisdiction laws

[2009-11-02 13::20:16] Natalie: lc2: Yes it is. strategy is about finding people capable of handling cache and tech issues (and figuring out how to pay them).

[2009-11-02 13::20:20] eekim: when you turn a caching server off, the data goes away

[2009-11-02 13::20:25] Natalie: It isn't about trying to understanding cache infrastructure.

[2009-11-02 13::20:26] eekim: that's a high-level implication

[2009-11-02 13::20:28] Platonides: AryehGregor2, this is mostly static, but is in fact dynamic

[2009-11-02 13::20:48] atglenn: I expect that if we convert more readers ot editors and the percentage of writes goes up this will impact our serving strategy. depends how successful we are at it

[2009-11-02 13::20:51] PeterKaminski: so even though we won't figure out the details of the caching here, it's important to flag changes that could affect caching here

[2009-11-02 13::20:53] AryehGregor2: Platonides, it's functionally static, in that you don't need to do any script processing to service almost any requests.

[2009-11-02 13::20:56] LauraHale: I have a mediawiki site and caching makes some pages usable for our small memory. If you have a category with 100,000 pages... or Special:WantedCategories with 8000 pages? Load tme can take forever.

[2009-11-02 13::21:07] eekim: atglenn: definitely

[2009-11-02 13::21:18] Platonides: agree, but you can't completely forget that fact

[2009-11-02 13::21:20] AryehGregor2: LauraHale, category with 100,000 pages should be fine, because it's paged . . . more of a #mediawiki issue, though.

[2009-11-02 13::21:32] Natalie: PeterKaminski: Is there any science behind any of this? It all looks like guesswork to me. You can't possibly know Wikipedia's growth for the next five years.

[2009-11-02 13::21:43] eekim: Natalie: we can't know. but we can scenario plan, right?

[2009-11-02 13::21:52] flyingparchment: Natalie: you can estimate it using statistics and informed knowledge

[2009-11-02 13::21:52] Natalie: Not really.

[2009-11-02 13::22:01] AryehGregor2: votes for the recommendation "Let's forget about the future, just give us more money when we ask for it and we'll all be fine"

[2009-11-02 13::22:05] Platonides: atglenn, I don't expect to get a percentage of readers changed to editors

[2009-11-02 13::22:10] flyingparchment: Natalie: you can't work it out to a precise figure, but that doesn't mean it shouldn't be considered

[2009-11-02 13::22:13] Platonides: people will read much more than edit

[2009-11-02 13::22:13] eekim: AryehGregor2: duly noted :-)

[2009-11-02 13::22:16] lc2: AryehGregor2: i second that motion

[2009-11-02 13::22:27] rainman-sr: eekim, the only pattern i've seen is that in october we get a sudden jump in request, apart from that there is not much sudden growth that needs any extra ahead-of-time planning

[2009-11-02 13::22:43] Platonides: even in big numbers, changing 1% would be quite hard

[2009-11-02 13::22:44] Natalie: flyingparchment: I was asking if there's any math going on. :-)

[2009-11-02 13::22:47] Mike_lifeguard: rainman-sr: You forget about media uploads.

[2009-11-02 13::22:48] » peteforsyth_ left the chat room.

[2009-11-02 13::22:50] rainman-sr: (given there are no new features that will eat up resources)

[2009-11-02 13::22:53] lc2: oh hi Mike_lifeguard

[2009-11-02 13::22:55] Natalie: I agree there should be. Just not sure there is.

[2009-11-02 13::23:04] eekim: so that's what i want: numbers

[2009-11-02 13::23:08] rainman-sr: Mike_lifeguard, clarify

[2009-11-02 13::23:11] eekim: let's be rigorous in planning

[2009-11-02 13::23:13] atglenn: Platonides: maybe not, but it's one of the things people are working on. We might as well think a bit about that possibility

[2009-11-02 13::23:19] AryehGregor2: What sort of numbers?

[2009-11-02 13::23:19] eekim: and if those numbers already exist, let's document that

[2009-11-02 13::23:27] AryehGregor2: There are lots of numbers!

[2009-11-02 13::23:32] Guest98456: Platonides- it is definitely possible that we may see an increase in the editor to reader ratio especially if we can make improvements on the usability front

[2009-11-02 13::23:41] Mike_lifeguard: rainman-sr: Well, we already mentioned that upload bots on commons have been told to slow down because media storage is getting too full

[2009-11-02 13::23:46] Platonides: there might be more one-time editors

[2009-11-02 13::23:46] » StrategyBot joined the chat room.

[2009-11-02 13::23:56] eekim: okay, so we've established that there are a lot of numbers already :-)

[2009-11-02 13::23:59] mark: numbers: http://www.nedworks.org/~mark/reqstats/reqstats-yearly.png

[2009-11-02 13::24:01] » Illyism joined the chat room.

[2009-11-02 13::24:02] Platonides: but involved editors... that would be harder

[2009-11-02 13::24:04] eekim: thanks, mark

[2009-11-02 13::24:18] AryehGregor2: What specific numbers do you want?

[2009-11-02 13::24:18] mark: some more numbers...

[2009-11-02 13::24:20] Natalie: So we know there are a lot of numbers and something might happen in the next five years. Good start.

[2009-11-02 13::24:26] mark: around this time in 2005 we were doing 3000 req/s

[2009-11-02 13::24:35] AryehGregor2: Natalie, I don't know, do you believe that 2012 Mayan prophecy thing?

[2009-11-02 13::24:36] rainman-sr: Mike_lifeguard, yes, but that is easy to fix... you just tell bots to stop eating resources.. you cannot tell people to get easy with reading articles

[2009-11-02 13::24:39] Skizzerz: Mike_lifeguard: would setting a limit on how many old versions of images can be kept in the filesystem help with that?

[2009-11-02 13::24:40] mark: so we had massive growth then

[2009-11-02 13::24:48] Natalie: AryehGregor2: Sure makes things easier if you do believe, doesn't it?

[2009-11-02 13::24:50] Amgine: <remembers that spike>

[2009-11-02 13::25:05] AryehGregor2: Skizzerz, I expect we can just get more disks.

[2009-11-02 13::25:11] Amgine: Mark: In part that was the resolution of network bottlenecks.

[2009-11-02 13::25:12] eekim: mark: those are combined cached and non-cached requests?

[2009-11-02 13::25:19] mark: no

[2009-11-02 13::25:27] Amgine: Is it possible we currently have network bottlenecks.

[2009-11-02 13::25:27] lc2: Skizzerz: there'd be gfdl issues there

[2009-11-02 13::25:31] mark: eekim: green is cached, the black line is cache misses

[2009-11-02 13::25:35] lc2: probably issues with other licenses too

[2009-11-02 13::25:37] mark: i.e. what is not cached and goes on to the apache servers

[2009-11-02 13::25:41] Skizzerz: hmm, yeah. didn't think about licenses

[2009-11-02 13::25:42] flyingparchment: Amgine: i don't remember that being the case

[2009-11-02 13::25:42] eekim: thx

[2009-11-02 13::25:54] Mike_lifeguard: Skizzerz: That isn't a viable solution

[2009-11-02 13::25:57] AryehGregor2: lc2, you don't need to keep the old versions, just the names of the contributors.

[2009-11-02 13::26:04] Platonides: for the filesystem, move to filearchive and distribute based on hash

[2009-11-02 13::26:21] lc2: AryehGregor2: and, as i recall, you'd need to document the changes too, right

[2009-11-02 13::26:25] Natalie: AryehGregor2: There's no reason to delete the old versions, either. Disk space is cheap enough.

[2009-11-02 13::26:27] AryehGregor2: Just get more disks. Seriously.

[2009-11-02 13::26:31] eekim: mark, how far back do you have numbers?

[2009-11-02 13::26:32] AryehGregor2: I don't think it's a long-term issue.

[2009-11-02 13::26:40] LauraHale: AryehGregor2: What about citation issues for older versions?

[2009-11-02 13::26:57] Natalie: Oh, you were getting there yourself.

[2009-11-02 13::26:58] mark: eekim: just one and a half year

[2009-11-02 13::27:04] AryehGregor2: LauraHale, just keep them. What's the problem?

[2009-11-02 13::27:04] Natalie: Long-term strategy should focus on the ways to have strictest adherence to the GFDL.

[2009-11-02 13::27:10] AryehGregor2: I bet the overwhelming majority of media only has one version anyway.

[2009-11-02 13::27:15] mark: the rest was from memory

[2009-11-02 13::27:17] AryehGregor2: Natalie, CC-BY-SA these days, you mean, surely.

[2009-11-02 13::27:30] Natalie: AryehGregor2: That would ruin the sarcasm.

[2009-11-02 13::27:37] Amgine: <gg>

[2009-11-02 13::27:37] AryehGregor2: Ah.

[2009-11-02 13::27:44] Platonides: AryehGregor2, it has

[2009-11-02 13::27:59] cimon: is there any way to isolate computational needs of editing from those of access by casual browsers?

[2009-11-02 13::28:10] lc2: cimon: yes, squids

[2009-11-02 13::28:14] AryehGregor2: cimon, you mean like by Squid caching?

[2009-11-02 13::28:16] eekim: i missed the licensing conversation

[2009-11-02 13::28:21] lc2: AryehGregor2: k bai

[2009-11-02 13::28:24] cimon: I mean totally isolate them.

[2009-11-02 13::28:26] eekim: what's the issue?

[2009-11-02 13::28:33] Platonides: you could use a different domain for that

[2009-11-02 13::28:47] cimon: not just buffer or deflect

[2009-11-02 13::28:49] » StrategyBot|Away joined the chat room.

[2009-11-02 13::28:51] lc2: cimon: i don't see how you can get more isolated than that

[2009-11-02 13::29:15] Platonides: anonymous readers are almost never hitting the backend

[2009-11-02 13::29:24] Mike_lifeguard: eekim: nothing, it was a joke

[2009-11-02 13::29:28] » StrategyBot left the chat room.

[2009-11-02 13::29:35] eekim: Mike_lifeguard: got it, thx

[2009-11-02 13::30:26] Rygir: Since we're talking performance here, there should be a shift towards more ajax style javascript requests rather than full page refreshes, so how is that going to affect things?

[2009-11-02 13::30:28] Natalie: A decent mechanism to view article views.

[2009-11-02 13::30:39] » StrategyBot|Away is now known as StrategyBot.

[2009-11-02 13::30:42] mark: erik zachte does that

[2009-11-02 13::30:44] Rygir: which would complicate counting article views

[2009-11-02 13::30:50] cimon: What I am getting at is stuff edited changes things in rendering, blah blah blah...

[2009-11-02 13::31:00] Natalie: mark: Is it in an accessible format? Is it per-article? I haven't seen much.

[2009-11-02 13::31:03] » tomaszf joined the chat room.

[2009-11-02 13::31:06] mark: via domas his stats, I think

[2009-11-02 13::31:13] flyingparchment: Natalie: http://stats.grok.se/

[2009-11-02 13::31:21] » nimish_g joined the chat room.

[2009-11-02 13::31:21] mark: yes, it is available in various places

[2009-11-02 13::31:23] Natalie: flyingparchment: Right. But that's not really a good plan going forward.

[2009-11-02 13::31:25] mark: flyingparchment's link, and some others

[2009-11-02 13::31:29] Guest98456: One of the things that the India task force needs to understand is how many indians are looking at enWikipedia- right now from my understanding we don't currently measure anything that would aproximate that

[2009-11-02 13::31:30] Natalie: (Relying on an external site.)

[2009-11-02 13::31:55] flyingparchment: Guest98456: "indians" or "people in india"? one of those is much less impossible

[2009-11-02 13::31:57] Platonides: Natalie, http://stats.grok.se/ is based on domas stats

[2009-11-02 13::31:58] mark: Guest98456: being worked on

[2009-11-02 13::32:03] Natalie: It's also difficult to pull the data in aggregate form (and Henrik shouldn't really have to host all of this himself).

[2009-11-02 13::32:05] mark: erik zachte will have stats of that very soon

[2009-11-02 13::32:07] Natalie: Platonides: See above.

[2009-11-02 13::32:11] Platonides: it could equally be at the toolserver

[2009-11-02 13::32:11] Guest98456: people in India

[2009-11-02 13::32:16] mark: I wrote a program to help him generate that last week

[2009-11-02 13::32:27] » Guest98456 is now known as Sarah.

[2009-11-02 13::32:50] » Sarah is now known as SarahS.

[2009-11-02 13::33:35] eekim: what other stats would be useful?

[2009-11-02 13::33:59] Amgine: Everything about everything, unfortunately.

[2009-11-02 13::34:18] LauraHale: conversion rates for editors. Comparisons for traffic across a category and subcategories.

[2009-11-02 13::34:24] eekim: Amgine, we need to do better than that :-)

[2009-11-02 13::34:32] PeterKaminski: eekim: some way to get an idea about ratio about viewing vs. editing, and trending especially as the interface changes

[2009-11-02 13::34:44] SarahS: Another thing that has come up amongst some of the smaller language Wikipedia's is a desire to know what content people are searching for but not finding articles on- This could really help an emerging wikipedia community decide where to prioritize there article creation efforts

[2009-11-02 13::35:00] » Chicablog joined the chat room.

[2009-11-02 13::35:39] SarahS: mark: that's great- it will be facinating to see that data

[2009-11-02 13::35:50] LauraHale: The category thing is important for identifying areas that need work. Or better understanding strange traffic patterns.

[2009-11-02 13::36:22] eekim: we've identified several good areas for measurement so far and reasons for wanting to measure those things

[2009-11-02 13::36:25] PeterKaminski: eekim: it would be good to find other sites comparable to wikipedia to help triangulate trends, too

[2009-11-02 13::36:26] » peteforsyth left the chat room.

[2009-11-02 13::36:30] eekim: do we have the infrastructure right now to measure those things?

[2009-11-02 13::36:35] LauraHale: Data needs some context. (And having a way to export Wikipedia traffic data to Mediawiki for other wikis to analyze would be awesome.)

[2009-11-02 13::36:46] Amgine: But that's not really technical; a lot of that is marketing.

[2009-11-02 13::36:50] eekim: if not, what sort of resources are we talking about?

[2009-11-02 13::37:16] tomaszf: SarahS: we actually do keep search logs from our fleet

[2009-11-02 13::37:27] LauraHale: Usability is tech. ;-)

[2009-11-02 13::37:29] SarahS: Usability should definately be part of tech

[2009-11-02 13::37:38] tomaszf: it's just a hot button issue on what legal wants to release

[2009-11-02 13::37:41] eekim: presumably, if we want to improve the tools, we should have some data supporting that

[2009-11-02 13::38:01] Mike_lifeguard: LauraHale: Why would you want to export that data "to MediaWiki" ?? That sounds like a horrible idea, since MediaWiki is a wiki, and not a statistics program.

[2009-11-02 13::38:05] flyingparchment: usability is not tech, but implementing it is

[2009-11-02 13::38:11] rainman-sr: tomaszf, actualy we don't atm because the process died when searchidx1 got restarted

[2009-11-02 13::38:13] » Chicablog left the chat room.

[2009-11-02 13::38:15] flyingparchment: tech people are crap at usability

[2009-11-02 13::38:28] tomaszf: rainman-sr: easy one to fix :)

[2009-11-02 13::38:41] eekim: flyingparchment: it doesn't have to be that way

[2009-11-02 13::38:42] LauraHale: Mike_lifeguard: I'd love to have some of that data because I run a history type wiki on MediaWiki. Historical data ad trends that we could comment on? Win.

[2009-11-02 13::38:47] eekim: plenty of tech folks build usable tools

[2009-11-02 13::38:47] tomaszf: and we'll have a permanent place to stash them once the new data sets file sever shows up

[2009-11-02 13::38:50] » Amgine_ joined the chat room.

[2009-11-02 13::38:51] PeterKaminski: LauraHale: i like your idea of exporting wikipedia traffic data for other wikis to analyze. it would be good the other way around, too -- to have a mediawiki plugin that other wikis could install that would help collect trend data

[2009-11-02 13::38:57] eekim: step one is to care

[2009-11-02 13::39:01] LauraHale: But at the moment, most of the tools available don't allow for exporting the data or pretty charts.

[2009-11-02 13::39:01] eekim: step two is to be data-driven

[2009-11-02 13::39:06] flyingparchment: eekim: usually by accident. and i think MW is proof that that doesn't happen much around here ;)

[2009-11-02 13::39:13] Mike_lifeguard: LauraHale: So you don't want to dump data into a wiki then. As I thought. :)

[2009-11-02 13::39:19] eekim: flyingparchment: so what could we do to change that?

[2009-11-02 13::39:46] flyingparchment: eekim: UI changes should be prototyped and shunted to a UI team before being deployed. the UI people can decide if it's okay or what needs to change

[2009-11-02 13::39:55] LauraHale: Mike_lifeguard: No. More like analyze traffic volume of say Twilight articles on Mediawiki on a page like www.fanhistory.com/wiki/Twilight

[2009-11-02 13::40:09] eekim: who are the UI people? are they employed by the Foundation? should there be room for open source usability?

[2009-11-02 13::40:38] SarahS: tomaszf: what would it take to make the logs into a usable format that wikipedians on smaller sites could easily analyze

[2009-11-02 13::40:46] flyingparchment: eekim: i don't know what open source usability is. it doesn't really matter who they are, although it seems important enough to employ someone to do it

[2009-11-02 13::40:50] lc2: SarahS: anonymising them is the real issue

[2009-11-02 13::40:59] Mike_lifeguard: flyingparchment: Sadly, UI complaints have always and continue to be given short shrift in bugzilla

[2009-11-02 13::41:01] Platonides: The meaning of 'usable format' should be determined first

[2009-11-02 13::41:20] lc2: that and the fact that would require shit to be done, which the foundation are not too great at imo

[2009-11-02 13::41:21] Skizzerz: I think the UI should be handled at least in part by the Usability Initiative, since they're pushing ahead with skin features that make editing and such easier

[2009-11-02 13::41:33] eekim: flyingparchment: but the open source nature of the tool changes planning. if you look at the other top five web sites, they all employ thousands of engineers

[2009-11-02 13::41:49] flyingparchment: Skizzerz: but the UI team needs to be someone who can provide quick turnaround, since their verdict will delay deployment

[2009-11-02 13::41:53] eekim: should that change, and if so, by how much?

[2009-11-02 13::41:53] flyingparchment: Skizzerz: can they do that at the moment?

[2009-11-02 13::42:02] Skizzerz: flyingparchment: not sure, I'm not in it

[2009-11-02 13::42:07] eekim: these are questions that we need to start exploring

[2009-11-02 13::42:10] Skizzerz: but they're at least making progress

[2009-11-02 13::42:17] SarahS: Platonides; that is a really great point- usable to whom? A developer? Someone with a college degree? someone in a developing country who is just learning how to access the web?

[2009-11-02 13::42:46] Skizzerz: the point is, any "teams" we form for these purposes would have to be largely comprised of volunteers

[2009-11-02 13::43:01] » Amgine_ left the chat room.

[2009-11-02 13::43:04] flyingparchment: eekim: someone like Apple probably has hundreds of people working in UI in some way. if we had 1, that seems like a reasonable ratio

[2009-11-02 13::43:04] Skizzerz: since the Foundation lacks the resources to hire hundreds of people

[2009-11-02 13::43:20] eekim: Skizzerz: agreed

[2009-11-02 13::43:23] lc2: but does have the resources to move its offices to the most expensive place in the country

[2009-11-02 13::43:23] PeterKaminski: i don't know how they're arranged, but i think a good example of usability is firefox

[2009-11-02 13::43:26] lc2: feels good man

[2009-11-02 13::43:37] atglenn: we lack the resources today

[2009-11-02 13::44:02] eekim: atglenn: to hire hundreds of people?

[2009-11-02 13::44:02] atglenn: or if not hundreds, then, say, 50?

[2009-11-02 13::44:18] » Amgine_ joined the chat room.

[2009-11-02 13::44:23] LauraHale: FireFox plugins for editing pages in the sidebar would be nice. :D

[2009-11-02 13::44:30] eekim: PeterKaminski: very good point. we should invite some folks from Firefox to participate in these discussions

[2009-11-02 13::44:40] lc2: atglenn: and if we add those resources, what then? you'd have to get the code into mediawiki

[2009-11-02 13::44:43] » Amgine left the chat room.

[2009-11-02 13::44:50] » Amgine_ left the chat room.

[2009-11-02 13::45:01] flyingparchment: lc2: UI shouldn't be writing code

[2009-11-02 13::45:01] lc2: at which point satan puts on his sweaters, and the AA gunners man their posts and scan the skies for livestock

[2009-11-02 13::45:10] atglenn: we would want several of those people to turn into people who can do code review

[2009-11-02 13::45:16] tomaszf: Sarahs: to have concrete ideas of what people want to use the data for. This is especially true of low volume wikis where a person search habits can be traced back to an actual user. Properly anonymizing and removing any identifiable markers is huge. We wouldn't want to botch it like AOL did years ago.

[2009-11-02 13::45:18] atglenn: I don't mean 50 ui people.

[2009-11-02 13::45:19] » Amgine joined the chat room.

[2009-11-02 13::45:25] Mike_lifeguard: atglenn: Currently, we seem to be having trouble filling even the small handful of open tech positions :\

[2009-11-02 13::45:33] atglenn: I mean 50 people that do development/code review/qc/ops/etc

[2009-11-02 13::46:06] SarahS: If the tech group can articulate a clear rational for why the Foundaiton needs to dramatically increase the number of tech people on staff as part of the strategy process then we can start to figure out how we might obtain these resources

[2009-11-02 13::46:06] eekim: exactly

[2009-11-02 13::46:13] atglenn: so, another question: how can we attract better and more candidates?

[2009-11-02 13::46:14] flyingparchment: i thought moving to SF was meant to provide an infinite pool of skilled people willing to work for nothing...

[2009-11-02 13::46:27] atglenn: aa large pool of people, yes

[2009-11-02 13::46:31] LauraHale: Very true on low level of users and internal search traffic. But at same time, modifying ToS could also happen.

[2009-11-02 13::46:33] atglenn: *nothing

[2009-11-02 13::46:45] rainman-sr: eekim, i think the general rule of software development is that is difficult to solve problems with adding more people to the team, becasue it takes lots of time and effort to fit it... but it does help to involve the existing people more, and we do seem to have a constant influx of volunteer developers

[2009-11-02 13::46:52] flyingparchment: maybe we should offer stock options.. that's how other startups do it

[2009-11-02 13::46:59] lc2: LauraHale: it doesn't even need to be a low level of users

[2009-11-02 13::47:00] Mike_lifeguard: or maybe not :)

[2009-11-02 13::47:01] SarahS: tomaszf: good point- that's tricky

[2009-11-02 13::47:02] eekim: rainman-sr: how can we do a better job of that?

[2009-11-02 13::47:03] lc2: LauraHale: cf. aol search

[2009-11-02 13::47:26] Mike_lifeguard: atglenn: But in all seriousness, do we even *have* a head of tech right now? I'm not sure how you guys aren't panicked

[2009-11-02 13::47:28] » peteforsyth joined the chat room.

[2009-11-02 13::47:37] SarahS: maybe what we need are just broad catagories

[2009-11-02 13::47:45] » Raymond_ is now known as Raymond_afk.

[2009-11-02 13::47:45] roberthl: Erik is acting CTO.

[2009-11-02 13::47:46] atglenn: who said we aren't panicked? :-P

[2009-11-02 13::47:46] eekim: Mike_lifeguard: very fair question

[2009-11-02 13::47:51] lc2: lol erik

[2009-11-02 13::47:54] tomaszf: Mike_lifeguard: erik is acting cto and were actively hiring a new one

[2009-11-02 13::47:57] PeterKaminski: Usability is important enough to pay market rate for. i don't think you have to have a large team, but it should be dedicated people

[2009-11-02 13::47:59] eekim: the Foundation has been evaluating CTO candidates

[2009-11-02 13::48:00] cimon: lol at atglenn

[2009-11-02 13::48:03] eekim: but what role with that person play?

[2009-11-02 13::48:09] » Illyism left the chat room.

[2009-11-02 13::48:12] eekim: it will almost certainly be different than brion's role

[2009-11-02 13::48:22] rainman-sr: eekim, well not i'm not sure ... it would help to have a cto i guess, and approach people more often than just let them do their thing and eventually leave

[2009-11-02 13::48:29] atglenn: we want a senior sw architect + a cto

[2009-11-02 13::48:30] mark: what's there to panick? ;) it's not hard to just do one's job :)

[2009-11-02 13::48:35] atglenn: to "replace" brion

[2009-11-02 13::48:46] Mike_lifeguard: eekim: True. My understanding is Brion originally wanted to hire a boss, but he left before even *that* got done.

[2009-11-02 13::48:47] Platonides: brion wasn't just a one :)

[2009-11-02 13::49:02] atglenn: actually, it is, when we do not all have the institutional knowledge you have, mark.

[2009-11-02 13::49:16] Mike_lifeguard: Is the plan still to have a techy-businessy person at the top with a techy-code-architecture guy under that?

[2009-11-02 13::49:49] atglenn: afaik yes ( Mike_lifeguard )

[2009-11-02 13::50:15] eekim: i want to start wrapping up this discussion

[2009-11-02 13::50:39] lc2: eekim: good luck with that, i don't even know what the topic is anymore

[2009-11-02 13::50:40] eekim: this strategic planning process is an opportunity to define a five-year vision

[2009-11-02 13::51:16] LauraHale: Yes.

[2009-11-02 13::51:56] eekim: mark, you earlier said that the uncertainty of future usage makes answering the resources question hard

[2009-11-02 13::52:05] LauraHale: Where the level of connection to the foundation is not always clear.

[2009-11-02 13::52:09] GerardM-: you asked if there is a chance of a big increase in usability ... Google wave as an editing front end

[2009-11-02 13::52:13] eekim: can we define a few scenarios that we can start exploring and possibly pulling in data from other comparable examples?

[2009-11-02 13::52:14] mark: obviously, any new/big feature can change things dramatically

[2009-11-02 13::52:32] GerardM-: that gets you both real time collaborative editing and real text editing

[2009-11-02 13::52:48] LauraHale: Google Wave and Wikipedia aren't operating for the same purpose.

[2009-11-02 13::52:48] mark: that also means that the value of existing numbers is limited

[2009-11-02 13::52:51] GerardM-: also we are going video

[2009-11-02 13::53:03] eekim: GerardM-: that raises a meta-question: are we good at innovation? what changes do we need to make so that we can do things like integrate Google Wave into MediaWiki (for example)?

[2009-11-02 13::53:19] GerardM-: eekim who is we

[2009-11-02 13::53:25] eekim: mark: it gives us a baseline for comparison

[2009-11-02 13::53:27] LauraHale: GerardM-: You think that some one will make a public wave to compete with Wikipedia?

[2009-11-02 13::53:31] » vvv left the chat room.

[2009-11-02 13::53:35] eekim: GerardM-: the Wikimedia "community" :-P

[2009-11-02 13::53:36] lc2: eekim: your answer is 1) no and 2) rewrite mediawiki

[2009-11-02 13::53:42] GerardM-: LauraHale: it is not about competing

[2009-11-02 13::53:44] LauraHale: And that users always WANT that level of interaction?

[2009-11-02 13::54:12] GerardM-: eekim: as long as we are not afraid of not invented here... we are good at it

[2009-11-02 13::54:22] eekim: lc2: what incremental changes could we start making to start being more innovative?

[2009-11-02 13::54:23] lc2: GerardM-: which we are

[2009-11-02 13::54:32] GerardM-: Hmmm

[2009-11-02 13::54:40] LauraHale: GerardM-: Even if not competing, people will still have preferences towards certain types of content and interaction for certain topics.

[2009-11-02 13::54:46] lc2: eekim: i don't know about institutional, cultural changes perhaps

[2009-11-02 13::54:57] GerardM-: the fact is that only our own software gets priority to be validated

[2009-11-02 13::54:59] Mike_lifeguard: GerardM-: Consensus from everyone I've heard from on that idea is it is a fairy tale

[2009-11-02 13::55:36] GerardM-: have you seen it, used it ? or is it only hearsay ?

[2009-11-02 13::55:54] Mike_lifeguard: GerardM-: I said "everyone I've heard from" didn't I?

[2009-11-02 13::56:01] LauraHale: I'm just thinking that FanFiction.Net offers on a relatively simplistic platform and people have tried to compete with them, offering increased interactivity and more features... and users still flock to not as feature rich FanFiction.Net for content AND lack of interaction.

[2009-11-02 13::56:05] lc2: eekim: like no matter how obviously unsuitable mediawiki is for a project like commons, there's an attachment to doing things in a certain way

[2009-11-02 13::56:09] GerardM-: I have used it ... I LOVE real text editing

[2009-11-02 13::56:22] flyingparchment: what is "real text editing"? WYSIWYG?

[2009-11-02 13::56:29] GerardM-: yep

[2009-11-02 13::56:41] eekim: lc2: okay great. so that's another hard question: should we be absolutely tied to Mediawiki for everything?

[2009-11-02 13::56:50] LauraHale: Real Text Editing only seems like it would be an issue for high traffic pages.

[2009-11-02 13::57:00] eekim: and how should we decide when/whether mediawiki extensions is the way to go or whether to go with a completely different tool?

[2009-11-02 13::57:08] SarahS: real text editing would be a HUGE boost to increase participation in developing countries and amongst less tech savvy populations

[2009-11-02 13::57:10] eekim: several wiktionary folks have suggested that mediawiki is a real obstacle

[2009-11-02 13::57:14] GerardM-: I am editing articles that are low traffic and I hate it

[2009-11-02 13::57:14] lc2: eekim: i don't think it's a hard question at all, but getting to there from here isn't going to happen, ever

[2009-11-02 13::57:23] cimon: LauraHale, it is a huge issue at entry level newcomers...

[2009-11-02 13::57:25] LauraHale: eekim: Why not just fork Mediawiki like AboutUs has done?

[2009-11-02 13::57:26] eekim: lc2: perhaps. so try me. :-)

[2009-11-02 13::57:39] flyingparchment: LauraHale: forking MW doesn't fix anything, we already control it

[2009-11-02 13::57:52] atglenn: :-D

[2009-11-02 13::57:54] Mike_lifeguard: LauraHale: Yes, that'd be the problem.

[2009-11-02 13::58:01] LauraHale: cimon: My issues with Wikipedia are more cultural and complex templates than real time ability to update.

[2009-11-02 13::58:16] GerardM-: eekim: if you want to grow traffic then you have to grow the articles that people want to read

[2009-11-02 13::58:19] flyingparchment: LauraHale: the issue here is that WYGIWYG is *extremely* difficult to do in MW without either massive user-visible changes, or a significant loss of functionality

[2009-11-02 13::58:23] Mike_lifeguard: eekim: Why do you think omegawiki exists? MediaWiki is totally unsuited to a dictionary.

[2009-11-02 13::58:34] lc2: eekim: let's say if i went and created a compatible frontend to wiktionary tomorrow, where would it go from there?

[2009-11-02 13::58:37] GerardM-: flyingparchment: who says ?

[2009-11-02 13::58:50] lc2: it'd never be accepted, or it might sit on the toolserver or some shit like that

[2009-11-02 13::58:52] cimon: LauraHale, yours are what they are, but they aren't objectively paramount.

[2009-11-02 13::58:53] freakolowsky: flyingparchment: yes, why*

[2009-11-02 13::58:55] flyingparchment: GerardM-: if i'm wrong, please show me the working wygiwyg solution for MW - many people have tried, no one succeeded

[2009-11-02 13::59:00] eekim: lc2: i'd like to hear other people's answers to your question

[2009-11-02 13::59:15] lc2: flyingparchment: see wikia, they're wysiwyg now

[2009-11-02 13::59:15] flyingparchment: remember 2-3 years ago when wikia were a month away from a complete implementation? that worked so well..

[2009-11-02 13::59:31] LauraHale: cimon: There has been survey work on that point that REAL TIME EDITING is the big issue for new contributors?

[2009-11-02 13::59:32] GerardM-: I blog about it on mediawikiwave.blogspot.com

[2009-11-02 14::00:15] freakolowsky: flyingparchment: FCK works quite well ... for basic stuff ... and that's actually all you need ..

[2009-11-02 14::00:17] lc2: flyingparchment: unfortunately the real time editing model is very, very different and incompatible with the mediawiki model

[2009-11-02 14::00:22] GerardM-: results from the past are no indication of performance of the future

[2009-11-02 14::01:08] flyingparchment: last time i looked at FCK, once you edited a page with it, you could never go back to wikitext editing.. have they fixed that, at least?

[2009-11-02 14::01:12] GerardM-: anyway ... if you want to grow traffic, you need to write content that people want to read

[2009-11-02 14::01:38] freakolowsky: flyingparchment: letmecheck ... have it on a site

[2009-11-02 14::01:45] PeterKaminski: "real editing" can mean many different things. I don't think we have a good problem statement that defines what we might want from adding "real editing" to wikipedia, so it's hard to know what to do, or how the community might react to possible solutions

[2009-11-02 14::02:00] GerardM-: we do not know what people do not find ... so we cannot write what is likely to be popular

[2009-11-02 14::02:06] cimon: LauraHale, that is a red herring. survey work vs. what your isssues happen to be at any time and point is a laughable comparison.

[2009-11-02 14::02:37] eekim: okay everyone

[2009-11-02 14::02:39] LauraHale: cimon: I'm not talking about just my feelings. I've talked to lots of people about this.

[2009-11-02 14::02:40] PeterKaminski: flyingparchment: the FCK that speaks native wikitext works pretty well on MediaWiki. It leaves behind a little cruft like
and  's, but it's output is not terrible

[2009-11-02 14::02:46] eekim: this has been a very helpful first gathering

[2009-11-02 14::03:06] cimon: LauraHale, which is survey work?

[2009-11-02 14::03:11] eekim: i'll post the log to the strategy wiki as well as a summary of this conversation

[2009-11-02 14::03:24] LauraHale: PeterKaminski: That editor can have issues. :/ I've seen it destroy pages. We discontinued on Fan History because it would insert large swaths of html. :/

[2009-11-02 14::03:28] atglenn: we should try to get domas to weigh in at one of these too

[2009-11-02 14::03:30] eekim: i encourage everyone to continue engaging with these questions at: http://strategy.wikimedia.org/wiki/Task_force/Technology_infrastructure,_interface,_and_innovation

[2009-11-02 14::03:49] LauraHale: cimon: Your statement and mine both lack grounding unless you had some data to back your assertation.

[2009-11-02 14::03:52] freakolowsky: flyingparchment: works fine at a glance ... will give it a more detailed test tomorow, but for now my vote is for FCK

[2009-11-02 14::04:02] cimon: LauraHale, just leave it.

[2009-11-02 14::04:11] eekim: many thanks to everyone for jumping in!

[2009-11-02 14::04:23] PeterKaminski: eekim, thanks for hosting

[2009-11-02 14::04:25] freakolowsky: peep peep cheereeo

[2009-11-02 14::04:33] eekim: bye everyone

[2009-11-02 14::04:38] freakolowsky: bye