G'Morning!

Woke up this morning feeling very refreshed; my first full

night of sleep in over a week, I think. Sang in French in

the shower, took a taxi to city center and then wardrove on

the bus to WSIS - as I suspected, there were not many APs,

except at the fancy hotels. There were even one or two WSIS

access points in downtown, which surprised me. The guy next

to me was a 28-year-old Bangladeshi who had just won an

award from WSIS for teaching a group of 15 kids about

technology and computer science. Apparently network

connections there are not very good - there are no submarine

cables to speak of, so everything's over VSAT. Local

infrastructure (LANs & WLANs) are pretty good, though,

which gives me confidence that once these guys get a good

backdrop, their country will be pretty much set, save for

the kind of political issues that have plagued other

single-drop, corrupt-government countries like Fiji and

Ghana. I managed to get a GPS fix, woot. That was fun.

TunisMed

I then checked out the parallel TunisMed conference just

outside of WSIS and was utterly boggled by its point. There

were trinkets, casinos, and a snack bar, in a wildly unthemed

hodgepodge of companies operating in or vaguely related to

the Mediterranean. (Hence the name.) I'm mildly disgusted by

the fact that there were game companies advertising casino

machines in an event run in conjunction with a summit to make

an information society. I left quickly for WSIS.

War Walking WSIS

By the way, I noticed when doing a warwalk of WSIS that there

were a few hundred dummy SSIDs, perhaps to thwart some kind

of wireless attack? It was not at all clear to me what those

were doing - they had names that looked like buffer overflow

attacks to me. See the

Netstumbler log for the results of my warwalk @ WSIS.

Wikipedia

I then strolled over to the Wikipedia presentation, in a

midsized room with maybe about 80 attendees.

Jimmy's theme for Wikipedia:

"Imagine that every human on earth is given access to the

sum knowledge of humankind." Funded by public donations,

small (50-100 euros). Now partnering with companies, like

gifts of servers from Yahoo, etc.

Wikipedia is using a new local chapter strategy.

The encyclopedia is organized by language but local

chapters are useful - Germany, France are active. Italy

and Poland coming online and other chapters are forming.

Wikipedia is freely licensed. It's now large - English version

has ~800,000 articles. (500m words!) Larger than Brittanica

and Encarta combined. German equal to Brauchaus. Over

100,000 articles in many major languages. 2.5m+ articles in

~120 languages.

WikiBooks is the most exciting part of Wikimedia foundation

to Jimmy. Wants to give the set of literature needed to get

the populace to a level of literacy where the encyclopedia

in their language actually makes sense.

Wikimedia Commons: generally language-neutral: images, sounds,

and video. All freely-licensed and easy to use. Infrastructure

point for the free culture movement.

Wikimedia foundation has TWO employees: Jimmy's assistant and

a paid fulltime developer. Everyone else (including Jimmy) is

volunteer.

WP is now a top 40 website.

Reach is more broad than NY Times, WSJ, MSNBC, Chicago Tribune,

all combined. 2.4 billion pageviews monthly.

No business deals or need to closely watch stats (other than for

amusement). This is about 2x about.com. (ed: whoa!) About.com

was sold for $410m!!! (ed: wikipedia then worth ~$1bn?)

>120 servers in Paris, Amsterdam, US, South Korea. All managed by

volunteers. Would be $ to do this if had to hire and pay volunteers.

Relatively small group of loud volunteers. Are these loud people

actually the contributors, or is it the case that a million people

are writing a sentence or two?

Tested hypothesis - 80% of the work by 20% of the people? No, more

concentrated than that:

• 50% of the edits by 0.7% of the volunteers
• 75% of the edits by 2% of the volunteer
• 746 people had done 72.8% of the edits

So the noisy people ARE the contributors, as it turns out. Not

really millions of strangers adding to Wikipedia sentence by

sentence.

Secret to Wikipedia - realtime peer review. Recent Changes page,

watchlist, versioning, easy reversion.

Can use people to replace software, not vice verse - e.g., Votes

for Deletion. Good to not automate - minority may have a better

reasoned argument. (ed: more like the US judicial system than

the US legislative system!) Tough debates surprisingly not over

the George Bush page, but instead more on which picture of

Eiffel Tower to use.

Mild aristocracy within community - reputation empowers.

Mild monarchy - Jim. Cited "I am the Queen of England" in German paper.

NOT a benevolent dictator model - deeply inappropriate for a

compendium of all human knowledge. Over time, power being reduced.

Neonazis tried to get some pages deleted by ganging into

Wikipedia and voting for deletion certain pages - they failed,

but what if they had succeeded? Answer: Jimmy blocks. Gives a

core to the core values. Flexible about social methodology,

value results more than methodology.

Freedom of speech critical for all cultures. Strong correlation

between successful wikipedias and cultures with free speech.

(ed - Of course, but wow! Neat to see.)

Consequence of silencing citizenry prevents them from producing

information, not just consuming information.

Wikipedia just blocked in China. Was happy to not be blocked

before, has assumed to be error. HOPING that this is just an

error, but not sure.

Affect of Chinese blocking on Wikipedia - Chinese can no longer

contribute to Chinese wikipedia (edits to cn wikipedia fell by

half since blocking). Mainland Chinese vs Taiwanese standpoint;

mainland voices are silences, so Wikipedia possibly developing

systemic bias towards Taiwan.

(ed Q: Doesn't this produce increasing justification?)

A: Perhaps, but we hope that not, due to Neutral Point of View

policies within Wikipedia. And we hope that the Chinese ban is

lifted shortly. NPOV is empotionally difficult but is what makes

a "good Wikipedian".

Conclusion? Repression of opposition ends up under representing

one's own point of view!

The multilingual aspect of Wikipedia is trying to give

minorities ability to express their viewpoint.

(ed Q: but if the view is in their language only, won't that

end up marginalizing their visibility to the average person?)

Cross-language surfing - different languages bring different

perspectives.

(ed Q: But doesn't this only work for multilingual readers?)

(ed Q: Doesn't this end up spreading the knowledge thinly

across multiple languages, making it impossible to get the

full perspective?)

Exmaple of translation of the week: broadcast to volunteers a

request to translate a given page into as many languages as possible!

Folks are now adding pronounciations of people's names in

various languages and common phrases in various languages.

Wikisource - public domain text, xlated multilang.

Two years ago, Wikipedia got Arabic version. Now ~7500,

hopefully 10k articles in next few months. Public domain Arabic

works getting added to Wikisource. An Arabic wictionary is out.

Arabic spoken in 25+ countries. This covers a diverse group of

speakers, differing wildly in opinions about many topics.

(Politics, religion, history) Discussion pages are very useful.

Amazingly enough, there are no Brittanica-like encyclopedia in

Arabic, so a successful future Arabic Wikipedia would be the

first Arabic encyclopedia of quality!

David Q: Encyclopedia gives one authoritative place to go, but

different projects and languages end up diluting the real value

of wikipedia - creates cultural walls around languages.

A: Well, people translate articles, so the good bits end up

getting propagated to all wikis. Also, some stuff may end up

coming across in different ways in different languages and

that's good. And many people can only contribute in their own

language.

(ed: I'm not satisfied with this, as most people are having to

learn English for other reasons anyhow, and whatever is lost by

forcing people to a single language that's not native is more

than gained by offering a SINGLE point of discussion and

interchange - in fact Wikimedia gave the example of Lithuanians

and Estonians banging out an agreement on certain articles in

the English wiki! Having other languages and other projects

like Wikinews hurts Wikimedia more than it helps.)

Wikipedia has made some decent amount of money in licensing

fees from Wikipedia trademark, such as in Germany. (ed: wow!)

How we get enough servers to keep the site running

happily is main issue. Monetary needs pretty small.

Just need to pay the bandwidth bills, get money from

public for that. Volunteering is main way this works.

(e.g., Red Cross / Crescent)

BBC News Q - Are there folks who may want more ($) from this?

A - career benefit, good on resume, better writer, better

tech. Will ppl get bored? Jimmy hopes not.

A Tunisian says - Arab world not homogenous. Not a good idea to mix it.

A - IS a good idea to mix! Need everyone involved to avoid biases.

Q - why no to ads?

A - Jimmy opposed, but won't say never.

Wants to leave things openended. Don't need the

money. If we're truning odwn millions a month turning

down oppty to distribute wikipedia to developing world

would be dumb. But right now, ad revenue would damage

grants and perceived independence. (eg, could never

take out negative comments about Yahoo - laughter)

For forseeable future, no ads.

Audience member - Ads change perspective, absolutely.

Audience member (Goethe Institute) - UNESCO affiliation possible?

A - yeah. but we're crazy Internet people (laughter), so our

influence is bigger than our capacity to network.

Hard as one guy out of his house to call up UNESCO!

Q - Biases in wisdom of wikipedia toward the literate

and the young, but wisdom found in the old and often

the illiterate. They can't contribute.

A - We'd love to have hundreds of Nepalese wikipedians

but we can't just yet. UNESCO partnership would be

helpful in this case. Example of Malian volunteer

putting in spoken knowledge into Bambalan / French

Wiki. Japanese wiki folks discuss, then change,

Americans do the opposite. (laughter) Negative version

will never be implemented, eg "not allowed to post if

you're not approved by elder".

Q - What if there were to be no figurehead for Wikipedia?

A - That's our hope. Maybe we'll need figurehead for

parades (laughter), but I hope not. Easier for the

Brits to grasp than Americans (Americans ask - where

is the Constitution? Brits used to process.) E.g.,

already with German wikipedia. (Germans have to work

it out themselves, since Jimmy doesn't speak German.)

A2 - We cannot handle the organization like a wiki.

It's inefficient. If a server goes down, it needs to

be fixed, shouldn't have to wait for the community to

agree to fix it.

Q - Why is there no disclaimer on Wikipedia? This

bugs researchers. This process can't be authoritative.

Peer review is much better for accuracy, and is what

actually happens with Britannica - can take it for

fact. Can't do that with Wikipedia. Something edited

by the public.

A - Deep question about Wikipedia & what it means to

be authoritative. Within the community, views are

traditional (accuracy, sources, neutrality). Difft

from standard processes, but same values. Better

than Britannica in some areas than worse. "People

will look at Britannica and say 'This was written

by one person and reviewed by two? How can I possibly

trust it?'" Not like a million random websites;

ongoing process of improvement. Not yet up to

traditional articles in encyc, but rapid improvement.

If we keep improving, hopeful of achieving very high

quality. Objective - Britannica or better quality.

A2 - We have academics who contribute, eg Wikisource.

Important to challenge academics.

A - Multiplier effect of acadmics and professionals

with volunteers. E.g., ornitholigist monitoring

bird articles by amateurs, just making sure they're

correct but don't have the time to add themselves.

Q - How do you get feedback form community?

A - It's a pretty noisy community. (laughter)

IRC, email, discussion boards. 40 IRC chans!

Q - (from University of Lille, France professor of linguistics)

When she doesn't know, tells students to use Wikipedia.

Q - BBC News. Was at Google, heard "even bad information is

good information". (ed: talking about hate pages)

Is the same true for Wikipedia?

A - We don't like bad information, but we expect to

have good information about bad topics. Attempt to

be tasteful. Not that titillating. Different from

search engine (also useful) - we don't post rants.

Q - quantitative measures for tracking neutrality?

A - Um, no. Hard to do. Involves human judgement.

C't did assessment of German Encarta, Brachaus, and

Wikipedia for a few articles and wikipedia fared

well, but very hard to do. Once computers can assess

accuracy or quality, we'll just let them write the

articles.

Q - Autodeletion of low-use articles to trim?

A - Um, no. No auto-deletion ever. For next version

of software, simple system for rating articles. We

plan to at first just gather data. Might not even

publish the data. (Where are we getting 5s? 1s?)

A system for identifying bad content would be

helpful. Even one page view a year is still useful.

Plenary Session

I then headed over to the plenary room. Most of the

talk was by very, very important and powerful people

who clearly have no idea about what the Internet is,

why it works, or how they'll benefit. They use very

big and generic words to say pretty much nothing, and

everyone claps. Most of the language concerns how

great everything is going and how important the summit

is. People are happy to feel important. Poor countries

tend to talk about how they don't like being poor, and

repressive regimes tend to talk about how dangerous

the Internet is.

Saudi delegation says it should be "WECANN" instead of

"ICANN". Out of technical considerations, I'm sure.

Libyan minister says the internet is used for plane

hijacking and needs to be regulated. People laughed

incredulously. What crack is he on?

Venezuala compares American control over the Internet to

the Spanish control of sea routes in the 19th century.

This is wildly off, since the Spanish actually got a cut

of the traffic, but the US government does not profit

from its oversight of ICANN.

Wireless pretty bad in most places in the conference, get

kicked off all the time, except for in the plenary hall,

where it's pretty decent.

Thoughts About Tunisia

Tunisia has pretty tremendous identity issues, due to it

history and location. It's in Africa, but its members

don't look or act African. It borders the mediterranean

but doesn't feel equal with Italy or Greece. It's populated

by Arabs, but due to Tunisian language being a mix between

Arabic and French, other Arabs don't understand Tunisian,

making Tunisia strictly a consumer (rather than a producer)

of Arabic culture, isolating them. Because of the bitterness

that remains between French and Tunisians, Tunisians feel as

French as Canadians feel American, or less.

So the Tunisians are neither truly French, nor truly Arabic,

nor truly Mediterranean, nor truly African. They want to be

included in these groups but don't know how. They are

consumptive members of these four societies and struggle to

contribute to any.

In my perspective, the Tunisians should give up on joining

one of the existing clubs. The troubles had so far are not

about to be reconciled any time soon. Instead, Tunisia

should band together with Algeria, Morrocco, and Libya to

create and embrace a "French North Africa" club that is

Muslim, French, Arabic, and bordering the Mediterranean.

This group could share intensely from a cultural perspective

and could serve as a gateway from Europe (and the world) to

Africa. I talked with a woman who had to take flights through

London to get from South Africa to West Africa, and flights

between African nations **usually* require a European hub,

making interstate travel tiring and expensive; the double-

whammy of earning less money and having to pay more money

for trade is much of what slows Africa down.