1967, important year, parents had photos from that time, would he be willing to contribute photos? Dror can provide legal and technical means to do so
Idea of project started, talked with someone from Israeli internet association, which had $; she was interested in Dror's ideas, they had facilities and resources
Initiated fieldtrips to different places in Israel and have them upload pictures, but people already have large collection of images in their homes. Just ask them to contribute? They may be willing? all they may want is credit?
Why not use Wikimedia Commons?
Commons is confusing, especially licensing and large number of license choices (many irrelevant, international)
Idea is that photographers only submit their own works. Does not work well for cases where third parties collect images from others and submit them.
Not flexible enough to give recognition to local archive (or museum) for their contribution. Would be nice for them to have a special page on localized Hebrew interface rather than commons.
Cannot provide special localized system of categorization
Commons does have Hebrew interface but less than ideal for right to left languages
Commons admins are often unaware of local nuances of copyright laws and regulations
Did pikiwiki give up on commons?
No.
Commons has storage space and images are compatible with Commons criteria
Pikiwiki site used as intermediary between contributors and Commons, with localized interface. Pikiwiki site does not keep copies of images on their servers.
Israeli copyright laws and conventions are modeled after British statute of 1911, but was amended by Israel when it was created and amended again last year. Many nuances (moral rights, crown copyrights, ...) of Israeli law.
Contributors can choose public domain or CC-BY-SA
Pikiwiki does screen images to ensure images are relevant and licensing okay; pikiwiki (wm isreal) does uploading, translates and uploads to commons
Wikipedia Academy at NIH
Frank Schulenberg
First one in 2006
Workshop vs. Academy
Workshop: 10-20 participants, 1-2 hours, organized by Wikipedians
Academy: 80-200 participants, 1-2 days, target group: subject matter experts, organized by local chapter who can sign contracts, etc.
Academy held in July 16 at NIH
Organizers: Bill Wedemeyer, John Broughton, Lennart Guldbrandsson, Tim Vickers, Mathias Schindler, (9 people - 7 volunteers, 2 staff members)
Agenda:
Introduction
Bill Wedemeyer, about Wikipedia, why contribute?
Broughton informal talking with staff members during day
After lunch, training sessions with about 60, for beginners
Can they contribute to Wikipedia during work hours? NIH working out policy regarding social sites
Organizers:
Be clear about your goals
Improve understanding of Wikipedia among participants (post survey: 89% say yes they understand Wikipedia better)
Enable participants to make simple edits to Wikipedia (65% have confidence in ability to make simple edits; 24% neutral; 6% strongly agree)
Improve credibility of Wikipedia, recruit experts within topic area as speakers at the academy (they had Tim Vickers Ph.D., Professor Bill Wedemeyer, author John Broughton at academy)
Seeking Health Information Online: Does Wikipedia Matter?
[1]
Communication is key
Via e-mail
Weekly IRC meetings, including some participation from an NIH person
Meta wiki used for coordination
Publicity
Blog posts
Twitter feed
News coverages
Put out pictures
Rob Halsell
Wikimedia server infrastructure
350 servers — very small for a project of this size
80,000 requests per second
Comscore rates Wikipedia as #4 most visited site
16 users, including 10 with root access (used to be 5)
4 data centers, 2 in Tampa, 2 in Amsterdam
data primarily resides in tampa, peer only, caching, and toolserver in Amsterdam; if Tampa center went out of service, huge problem!!!
building out new data center (e.g in case hurricane)
servers originally in Tampa because Jimmy lived there, but it's not an ideal location, no good rates, not best network
moving to Ashburn, Virginia
In Amsterdam, KNAMS (Kenniset - no pay for internet access there, thanks to mark) and ESAMS (toolserver and caching only)
Rob does not do software development
He does site requests, extensions, groups, config changes requiring shell access, repair, upgrade broken hardware replacement, deploy new hardware and setup, tweaking
He works with bugzilla, which is the place for site requests
Deployment
very time consuming, will take 4 admins a months time (more?)
Inventory management
use racktables to keep track of all servers
developing own extensions to do this within mediawiki
yearly audits require accurate accounting for hardware, software and spending.
if not logged in, don't hit anything past the caching
cache deployed in both Tampa and Amsterdam
do not need to be recompiled by apache
apache only deployed in Tampa
when page changes, hit apache and database
memcache runs on apaches, stores recently fetched database data in resident memory. no need to hit database all the time
mysql servers in Tampa (30 servers), broken into four clusters. s1 for enwiki, s2 for 19 next largest wikis, s3 & s4 for other wikis
each cluster has master and multiple replication servers
everything runs on ubuntu (as possible)
run own netboot/pxe server as well as our own apt repository
create some custom packages for our deploymnent
some servers run nothing but images
use nagios to monitor servers, showing when they go down, this is open
[2]
ganglia, distributed monitoring
torrus, snmp monitooring and graphing software, see how much power we are drawing and data sent through switch
[3]
racktables for datacenter, no url needed it is locked and no one can login but employees
wikitech wiki, explains how to do things, documentation
Squid:
has it configured really well
nervous about moving to varnish, but perhaps at some point
Need site backup infrastructure, in case one center goes down
Toolserver
started in 2005, after frankfurt wikimania, with one server
2006 - new server, one for database replication and one for tools
2007 - 3 master database servers, also mirroring text in amsterdam, though not directly part of toolserver and not directly accessible to toolserver users (privacy concerns, including deleted revisions and private wikis)
2007 - complete backup of the site, in sense that data is not lost when tampa goes down, but not meant for running the site if tampa goes down
2008 - server added for stable tools, better reviewed, multiple maintainers
2009 - better database servers, new webserver to handle load better, new data for home (user) directories, osm servers including one toolserver (note, get diagram from Daniel of server schematic)
Now:
4 db servers, 32 gb ram each and 16 disks
3 login servers, 8/24 gb ram, 8 cores
web server, 8 cores, 8 gb ram
2 misc servers, 8 cores, 32 gb ram
2 smaller servers
Cool stuff:
variety of tools, bots, statistics, research, web tools -- wikiminiatlas, geohack
222 users currently
1/2 million pages served per day (12 gb of data)
Use:
php, python, perl, java, c#, ruby...
solaris, linux
zeus ws, tomcat, sjsws (don't use apache)
mediawiki, confluence
svn, jira, fisheye
ldap, nfs, zfs
mysql with live copy of db
Improved reliability:
2 copies of each db
wmf grant $40k for this
redundant /home directory
Internationalization:
cooperation with translatewiki.net
commons i18n library and database
Usability Initiative
Wikimedia usability project in 2005 (by Wikimedia Germany?), using same logo for this project
aiming for clearest path and remove barriers for editing
promote public participation, including from subject matter experts
erik (responsible for making project possible), brion (tech support), frank (outreach)
team:
parul - interaction designer, from yahoo!
trevor - software developer and gui designer
nimish - statistical analysis and linguistics
roan - software developer
hannes tank - designer intern
aaron wright - software developer volunteer
Usability study:
recruited participants,
3000 applications, narrowed down to 15 participants by demographics (age, gender, editing experience)
participant quotes -
"easier than i thought, just have to get used to it"
"it's really hard to see what i changed"
"i feel lost. did someone clean it up on me" (participant is a database administrator)
"it's much easier to add my content to a blog" (software developer)
"i'm just gonna keep trying here" (masters student in education)
"i'm not a programmer. i know the letters p-h-p" (business student)
Solution:
high impact-complex
simplification - collapse template markup, can expand if advanced user
differentiation -
streamlining - save preview workflow, flagged revisions, need to be integrated and presented in simple, understandable way
high-impact-simple
cleaning
navigation - acai release (improved navigation with vector skin)
feedback: watch needs to be more visible than how it's in the vector skin, without adding new tab
improved search interface
tools
acai toolbar, can be expanded for advanced functionality, and has special characters and help (cheat sheets)
low-impact-complex
user profiling
low-impact-simple
help
messaging
style guide
acai feedback survey:
53678 user enabled it
44018 still have it enabled
80% retained it on English
less retain it in Japanese and right to left languages
overall 75% retention rate
babaco release
new color palette
will be prototyping it
similar colors (red and blue links but more subtle)
keep contents on edit page, right aligned (tab, alternate with help tab, to the right of edit box)
editing help also on edit page (as tab, alternate with contents tab)
insert table wizard
insert media wizard (by michael dale)
citron release (early next year):
content folding
side-by-side preview
wysiwyg?
why not?
not yet, but apply easy solutions now that will have impact
please try beta!
need help with browser testing, feedback, and translation!!!
questions:
official repository of skins? scope of grant only to create one skin, but can have repository on mediawiki.org
usability testing with new tools and skin, (2?) tests will happen
track # of people who click edit and save? 1 out of 4 (according to Erik Zachte), need to subtract people who click red links who didn't intend to edit, other filters can be applied.
templates really scare new users - content folding
considering accessibility issues? mediawiki has shortcuts, but not addressed in grant, but trying not to take shortcuts away. would like to, but...
more about testing methods, usability test subjects are Americans - 10 were in s.f., 5 were remote; couldn't do languages other than english because of usability team limited language skills. reached out to other parts of U.S. but not international. would like help from chapters to do more testing.
semantic forms extension for templates - in discussion with yaron, would like to separate semantic wiki and template management. semantic is beyond; content folding is similar
Angela
wikia usability
wikia is working on wysiwyg
prototype
category select (add category button), ajax, also can select from existing categories (suggestions)
visible save bar, even on long pages, hovers at bottom of screen, can scroll past if you want to see footer
editing tips next to edit box
important links in sidebar
write an article
add an image
moved less-used links to footer (what links here, related changes, permanent link)
sidebar summary of activity - latest activity
action icons, separated from page tabs
reduce # of links on page
search suggest
link suggest for wikilinks (as you type)
talk pages
"leave message" link
contents page - hidden by default
more obvious section edit button
quick action panel - edit, move, add categories without visiting the page - good for advanced users (which wiki is this enabled?)
image placeholder, can be placed anywhere in article, links to image upload page
Discussion
Wikipedia: show article contributors on article page?
looking at wikitrust (says Erik Moeller)
Wikipedia: flagged revisions and patrolling recent changes, usability for how one parses the change?
Wikia: how far are we from being able to edit pages right on the screen (e.g. ajax) - question from domas
Wikia: being able to see who is working with who? - wikia uses social profile, but not very advanced yet
General understanding of wiki platform by users - lots of people don't know Wikipedia edited by users. Has there been any research into this.
MediaWiki usability: talk page (ting chen) - people complaining about it, it's confusing, can edit others' comments, any usability study of talk pages.
"this thread has been summarized as follows" feature
Wikia: add new page special page
Andrew: how it looks without javascript? wrote it without javascript to begin with, then added javascript. there would be a page reload, but it works but not fancy
my questions:
quick action panel - answers wikia
syntax highlighting - wish to do it, but lesser priority than content folding. it's difficult to do
social barriers to participation:
jack / wikihow: wants to "undo" adding of template feature; added when about to leave note to new editor (less than 50 edits), warns to make sure to leave good, constructive advice
naoko: german wikipedia has a mentoring system and it's working pretty well. other languages could learn from them. someone new users can talk to.
angela: lots of policy changes, educate users
wik-ed (editor) is a gadget now - inspiration for content folding and syntax highlighting
make language preference option more prominent
andrew: redesign was behind the scene, cleaning up the code, but not aware of any effort to make it more prominent
adding photos: multimedia usability project, grant from ford foundation
Roan: michael dale's work on upload and add media wizard
drafts extension: almost ready, with one of new usability releases.
Collaborative video
Erik
3,000 videos on wikipedia, very small compared to youtube, mainly animal clips, historic clips, etc. that support educational content on wikipedia
format: ogg theora, open source format and no patent restrictions
format is reason why so few videos, ogg is not widely understood but this is changing
attempt to standardize video, html5: new tags <video>, <audio>
collaboration with kaltura, who is developing open source technology that uses flash, we want to develop something completely open source, they are sponsoring michael dale for this work, and other people
open video conference in nyc, several hundred people, interesting discussion, other video sites talking about adapting ogg (daily motion, internet archive)
beside fact format is open, other primary reason is ability to use video much in way using images on the web, with programs that modify video like javascript can modify images
mozilla has created cool demos of supporting ogg (e.g. motion detection, recognize patterns in videos)
Shay
open video alliance
[4] - not an official organization, but brings organizations and individuals together
open video conference: 900 people, 200 people submitted to speak
video 1.0 (television programming, films, live events, advertisements, video on-demand) vs. video 2.0 (also webisodes, democratization programming, mash-up and remixing, video sharing, ...) - audience as active participant
Michael
HTML5 video player
make it easy to embed video in browser and not worry about which version of browser user is using. these questions can be answered with javascript, and mapped to uniform interface
can update the theme (theme roller)
uses jquery ui
mashable player that people can customize to their needs
opencongress.org using video player with slightly different look and feel
new skin, more accessible menu items
handling can be added for video tag to adapt for older browsers (e.g. using cortado java applet video player)
firefogg
browser extension to select high quality source and we handle ogg encoding
also improve uploading via http
can upload in chunks, helps if internet connection breaks
scriptloader
make ogg - super lots of javascript loading (bad!)
new upload api
handle upload, provide upload status feedback, etc.
integration with mediawiki
can download files easily, like images - "save video as..." when right clicking
add media wizard
easily find images on commons, can crop image and insert into page
can be imported from commons, for non wikimedia wiki sites
can import clips from remote sites like internet archive
upload wizard - is a prototype
stand-alone tool, could be incorporated into a blog
sequencer
sequence editing - rearrange video and put clips together, use video template, wizard for editing clip properties (title, date), can drag and drop and import resources from various sources, can create high quality metadata with ease, add clip transitions
timed text editing
Questions
timeline for deployment? - deployment plan, upload api is in trunk tree, allow other features to be implemented, iterative deployment of small things, next thing is add media wizard, leading up to sequence editor
upload api to go live with next scap (this week?), enabling chunk uploads with firefogg
ford foundation project, build upon michael's work, to streamline upload process, make licensing clear, and also tools for assessing content (e.g. deletion)
how to apply wiki policies (no original research) to video, would we want interviews in articles?
wiki@home - donate spare cpu cycles to help with resource intensive operations
Covering Wikipedia
Andrew Lih
Panelists:
Andrew
Liam
Phoebe
Noam
Jay
Outlets:
signpost
wikipedia weekly
wikivoices
openwikiblogplanet
planet wikipedia
wikizine
questions:
where to find interesting stories
where to find interesting people
how to determine what to cover and what not to cover
how to portray things accurately and not just from our perspective
how to represent complex system to outside world
tough issues:
explaining policy, accumulated culture
explain arbitration fairly to outside world
Critical issues
meaning of editor and admin
how powerful is an admin
audience?
what would you like to see better coverage of?
Phoebe
How Wikipedia Works - book co-authored with Charles Matthews
How to enable people to participate, without coloring situation with own individual experience (e.g. our own arguments and frustrations)
Wikipedia Signpost - started by Michael Snow in 2005
Jay
press releases
annual report
blog
convey that wikipedia is run by volunteers and foundation is non-profit
Noam
essjay story, seigenthaler...
wants to know and report about changes in wikipedia (e.g. flagged revisions)
posting images on wikipedia story
google news linking to wikipedia was a story
david rohde story
wikipedia is powerful player that needs to be covered, similar as google is covered
must keep stories tight, short, and engaging
wikipedia debates, statistics, and people are public and accessible, but challenge is the mass of information and people
explain jimmy wales role is difficult
people will find fault with what he does, but he does the best he can
covering the strategy project, hard to explain and understand that
Phoebe
even Phoebe has difficulty knowing and understanding what's going on with flagged revisions, even though she reads the discussion boards, mailing lists, talks to brion, etc.
Liam
Wikipedia Weekly (3 years old), but not weekly.
does not intend to be neutral, but insert own point of view into the discussions
and there is debate, with liam and andrew taking different points of view
Phoebe: Signpost aims to be neutral
Liam also did academic thesis about wikipedia for academics, wikipedia as a subject of study was questioned a few years ago but now is not questioned.
meaning of user - edit it, read it, cite it?
many people in community don't know they are in the community
Questions
david rohde case, what/how to decide not to report on it
phoebe: to report things that might be damaging? damaging to the foundation?
mattias: paper from ap, identifies Wikipedia as a competitor? implication for how media covers Wikipedia?
noam: inverse relationship to that, as a competitor, media gives more respect; there are some exceptions, however.
phoebe: signpost is looking for writers
noam: linking to wikipedia articles? it is being debated. currently don't link, unless specifically related to a nytimes article. a lot of pressure to link, but right now very ad hoc, might be done by reporters
Sue
the year in review, the year ahead
2007-2008 (setting up, hired lawyer)
2008-2009 (move to sf, hire fundraising staff and seek out support for wikimedia projects)
Sue has traveled and met with people, including visiting India.
participation chart
Stanton usability project
quality - teachers working with teachers
wikipedia academy - mathematics
grad students
wikipedia academy - at NIH
sue is accountable to the board and to the community
this is a systematic effort, to be done together, with resources, infrastructure and expertise
how you can be involved!
call for proposals, several hundred so far
rate and comment on proposals
if you have time and interest, and particular expertise, there are opportunities to serve on a task force
provide research support (bridgespan's slides available in fact base on strategy wiki)
always great to have friendly, constructive, helpful people involved to facilitate conversations, maintain wiki, conduct interviews, etc.
call for participation will be open soon, and open for ~10 days.
thank you to the community!
Board panel
Jan Bart, SJ, Ting, Jimmy, Kat, Arno, Michael, Domas
Questions:
Brianna: where is Matt Halprin? Michael: he could only be here on Wednesday. Wikipedia Weekly conducted an interview with him. Stu West is also not here, due to work schedule. Stu did join board meeting via Skype.
Lodewik: new board members, role of board? SJ: role of community representatives as bridging gap between what community does and foundation does are in line. Board is stabilizing force, and maintains vision. Arno: agrees with SJ, adds that for him, the board has responsibility for foundation and for the wikimedia movement (not just editors, but as a whole, broader sense).
IRC - Ragesoss: Omidyar Network grant - What are the specific targets for percent from individual donations, reach, and participation rates? Michael: we draw on contributions of large network of people, not all come from single concentrated source, in terms of project contributions or financial aspect. targets in grant relate to fundraising, having us work to strengthen and build percentage of revenues supplied by general small donations and the way that we be able to work in the past so that we are not reliant on the good graces of one or a handful of major donors. that's one aspect of targets they are interested in. also, the progress of the strategic planning initiative is one reason we were interested in bringing matt on the board because of his knowledge and experience in that area. omidyar is interested in seeing we make progress in that regard, so we have a roadmap not just the board and staff, but community know where we want to go in the next five year.
Dror: Damian's story of establishing chapter in Venezuela, very hard and many people enthusiastic to get together and form chapter, but lot of physical and technical problems. gathering in one place is problematic. one way the foundation can help to establish such chapters or local organizations or even meetups where there are such difficulties? Ting: think that we have started in last year with work of foundation to reach out and make yearly programs organized by foundation; foundation can help volunteers to do things, people can always ask foundation for help. Michael: most chapters we have so far are organized on their own momentum and not needed much help to do that. there are two things we can do. 1) late last year board decided on subnational chapters, which can help overcome problem of great distances 2) help potential indian chapter to organize. advisory board member has helped arrange support from internet society in india which is providing support to local wikipedians. he hopes those two thing can provide path for other chapters organizing in future. Jan Bart: more and more clear that chapter with legal entity is not always best solution. need to explore other options, some places it may not be good idea to form legal entity. will work on that issue.
audience question: impression of wikimania this year. jan bart: lots of energy , lots of ideas, good to meet people behind mailing lists. strategy project great. Ting: this is the wikimania that ting attended that had highest quality of presentations. sign our community is growing big and major and find it really good. Domas: lounge areas were awesome, great for talking to people. SJ: he's been to all five wikimanias. amazing here. warmth of people hanging out in general space, high quality of speakers and their data, organizers for simultaneous translation. very much enjoyed spanish language sessions on education. Arno: co-organizer in frankfurt. impressed by incredible amount of people and diversity, and thoughtful talks, and spirit. very impressed that this year is like frankfurt, amazing job by organizers. kat: really amazed how much people enjoy what they do and like to share it, what they are excited about. being able to interact with people in informal setting, get more, rather than formal conversation jimbo: almost everything filmed and online. have that human connection, really looks forward to the party. michael: great time at conference, too many sessions to choose from, fantastic venue and organization, it was perfect.
in three years, reach and participation - jimmy: we will be bigger, he is really passionate about growth in developing world. not a lot of room to grow in english, since we already are a top google search but reach can grow a lot in other places. jan bart: wrote strategy proposal, worries about participation, andrew lih and kat say about going back to core values, same talk could be held last year and this year, hope it won't be needed next year. everyone appreciates editors, but not necessarily appreciated doing other things like welcoming new users. (Mattias' baby crys) Michael: participation ... in three - five years, wants to see more avenues and easier ways for participation. usability project or work on open and collaborative video, (encourages people to see Michael Dale's presentation video), more ways for people to contribute. Ting: wikipedia dominates projects, but there are other projects. two teachers from university in america had adopted method of teaching with, asked students to write textbook for themselves, afterwards asked students to ask about method? most students said yes. such ideas can expand, and other projects may grow in future.
journalist in buenos aires, not an expert, no one at board table speaking spanish. Jimbo: we have dutch, chinese, french, lithuanian; kat: lo siento no hablo espanol. SJ: we did have a spanish speaking board candidate this year.
tech expertise on board - domas: quite a few board members have computer science background, he's not completely uncomfortable about leaving board. ways to communicate with board, not a big issue. domas may still be hanging around and won't go away.
no session about iphone and mobile phones, iphone application launched just this month Jay: there was one presentation; Jan Bart: depends on people and getting space, imagine there are tons of things going on, it could deserve more focus what we are doing in mobile space. Domas: iphone POV, android, he loves using mobile platform, this is a key focus and people working on it. iphone is just one thing. mobile web version is getting use, especially good for areas without great internet but have mobile internet. android users in future. Kat: unfortunate Stu is not here, he has collected data in hist background, what equipment people use in getting information, help get comscore data for different countries.
argentinian - how the board thinks traditional media deal with new media (wikipedia), what role for traditional media in future years - Michael: don't know what future holds for traditional media, some of them this is very challenging time because of wikipedia or because of business model which have nothing particular to do with us, we will always have some sort of media around and not sure what that will look like in future, but something we as partof this movement depend on. we have symbiatic relation with it because we don't do original research and rely on other sources, books, media, whatever it may be. that's the environment in which we work and can benefit from and hope that traditional organizations work out challenges and can find benefits they can draw from wikimedia projects. our mission is to share knowledge, including people who think they already have other sources and produce journalism.
since making first edit, wikipedia is becoming more and more complex, trying to see what would happen in growing complexity, as you may see, on board there is imbalance in diversity (e.g. gender). effort to improve diversity. Jimbo: not sure about regional imbalance... yes he does agree. it's an american and european board for most part. Ting is chinese but lives in europe. gender balance is definitely a concern. it's complicated and difficult topic. need to do better job. Ting: important for people to think about then when evaluating candidates.
IRC question- how should wikipedia movement be explained beyond wikipedia, to attract 300 million participants in future - Kat: needs to start with shared values and articulate what we are doing and why. more people will want to get involved with movement. that's the first step. Jimbo: most important, work regarding usability. he attended wikihow (Jack) presentation, talked about wysywyg, including wikia, see dramatic change in ratio of users that click edit and then save their work. that would reduce technical barrier to participation. SJ: everyone has something to teach, and can teach others. move beyond it's a place to read, but it's a place to share with others. Jan: need to be more friendly community and actively practice values.
time is short, need to wind down session, need to move to closing
Michael: one last round of applause for domas for his service on the board
1967, important year, parents had photos from that time, would he be willing to contribute photos? Dror can provide legal and technical means to do so
Idea of project started, talked with someone from Israeli internet association, which had $; she was interested in Dror's ideas, they had facilities and resources
Initiated fieldtrips to different places in Israel and have them upload pictures, but people already have large collection of images in their homes. Just ask them to contribute? They may be willing? all they may want is credit?
Why not use Wikimedia Commons?
Commons is confusing, especially licensing and large number of license choices (many irrelevant, international)
Idea is that photographers only submit their own works. Does not work well for cases where third parties collect images from others and submit them.
Not flexible enough to give recognition to local archive (or museum) for their contribution. Would be nice for them to have a special page on localized Hebrew interface rather than commons.
Cannot provide special localized system of categorization
Commons does have Hebrew interface but less than ideal for right to left languages
Commons admins are often unaware of local nuances of copyright laws and regulations
Did pikiwiki give up on commons?
No.
Commons has storage space and images are compatible with Commons criteria
Pikiwiki site used as intermediary between contributors and Commons, with localized interface. Pikiwiki site does not keep copies of images on their servers.
Israeli copyright laws and conventions are modeled after British statute of 1911, but was amended by Israel when it was created and amended again last year. Many nuances (moral rights, crown copyrights, ...) of Israeli law.
Contributors can choose public domain or CC-BY-SA
Pikiwiki does screen images to ensure images are relevant and licensing okay; pikiwiki (wm isreal) does uploading, translates and uploads to commons
Wikipedia Academy at NIH
Frank Schulenberg
First one in 2006
Workshop vs. Academy
Workshop: 10-20 participants, 1-2 hours, organized by Wikipedians
Academy: 80-200 participants, 1-2 days, target group: subject matter experts, organized by local chapter who can sign contracts, etc.
Academy held in July 16 at NIH
Organizers: Bill Wedemeyer, John Broughton, Lennart Guldbrandsson, Tim Vickers, Mathias Schindler, (9 people - 7 volunteers, 2 staff members)
Agenda:
Introduction
Bill Wedemeyer, about Wikipedia, why contribute?
Broughton informal talking with staff members during day
After lunch, training sessions with about 60, for beginners
Can they contribute to Wikipedia during work hours? NIH working out policy regarding social sites
Organizers:
Be clear about your goals
Improve understanding of Wikipedia among participants (post survey: 89% say yes they understand Wikipedia better)
Enable participants to make simple edits to Wikipedia (65% have confidence in ability to make simple edits; 24% neutral; 6% strongly agree)
Improve credibility of Wikipedia, recruit experts within topic area as speakers at the academy (they had Tim Vickers Ph.D., Professor Bill Wedemeyer, author John Broughton at academy)
Seeking Health Information Online: Does Wikipedia Matter?
[1]
Communication is key
Via e-mail
Weekly IRC meetings, including some participation from an NIH person
Meta wiki used for coordination
Publicity
Blog posts
Twitter feed
News coverages
Put out pictures
Rob Halsell
Wikimedia server infrastructure
350 servers — very small for a project of this size
80,000 requests per second
Comscore rates Wikipedia as #4 most visited site
16 users, including 10 with root access (used to be 5)
4 data centers, 2 in Tampa, 2 in Amsterdam
data primarily resides in tampa, peer only, caching, and toolserver in Amsterdam; if Tampa center went out of service, huge problem!!!
building out new data center (e.g in case hurricane)
servers originally in Tampa because Jimmy lived there, but it's not an ideal location, no good rates, not best network
moving to Ashburn, Virginia
In Amsterdam, KNAMS (Kenniset - no pay for internet access there, thanks to mark) and ESAMS (toolserver and caching only)
Rob does not do software development
He does site requests, extensions, groups, config changes requiring shell access, repair, upgrade broken hardware replacement, deploy new hardware and setup, tweaking
He works with bugzilla, which is the place for site requests
Deployment
very time consuming, will take 4 admins a months time (more?)
Inventory management
use racktables to keep track of all servers
developing own extensions to do this within mediawiki
yearly audits require accurate accounting for hardware, software and spending.
if not logged in, don't hit anything past the caching
cache deployed in both Tampa and Amsterdam
do not need to be recompiled by apache
apache only deployed in Tampa
when page changes, hit apache and database
memcache runs on apaches, stores recently fetched database data in resident memory. no need to hit database all the time
mysql servers in Tampa (30 servers), broken into four clusters. s1 for enwiki, s2 for 19 next largest wikis, s3 & s4 for other wikis
each cluster has master and multiple replication servers
everything runs on ubuntu (as possible)
run own netboot/pxe server as well as our own apt repository
create some custom packages for our deploymnent
some servers run nothing but images
use nagios to monitor servers, showing when they go down, this is open
[2]
ganglia, distributed monitoring
torrus, snmp monitooring and graphing software, see how much power we are drawing and data sent through switch
[3]
racktables for datacenter, no url needed it is locked and no one can login but employees
wikitech wiki, explains how to do things, documentation
Squid:
has it configured really well
nervous about moving to varnish, but perhaps at some point
Need site backup infrastructure, in case one center goes down
Toolserver
started in 2005, after frankfurt wikimania, with one server
2006 - new server, one for database replication and one for tools
2007 - 3 master database servers, also mirroring text in amsterdam, though not directly part of toolserver and not directly accessible to toolserver users (privacy concerns, including deleted revisions and private wikis)
2007 - complete backup of the site, in sense that data is not lost when tampa goes down, but not meant for running the site if tampa goes down
2008 - server added for stable tools, better reviewed, multiple maintainers
2009 - better database servers, new webserver to handle load better, new data for home (user) directories, osm servers including one toolserver (note, get diagram from Daniel of server schematic)
Now:
4 db servers, 32 gb ram each and 16 disks
3 login servers, 8/24 gb ram, 8 cores
web server, 8 cores, 8 gb ram
2 misc servers, 8 cores, 32 gb ram
2 smaller servers
Cool stuff:
variety of tools, bots, statistics, research, web tools -- wikiminiatlas, geohack
222 users currently
1/2 million pages served per day (12 gb of data)
Use:
php, python, perl, java, c#, ruby...
solaris, linux
zeus ws, tomcat, sjsws (don't use apache)
mediawiki, confluence
svn, jira, fisheye
ldap, nfs, zfs
mysql with live copy of db
Improved reliability:
2 copies of each db
wmf grant $40k for this
redundant /home directory
Internationalization:
cooperation with translatewiki.net
commons i18n library and database
Usability Initiative
Wikimedia usability project in 2005 (by Wikimedia Germany?), using same logo for this project
aiming for clearest path and remove barriers for editing
promote public participation, including from subject matter experts
erik (responsible for making project possible), brion (tech support), frank (outreach)
team:
parul - interaction designer, from yahoo!
trevor - software developer and gui designer
nimish - statistical analysis and linguistics
roan - software developer
hannes tank - designer intern
aaron wright - software developer volunteer
Usability study:
recruited participants,
3000 applications, narrowed down to 15 participants by demographics (age, gender, editing experience)
participant quotes -
"easier than i thought, just have to get used to it"
"it's really hard to see what i changed"
"i feel lost. did someone clean it up on me" (participant is a database administrator)
"it's much easier to add my content to a blog" (software developer)
"i'm just gonna keep trying here" (masters student in education)
"i'm not a programmer. i know the letters p-h-p" (business student)
Solution:
high impact-complex
simplification - collapse template markup, can expand if advanced user
differentiation -
streamlining - save preview workflow, flagged revisions, need to be integrated and presented in simple, understandable way
high-impact-simple
cleaning
navigation - acai release (improved navigation with vector skin)
feedback: watch needs to be more visible than how it's in the vector skin, without adding new tab
improved search interface
tools
acai toolbar, can be expanded for advanced functionality, and has special characters and help (cheat sheets)
low-impact-complex
user profiling
low-impact-simple
help
messaging
style guide
acai feedback survey:
53678 user enabled it
44018 still have it enabled
80% retained it on English
less retain it in Japanese and right to left languages
overall 75% retention rate
babaco release
new color palette
will be prototyping it
similar colors (red and blue links but more subtle)
keep contents on edit page, right aligned (tab, alternate with help tab, to the right of edit box)
editing help also on edit page (as tab, alternate with contents tab)
insert table wizard
insert media wizard (by michael dale)
citron release (early next year):
content folding
side-by-side preview
wysiwyg?
why not?
not yet, but apply easy solutions now that will have impact
please try beta!
need help with browser testing, feedback, and translation!!!
questions:
official repository of skins? scope of grant only to create one skin, but can have repository on mediawiki.org
usability testing with new tools and skin, (2?) tests will happen
track # of people who click edit and save? 1 out of 4 (according to Erik Zachte), need to subtract people who click red links who didn't intend to edit, other filters can be applied.
templates really scare new users - content folding
considering accessibility issues? mediawiki has shortcuts, but not addressed in grant, but trying not to take shortcuts away. would like to, but...
more about testing methods, usability test subjects are Americans - 10 were in s.f., 5 were remote; couldn't do languages other than english because of usability team limited language skills. reached out to other parts of U.S. but not international. would like help from chapters to do more testing.
semantic forms extension for templates - in discussion with yaron, would like to separate semantic wiki and template management. semantic is beyond; content folding is similar
Angela
wikia usability
wikia is working on wysiwyg
prototype
category select (add category button), ajax, also can select from existing categories (suggestions)
visible save bar, even on long pages, hovers at bottom of screen, can scroll past if you want to see footer
editing tips next to edit box
important links in sidebar
write an article
add an image
moved less-used links to footer (what links here, related changes, permanent link)
sidebar summary of activity - latest activity
action icons, separated from page tabs
reduce # of links on page
search suggest
link suggest for wikilinks (as you type)
talk pages
"leave message" link
contents page - hidden by default
more obvious section edit button
quick action panel - edit, move, add categories without visiting the page - good for advanced users (which wiki is this enabled?)
image placeholder, can be placed anywhere in article, links to image upload page
Discussion
Wikipedia: show article contributors on article page?
looking at wikitrust (says Erik Moeller)
Wikipedia: flagged revisions and patrolling recent changes, usability for how one parses the change?
Wikia: how far are we from being able to edit pages right on the screen (e.g. ajax) - question from domas
Wikia: being able to see who is working with who? - wikia uses social profile, but not very advanced yet
General understanding of wiki platform by users - lots of people don't know Wikipedia edited by users. Has there been any research into this.
MediaWiki usability: talk page (ting chen) - people complaining about it, it's confusing, can edit others' comments, any usability study of talk pages.
"this thread has been summarized as follows" feature
Wikia: add new page special page
Andrew: how it looks without javascript? wrote it without javascript to begin with, then added javascript. there would be a page reload, but it works but not fancy
my questions:
quick action panel - answers wikia
syntax highlighting - wish to do it, but lesser priority than content folding. it's difficult to do
social barriers to participation:
jack / wikihow: wants to "undo" adding of template feature; added when about to leave note to new editor (less than 50 edits), warns to make sure to leave good, constructive advice
naoko: german wikipedia has a mentoring system and it's working pretty well. other languages could learn from them. someone new users can talk to.
angela: lots of policy changes, educate users
wik-ed (editor) is a gadget now - inspiration for content folding and syntax highlighting
make language preference option more prominent
andrew: redesign was behind the scene, cleaning up the code, but not aware of any effort to make it more prominent
adding photos: multimedia usability project, grant from ford foundation
Roan: michael dale's work on upload and add media wizard
drafts extension: almost ready, with one of new usability releases.
Collaborative video
Erik
3,000 videos on wikipedia, very small compared to youtube, mainly animal clips, historic clips, etc. that support educational content on wikipedia
format: ogg theora, open source format and no patent restrictions
format is reason why so few videos, ogg is not widely understood but this is changing
attempt to standardize video, html5: new tags <video>, <audio>
collaboration with kaltura, who is developing open source technology that uses flash, we want to develop something completely open source, they are sponsoring michael dale for this work, and other people
open video conference in nyc, several hundred people, interesting discussion, other video sites talking about adapting ogg (daily motion, internet archive)
beside fact format is open, other primary reason is ability to use video much in way using images on the web, with programs that modify video like javascript can modify images
mozilla has created cool demos of supporting ogg (e.g. motion detection, recognize patterns in videos)
Shay
open video alliance
[4] - not an official organization, but brings organizations and individuals together
open video conference: 900 people, 200 people submitted to speak
video 1.0 (television programming, films, live events, advertisements, video on-demand) vs. video 2.0 (also webisodes, democratization programming, mash-up and remixing, video sharing, ...) - audience as active participant
Michael
HTML5 video player
make it easy to embed video in browser and not worry about which version of browser user is using. these questions can be answered with javascript, and mapped to uniform interface
can update the theme (theme roller)
uses jquery ui
mashable player that people can customize to their needs
opencongress.org using video player with slightly different look and feel
new skin, more accessible menu items
handling can be added for video tag to adapt for older browsers (e.g. using cortado java applet video player)
firefogg
browser extension to select high quality source and we handle ogg encoding
also improve uploading via http
can upload in chunks, helps if internet connection breaks
scriptloader
make ogg - super lots of javascript loading (bad!)
new upload api
handle upload, provide upload status feedback, etc.
integration with mediawiki
can download files easily, like images - "save video as..." when right clicking
add media wizard
easily find images on commons, can crop image and insert into page
can be imported from commons, for non wikimedia wiki sites
can import clips from remote sites like internet archive
upload wizard - is a prototype
stand-alone tool, could be incorporated into a blog
sequencer
sequence editing - rearrange video and put clips together, use video template, wizard for editing clip properties (title, date), can drag and drop and import resources from various sources, can create high quality metadata with ease, add clip transitions
timed text editing
Questions
timeline for deployment? - deployment plan, upload api is in trunk tree, allow other features to be implemented, iterative deployment of small things, next thing is add media wizard, leading up to sequence editor
upload api to go live with next scap (this week?), enabling chunk uploads with firefogg
ford foundation project, build upon michael's work, to streamline upload process, make licensing clear, and also tools for assessing content (e.g. deletion)
how to apply wiki policies (no original research) to video, would we want interviews in articles?
wiki@home - donate spare cpu cycles to help with resource intensive operations
Covering Wikipedia
Andrew Lih
Panelists:
Andrew
Liam
Phoebe
Noam
Jay
Outlets:
signpost
wikipedia weekly
wikivoices
openwikiblogplanet
planet wikipedia
wikizine
questions:
where to find interesting stories
where to find interesting people
how to determine what to cover and what not to cover
how to portray things accurately and not just from our perspective
how to represent complex system to outside world
tough issues:
explaining policy, accumulated culture
explain arbitration fairly to outside world
Critical issues
meaning of editor and admin
how powerful is an admin
audience?
what would you like to see better coverage of?
Phoebe
How Wikipedia Works - book co-authored with Charles Matthews
How to enable people to participate, without coloring situation with own individual experience (e.g. our own arguments and frustrations)
Wikipedia Signpost - started by Michael Snow in 2005
Jay
press releases
annual report
blog
convey that wikipedia is run by volunteers and foundation is non-profit
Noam
essjay story, seigenthaler...
wants to know and report about changes in wikipedia (e.g. flagged revisions)
posting images on wikipedia story
google news linking to wikipedia was a story
david rohde story
wikipedia is powerful player that needs to be covered, similar as google is covered
must keep stories tight, short, and engaging
wikipedia debates, statistics, and people are public and accessible, but challenge is the mass of information and people
explain jimmy wales role is difficult
people will find fault with what he does, but he does the best he can
covering the strategy project, hard to explain and understand that
Phoebe
even Phoebe has difficulty knowing and understanding what's going on with flagged revisions, even though she reads the discussion boards, mailing lists, talks to brion, etc.
Liam
Wikipedia Weekly (3 years old), but not weekly.
does not intend to be neutral, but insert own point of view into the discussions
and there is debate, with liam and andrew taking different points of view
Phoebe: Signpost aims to be neutral
Liam also did academic thesis about wikipedia for academics, wikipedia as a subject of study was questioned a few years ago but now is not questioned.
meaning of user - edit it, read it, cite it?
many people in community don't know they are in the community
Questions
david rohde case, what/how to decide not to report on it
phoebe: to report things that might be damaging? damaging to the foundation?
mattias: paper from ap, identifies Wikipedia as a competitor? implication for how media covers Wikipedia?
noam: inverse relationship to that, as a competitor, media gives more respect; there are some exceptions, however.
phoebe: signpost is looking for writers
noam: linking to wikipedia articles? it is being debated. currently don't link, unless specifically related to a nytimes article. a lot of pressure to link, but right now very ad hoc, might be done by reporters
Sue
the year in review, the year ahead
2007-2008 (setting up, hired lawyer)
2008-2009 (move to sf, hire fundraising staff and seek out support for wikimedia projects)
Sue has traveled and met with people, including visiting India.
participation chart
Stanton usability project
quality - teachers working with teachers
wikipedia academy - mathematics
grad students
wikipedia academy - at NIH
sue is accountable to the board and to the community
this is a systematic effort, to be done together, with resources, infrastructure and expertise
how you can be involved!
call for proposals, several hundred so far
rate and comment on proposals
if you have time and interest, and particular expertise, there are opportunities to serve on a task force
provide research support (bridgespan's slides available in fact base on strategy wiki)
always great to have friendly, constructive, helpful people involved to facilitate conversations, maintain wiki, conduct interviews, etc.
call for participation will be open soon, and open for ~10 days.
thank you to the community!
Board panel
Jan Bart, SJ, Ting, Jimmy, Kat, Arno, Michael, Domas
Questions:
Brianna: where is Matt Halprin? Michael: he could only be here on Wednesday. Wikipedia Weekly conducted an interview with him. Stu West is also not here, due to work schedule. Stu did join board meeting via Skype.
Lodewik: new board members, role of board? SJ: role of community representatives as bridging gap between what community does and foundation does are in line. Board is stabilizing force, and maintains vision. Arno: agrees with SJ, adds that for him, the board has responsibility for foundation and for the wikimedia movement (not just editors, but as a whole, broader sense).
IRC - Ragesoss: Omidyar Network grant - What are the specific targets for percent from individual donations, reach, and participation rates? Michael: we draw on contributions of large network of people, not all come from single concentrated source, in terms of project contributions or financial aspect. targets in grant relate to fundraising, having us work to strengthen and build percentage of revenues supplied by general small donations and the way that we be able to work in the past so that we are not reliant on the good graces of one or a handful of major donors. that's one aspect of targets they are interested in. also, the progress of the strategic planning initiative is one reason we were interested in bringing matt on the board because of his knowledge and experience in that area. omidyar is interested in seeing we make progress in that regard, so we have a roadmap not just the board and staff, but community know where we want to go in the next five year.
Dror: Damian's story of establishing chapter in Venezuela, very hard and many people enthusiastic to get together and form chapter, but lot of physical and technical problems. gathering in one place is problematic. one way the foundation can help to establish such chapters or local organizations or even meetups where there are such difficulties? Ting: think that we have started in last year with work of foundation to reach out and make yearly programs organized by foundation; foundation can help volunteers to do things, people can always ask foundation for help. Michael: most chapters we have so far are organized on their own momentum and not needed much help to do that. there are two things we can do. 1) late last year board decided on subnational chapters, which can help overcome problem of great distances 2) help potential indian chapter to organize. advisory board member has helped arrange support from internet society in india which is providing support to local wikipedians. he hopes those two thing can provide path for other chapters organizing in future. Jan Bart: more and more clear that chapter with legal entity is not always best solution. need to explore other options, some places it may not be good idea to form legal entity. will work on that issue.
audience question: impression of wikimania this year. jan bart: lots of energy , lots of ideas, good to meet people behind mailing lists. strategy project great. Ting: this is the wikimania that ting attended that had highest quality of presentations. sign our community is growing big and major and find it really good. Domas: lounge areas were awesome, great for talking to people. SJ: he's been to all five wikimanias. amazing here. warmth of people hanging out in general space, high quality of speakers and their data, organizers for simultaneous translation. very much enjoyed spanish language sessions on education. Arno: co-organizer in frankfurt. impressed by incredible amount of people and diversity, and thoughtful talks, and spirit. very impressed that this year is like frankfurt, amazing job by organizers. kat: really amazed how much people enjoy what they do and like to share it, what they are excited about. being able to interact with people in informal setting, get more, rather than formal conversation jimbo: almost everything filmed and online. have that human connection, really looks forward to the party. michael: great time at conference, too many sessions to choose from, fantastic venue and organization, it was perfect.
in three years, reach and participation - jimmy: we will be bigger, he is really passionate about growth in developing world. not a lot of room to grow in english, since we already are a top google search but reach can grow a lot in other places. jan bart: wrote strategy proposal, worries about participation, andrew lih and kat say about going back to core values, same talk could be held last year and this year, hope it won't be needed next year. everyone appreciates editors, but not necessarily appreciated doing other things like welcoming new users. (Mattias' baby crys) Michael: participation ... in three - five years, wants to see more avenues and easier ways for participation. usability project or work on open and collaborative video, (encourages people to see Michael Dale's presentation video), more ways for people to contribute. Ting: wikipedia dominates projects, but there are other projects. two teachers from university in america had adopted method of teaching with, asked students to write textbook for themselves, afterwards asked students to ask about method? most students said yes. such ideas can expand, and other projects may grow in future.
journalist in buenos aires, not an expert, no one at board table speaking spanish. Jimbo: we have dutch, chinese, french, lithuanian; kat: lo siento no hablo espanol. SJ: we did have a spanish speaking board candidate this year.
tech expertise on board - domas: quite a few board members have computer science background, he's not completely uncomfortable about leaving board. ways to communicate with board, not a big issue. domas may still be hanging around and won't go away.
no session about iphone and mobile phones, iphone application launched just this month Jay: there was one presentation; Jan Bart: depends on people and getting space, imagine there are tons of things going on, it could deserve more focus what we are doing in mobile space. Domas: iphone POV, android, he loves using mobile platform, this is a key focus and people working on it. iphone is just one thing. mobile web version is getting use, especially good for areas without great internet but have mobile internet. android users in future. Kat: unfortunate Stu is not here, he has collected data in hist background, what equipment people use in getting information, help get comscore data for different countries.
argentinian - how the board thinks traditional media deal with new media (wikipedia), what role for traditional media in future years - Michael: don't know what future holds for traditional media, some of them this is very challenging time because of wikipedia or because of business model which have nothing particular to do with us, we will always have some sort of media around and not sure what that will look like in future, but something we as partof this movement depend on. we have symbiatic relation with it because we don't do original research and rely on other sources, books, media, whatever it may be. that's the environment in which we work and can benefit from and hope that traditional organizations work out challenges and can find benefits they can draw from wikimedia projects. our mission is to share knowledge, including people who think they already have other sources and produce journalism.
since making first edit, wikipedia is becoming more and more complex, trying to see what would happen in growing complexity, as you may see, on board there is imbalance in diversity (e.g. gender). effort to improve diversity. Jimbo: not sure about regional imbalance... yes he does agree. it's an american and european board for most part. Ting is chinese but lives in europe. gender balance is definitely a concern. it's complicated and difficult topic. need to do better job. Ting: important for people to think about then when evaluating candidates.
IRC question- how should wikipedia movement be explained beyond wikipedia, to attract 300 million participants in future - Kat: needs to start with shared values and articulate what we are doing and why. more people will want to get involved with movement. that's the first step. Jimbo: most important, work regarding usability. he attended wikihow (Jack) presentation, talked about wysywyg, including wikia, see dramatic change in ratio of users that click edit and then save their work. that would reduce technical barrier to participation. SJ: everyone has something to teach, and can teach others. move beyond it's a place to read, but it's a place to share with others. Jan: need to be more friendly community and actively practice values.
time is short, need to wind down session, need to move to closing
Michael: one last round of applause for domas for his service on the board