This paper (first reference) is the result of a class project I was part of
almost two years ago for CSCI 5417 Information Retrieval Systems. It builds
on a class project I did in CSCI 5832 Natural Language Processing and which
I presented at Wikimania '07. The project was very late as we didn't send
the final paper in until the day before new years. This technical report was
never really announced that I recall so I thought it would be interesting to
look briefly at the results. The goal of this paper was to break articles
down into surface features and latent features and then use those to study
the rating system being used, predict article quality and rank results in a
search engine. We used the [[random forests]] classifier which allowed us to
analyze the contribution of each feature to performance by looking directly
at the weights that were assigned. While the surface analysis was performed
on the whole english wikipedia, the latent analysis was performed on the
simple english wikipedia (it is more expensive to compute). = Surface
features = * Readability measures are the single best predictor of quality
that I have found, as defined by the Wikipedia Editorial Team (WET). The
[[Automated Readability Index]], [[Gunning Fog Index]] and [[Flesch-Kincaid
Grade Level]] were the strongest predictors, followed by length of article
html, number of paragraphs, [[Flesh Reading Ease]], [[Smog Grading]], number
of internal links, [[Laesbarhedsindex Readability Formula]], number of words
and number of references. Weakly predictive were number of to be's, number
of sentences, [[Coleman-Liau Index]], number of templates, PageRank, number
of external links, number of relative links. Not predictive (overall - see
the end of section 2 for the per-rating score breakdown): Number of h2 or
h3's, number of conjunctions, number of images*, average word length, number
of h4's, number of prepositions, number of pronouns, number of interlanguage
links, average syllables per word, number of nominalizations, article age
(based on page id), proportion of questions, average sentence length. :*
Number of images was actually by far the single strongest predictor of any
class, but only for Featured articles. Because it was so good at picking out
featured articles and somewhat good at picking out A and G articles the
classifier was confused in so many cases that the overall contribution of
this feature to classification performance is zero. :* Number of external
links is strongly predictive of Featured articles. :* The B class is highly
distinctive. It has a strong "signature," with high predictive value
assigned to many features. The Featured class is also very distinctive. F, B
and S (Stop/Stub) contain the most information.
:* A is the least distinct class, not being very different from F or G. =
Latent features = The algorithm used for latent analysis, which is an
analysis of the occurence of words in every document with respect to the
link structure of the encyclopedia ("concepts"), is [[Latent Dirichlet
Allocation]]. This part of the analysis was done by CS PhD student Praful
Mangalath. An example of what can be done with the result of this analysis
is that you provide a word (a search query) such as "hippie". You can then
look at the weight of every article for the word hippie. You can pick the
article with the largest weight, and then look at its link network. You can
pick out the articles that this article links to and/or which link to this
article that are also weighted strongly for the word hippie, while also
contributing maximally to this articles "hippieness". We tried this query in
our system (LDA), Google (site:en.wikipedia.org hippie), and the Simple
English Wikipedia's Lucene search engine. The breakdown of articles occuring
in the top ten search results for this word for those engines is: * LDA
only: [[Acid rock]], [[Aldeburgh Festival]], [[Anne Murray]], [[Carl
Radle]], [[Harry Nilsson]], [[Jack Kerouac]], [[Phil Spector]], [[Plastic
Ono Band]], [[Rock and Roll]], [[Salvador Allende]], [[Smothers brothers]],
[[Stanley Kubrick]]. * Google only: [[Glam Rock]], [[South Park]]. * Simple
only: [[African Americans]], [[Charles Manson]], [[Counterculture]], [[Drug
use]], [[Flower Power]], [[Nuclear weapons]], [[Phish]], [[Sexual
liberation]], [[Summer of Love]] * LDA & Google & Simple: [[Hippie]],
[[Human Be-in]], [[Students for a democratic society]], [[Woodstock
festival]] * LDA & Google: [[Psychedelic Pop]] * Google & Simple: [[Lysergic
acid diethylamide]], [[Summer of Love]] ( See the paper for the articles
produced for the keywords philosophy and economics ) = Discussion /
Conclusion = * The results of the latent analysis are totally up to your
perception. But what is interesting is that the LDA features predict the WET
ratings of quality just as well as the surface level features. Both feature
sets (surface and latent) both pull out all almost of the information that
the rating system bears. * The rating system devised by the WET is not
distinctive. You can best tell the difference between, grouped together,
Featured, A and Good articles vs B articles. Featured, A and Good articles
are also quite distinctive (Figure 1). Note that in this study we didn't
look at Start's and Stubs, but in earlier paper we did. :* This is
interesting when compared to this recent entry on the YouTube blog. "Five
Stars Dominate Ratings"
I think a sane, well researched (with actual subjects) rating system
well within the purview of the Usability Initiative. Helping people find and
create good content is what Wikipedia is all about. Having a solid rating
system allows you to reorganized the user interface, the Wikipedia
namespace, and the main namespace around good content and bad content as
needed. If you don't have a solid, information bearing rating system you
don't know what good content really is (really bad content is easy to spot).
:* My Wikimania talk was all about gathering data from people about articles
and using that to train machines to automatically pick out good content. You
ask people questions along dimensions that make sense to people, and give
the machine access to other surface features (such as a statistical measure
of readability, or length) and latent features (such as can be derived from
document word occurence and encyclopedia link structure). I referenced page
262 of Zen and the Art of Motorcycle Maintenance to give an example of the
kind of qualitative features I would ask people. It really depends on what
features end up bearing information, to be tested in "the lab". Each word is
an example dimension of quality: We have "*unity, vividness, authority,
economy, sensitivity, clarity, emphasis, flow, suspense, brilliance,
precision, proportion, depth and so on.*" You then use surface and latent
features to predict these values for all articles. You can also say, when a
person rates this article as high on the x scale, they also mean that it has
has this much of these surface and these latent features.
= References =
- DeHoust, C., Mangalath, P., Mingus., B. (2008). *Improving search in
Wikipedia through quality and concept discovery*. Technical Report.
- Rassbach, L., Mingus., B, Blackford, T. (2007). *Exploring the
feasibility of automatically rating online article quality*. Technical
I have asked and received permission to forward to you all this most
excellent bit of news.
The linguist list, is a most excellent resource for people interested in the
field of linguistics. As I mentioned some time ago they have had a funding
drive and in that funding drive they asked for a certain amount of money in
a given amount of days and they would then have a project on Wikipedia to
learn what needs doing to get better coverage for the field of linguistics.
What you will read in this mail that the total community of linguists are
asked to cooperate. I am really thrilled as it will also get us more
linguists interested in what we do. My hope is that a fraction will be
interested in the languages that they care for and help it become more
relevant. As a member of the "language prevention committee", I love to get
more knowledgeable people involved in our smaller projects. If it means that
we get more requests for more projects we will really feel embarrassed with
all the new projects we will have to approve because of the quality of the
Incubator content and the quality of the linguistic arguments why we should
approve yet another language :)
NB Is this not a really clever way of raising money; give us this much in
this time frame and we will then do this as a bonus...
---------- Forwarded message ----------
From: LINGUIST Network <linguist(a)linguistlist.org>
Date: Jun 18, 2007 6:53 PM
Subject: 18.1831, All: Call for Participation: Wikipedia Volunteers
LINGUIST List: Vol-18-1831. Mon Jun 18 2007. ISSN: 1068 - 4875.
Subject: 18.1831, All: Call for Participation: Wikipedia Volunteers
Moderators: Anthony Aristar, Eastern Michigan U <aristar(a)linguistlist.org>
Helen Aristar-Dry, Eastern Michigan U <hdry(a)linguistlist.org>
Reviews: Laura Welcher, Rosetta Project
The LINGUIST List is funded by Eastern Michigan University,
and donations from subscribers and publishers.
Editor for this issue: Ann Sawyer <sawyer(a)linguistlist.org>
To post to LINGUIST, use our convenient web form at
From: Hannah Morales < hannah(a)linguistlist.org >
Subject: Wikipedia Volunteers
-------------------------Message 1 ----------------------------------
Date: Mon, 18 Jun 2007 12:49:35
From: Hannah Morales < hannah(a)linguistlist.org >
Subject: Wikipedia Volunteers
As you may recall, one of our Fund Drive 2007 campaigns was called the
"Wikipedia Update Vote." We asked our viewers to consider earmarking their
donations to organize an update project on linguistics entries in the
English-language Wikipedia. You can find more background information on this
The speed with which we met our goal, thanks to the interest and generosity
our readers, was a sure sign that the linguistics community was enthusiastic
about the idea. Now that summer is upon us, and some of you may have a bit
leisure time, we are hoping that you will be able to help us get started on
Wikipedia project. The LINGUIST List's role in this project is a purely
organizational one. We will:
*Help, with your input, to identify major gaps in the Wikipedia materials or
pages that need improvement;
*Compile a list of linguistics pages that Wikipedia editors have identified
"in need of attention from an expert on the subject" or " does not cite any
references or sources," etc;
*Send out periodical calls for volunteer contributors on specific topics or
*Provide simple instructions on how to upload your entries into Wikipedia;
*Keep track of our project Wikipedians;
*Keep track of revisions and new entries;
*Work with Wikimedia Foundation to publicize the linguistics community's
We hope you are as enthusiastic about this effort as we are. Just to help us
get started looking at Wikipedia more critically, and to easily identify an
needing improvement, we suggest that you take a look at the List of
Many people are not listed there; others need to have more facts and
added. If you would like to participate in this exciting update effort,
respond by sending an email to LINGUIST Editor Hannah Morales at
hannah(a)linguistlist.org, suggesting what your role might be or which
entries you feel should be updated or added. Some linguists who saw our
on the Internet have already written us with specific suggestions, which we
share with you soon.
This update project will take major time and effort on all our parts. The
result will be a much richer internet resource of information on the breadth
depth of the field of linguistics. Our efforts should also stimulate
students to consider studying linguistics and to educate a wider public on
we do. Please consider participating.
Editor, Wikipedia Update Project
Linguistic Field(s): Not Applicable
LINGUIST List: Vol-18-1831
It is an honor to announce that the Affiliations Committee has resolved
 recognizing the Wikimedia User Group China as a Wikimedia User
Group; their main focus areas are getting more chinese people know and
use Wikipedia, encouraging people to become contributors to the
different Wikimedia projects, and maintain the community healthy and
growing. Let's welcome the newest member of the family of affiliates
-and the fourth from the Sinosphere!
"*Jülüjain wane mmakat* ein kapülain tü alijunakalirua jee wayuukanairua
junain ekerolaa alümüin supüshuwayale etijaanaka. Ayatashi waya junain."
Carlos M. Colina
Vicepresidente, A.C. Wikimedia Venezuela | RIF J-40129321-2 |
Chair, Wikimedia Foundation Affiliations Committee
This is a personal note to clarify a some questions that recently came up,
specifically in the context of my role as the incoming ED.
My partner Wil and I are partners in our private lives. We have always both
been extremely independent, and we respect that in each other. That said we
have different roles: I am the Executive Director with responsibilities
towards the Foundation and the movement, and he is an independent community
member with his own voice.
I make my decisions using my own professional judgement in conjunction with
input from the community and staff. I don’t consult Wil on these matters,
ask him to do anything on my behalf or monitor his engagements with the
community. When I speak here, it is in my capacity as an ED.
Wil, on the other hand, has a very strong personal interest in the
community and agreat deal of curiosity about how the Wikimedia
projectswork. It is very important to him that he remains an
able to speak with his own voice and ask his own questions. He does not
take direction from me. He will not work for the WMF or engage with the WMF
I hope this addresses some of the questions and draws distinction between
my role as ED and Wil’s participation as an independent member. If you have
any questions for Wil you can reach him directly. If you have any questions
for me or the WMF, you can get a hold of me by email or on my talk page.
This is a response to Martin's note here:
.. and also a more general update on the next steps regarding disputes
about deployments. As you may have seen, Lila has also posted an
update to her talk page, here:
I want to use this opportunity to respond to Martin's and other
people's criticisms, and to talk about next steps from WMF’s
perspective following discussions with Lila and the team. I’m also
sending a copy of this note to all the stewards, to better involve
them in the process going forward.
I am -- genuinely -- sorry that this escalation occurred. We would
have preferred to avoid it.
I would like to recap how we find ourselves in this situation: As
early as July, we stated that the Wikimedia Foundation reserves the
right to determine the final configuration of the MediaViewer feature,
and we explicitly included MediaWiki: namespace hacks in that
statement.  When an admin implemented a hack to disable
MediaViewer, another local admin reverted the edit. The original admin
reinstated it. We then reverted it with a clear warning that we may
limit editability of the page.  The original admin reinstated the
hack. This is when we protected the page.
Because all admins have equal access to the MediaWiki: namespace,
short of desysopping, there are few mechanisms to actually prevent
edit wars about the user experience for millions of readers.
Desysopping actions could have gotten just as messy -- and we felt
that waiting for a "better hack" to come along (the likeliest eventual
outcome of doing nothing) or disabling the feature ourselves would not
be any better, either from a process or outcome standpoint.
Our processes clearly need to be improved to avoid these situations in
the future. We recognize that simply rejecting a community request
rather than resolving a conflict together is not the right answer.
We’ve been listening to feedback, and we’ve come to the following
- We intend to undertake a review of our present processes immediately
and propose a new approach that allows for feedback at more critical
and relevant junctures in the next 90 days. This will be a transparent
process that includes your voices.
- As the WMF, we need to improve the process for managing changes that
impact all users. That includes the MediaWiki: namespace. For WMF to
fulfill its role of leading consistent improvements to the user
experience across Wikimedia projects, we need to be able to review
code and manage deployments. This can be done in partnership with
trusted volunteers, but WMF needs to be able to make an ultimate
determination after receiving community feedback regarding production
changes that impact all users.
- We are prepared to unprotect MediaWiki:Common.js on German Wikipedia
and enter constructive, open-ended conversations about the way
forward, provided we can mutually agree to do so on the basis of the
current consistent configuration -- for now. We would like to request
a moratorium on any attempts to disable the feature during this
conflict resolution process. The goal would be to make a final,
cross-wiki determination regarding this specific feature, in
partnership with the community, within at most 90 days.
With regard to the German Wikipedia situation, we’d like to know if
stewards want to at all be involved in this process: In a situation
like this, it can be helpful to have a third party support the
conversation. Stewards are accountable to "valid community consensus
within the bounds of the Foundation's goals" , which seems to be
precisely the intersection of concerns at issue here. We would like to
suggest an IRC meeting with stewards ASAP to talk about the specific
question of stewards’ involvement, if any. If stewards prefer not to
be involved, we understand, but it's probably a good idea to have a
sync-up conversation regardless.
I hope we can move forward in good faith from here, and find better
ways to work together. As Lila has expressed, we believe there is a
need for a clear understanding of our role. It is as follows:
Managing software development, site configuration and deployment is a
core WMF responsibility. The community leads in the development of
content; the Wikimedia Foundation leads in the development of
Because these processes are deeply interdependent, we need to develop
better protocols for timely feedback and resolution of disagreements.
At the same time Lila’s and the Board’s statements make it very clear
that the WMF will not accept RfCs or votes as the sole determining
factor in global software deployments.
This means that technology and UX changes should not be decided by
vote or poll and then disabled at-will: where we disagree, we need to
talk to each other (and yes, that means a more judicious application
of RESOLVED WONTFIX on our end, as well!). We need to ensure a
process where the community voice is heard earlier at critical
junctions in the product development. All of this is consistent with
the principle of "shared power" articulated in the Guiding Principles
 approved by the Board of Trustees.
At the same time, as noted above and earlier, the superprotection
feature should be replaced with a better mechanism for code review and
deployment in the MediaWiki: namespace. This is discussed in  and
ideas and suggestions are welcome. Let’s be upfront about control
structures for production changes to avoid misunderstandings and
ambiguity in the future.
We are exploring options on how to improve dispute resolution
mechanisms -- whether it’s e.g. a standing working group or a better
protocol for responding to RfCs and engaging in discussions. We've
started a brainstorming page, here, which we hope will usefully inform
the process of conflict resolution regarding German Wikipedia, as
well, so we can arrive at a more concrete conflict resolution process
soon. Your thoughts/suggestions are welcome, so we can (in NPOV style)
look at different possibilities (e.g. workgroups, committees, votes,
surveys) that have been discussed in the past:
We’re absolutely not saying that WMF simply wants to be able to
enforce its decisions: we completely understand there need to be
mechanisms for the community to influence decisions and outcomes at
all stages of the development and release of software. We need to
arrive at this process together.
Again, we are sorry that this escalation occurred - and we hope we can
move forward constructively from here.
VP of Engineering and Product Development, Wikimedia Foundation
1. Where can I find a response from either the WMF board or WMF
funding/finance to the criticisms of a lack of transparency or the
apparent failure of the project to deliver value for the donor's money
as raised in this blog post?
2. Where can I read an officially recognized report for the outcomes
of this project in terms of value for Wikimedia projects? Obviously we
do not want to rely on second-hand analysis when reports to the WMF
are a requirement for such projects.
We know NSA wants Wikipedia data, as Wikipedia is listed in one of the
That slide is about HTTP, and the tech staff are moving the
user/reader base to HTTPS.
As we learn more about the NSA programs, we need to consider vectors
other than HTTP for the NSA to obtain the data they want. And the
userbase needs to be aware of the current risks.
One question from the "Dells are backdored"[sic] thread that is worth
separate consideration is:
Are the Wikimedia transit links encrypted, especially for database replication?
MySQL has replication over SSL, so I assume the answer is Yes.
If not, is this necessary or useful, and feasible ?
However we also need to consider that SSL and other encryption may be
useless against NSA/etc, which means replicating non-public data
should be avoided wherever possible, as it becomes a single point of
Given how public our system is, we don't have a lot of non-public
data, so we might be able to design the architecture so that
information isnt replicated, and also ensure it isnt accessed over
insecure links. I think the only parts of the dataset that are
private & valuable are
* passwords/login cookies,
* checkuser info - IPs and useragents,
* WMF analytics, which includes readers iirc, and
* hidden/deleted edits
* private wikis and mailing lists
Have I missed any?
Are passwords and/or checkuser info replicated?
Is there a data policy on WMF analytics data which prevents it flowing
over insecure links, and limits what is collected and ensures
destruction of the data within reasonable timeframes? i.e. how about
not using cookies to track analytics of readers who are on HTTP
instead of HTTPS?
The private wikis can be restricted to https, depending on the value
of the data on those wikis in the wrong hands. The private mailing
lists will be harder to secure, and at least the English Wikipedia
arbcom list contain a lot of valuable data about contributors.
Regarding hidden/deleted edits, the replication isnt the only source
of this data. All edits are also exposed via Recent Changes
(https/api/etc) as they occur, and the value of these edits is
determined by the fact they are hidden afterwards (e.g. don't appear
in dumps). Is there any way to control who is effectively capturing
all edits via Recent Changes?
For some months, Twitter have been blocking most URLs in their direct
messages (DMs), supposedly as an anti-spam measure.
Do we have someone who has a contact there, who could ask them to
whitelist Wikimedia project URLs in DMs?
to increase accountability and create more opportunities for course
corrections and resourcing adjustments as necessary, Sue's asked me
and Howie Fung to set up a quarterly project evaluation process,
starting with our highest priority initiatives. These are, according
to Sue's narrowing focus recommendations which were approved by the
- Visual Editor
- Mobile (mobile contributions + Wikipedia Zero)
- Editor Engagement (also known as the E2 and E3 teams)
- Funds Dissemination Committe and expanded grant-making capacity
I'm proposing the following initial schedule:
- Editor Engagement Experiments
- Visual Editor
- Mobile (Contribs + Zero)
- Editor Engagement Features (Echo, Flow projects)
- Funds Dissemination Committee
We’ll try doing this on the same day or adjacent to the monthly
metrics meetings , since the team(s) will give a presentation on
their recent progress, which will help set some context that would
otherwise need to be covered in the quarterly review itself. This will
also create open opportunities for feedback and questions.
My goal is to do this in a manner where even though the quarterly
review meetings themselves are internal, the outcomes are captured as
meeting minutes and shared publicly, which is why I'm starting this
discussion on a public list as well. I've created a wiki page here
which we can use to discuss the concept further:
The internal review will, at minimum, include:
Team members and relevant director(s)
So for example, for Visual Editor, the review team would be the Visual
Editor / Parsoid teams, Sue, me, Howie, Terry, and a minute-taker.
I imagine the structure of the review roughly as follows, with a
duration of about 2 1/2 hours divided into 25-30 minute blocks:
- Brief team intro and recap of team's activities through the quarter,
compared with goals
- Drill into goals and targets: Did we achieve what we said we would?
- Review of challenges, blockers and successes
- Discussion of proposed changes (e.g. resourcing, targets) and other
- Buffer time, debriefing
Once again, the primary purpose of these reviews is to create improved
structures for internal accountability, escalation points in cases
where serious changes are necessary, and transparency to the world.
In addition to these priority initiatives, my recommendation would be
to conduct quarterly reviews for any activity that requires more than
a set amount of resources (people/dollars). These additional reviews
may however be conducted in a more lightweight manner and internally
to the departments. We’re slowly getting into that habit in
As we pilot this process, the format of the high priority reviews can
help inform and support reviews across the organization.
Feedback and questions are appreciated.
VP of Engineering and Product Development, Wikimedia Foundation
Support Free Knowledge: https://wikimediafoundation.org/wiki/Donate