Hacker News new | past | comments | ask | show | jobs | submit login
Redesigning Wikipedia: The Athena Project (wikipedia.org)
119 points by friggeri on Aug 13, 2012 | hide | past | favorite | 26 comments

The biggest problems I have with wikipedia are:

1) deletionists 2) secret editor cabals

Wikipedia is stunningly open to new editors, unless you create a page that someone else doesn't like. I've seen multiple people get turned off from Wikipedia when they create two or three pages and each gets hit with a speedy deletion within days.

The other big problem with Wikipedia, for slightly more advanced editors, is that the ethics of the site frown on recruiting allies for revision wars...and yet there is tons of evidence that there's a vast world of behind-the-scenes editor politics, factions, rivalries, and Joe User can make a change, or weigh in on a controversy, and a dozen editors can jump in at once.

After being burned a few times I've avoided the cliquey corners of wikipedia and mostly just work on things like spacecraft docking standards...but not everyone is content to work on out of the limelight areas.

I think it was Clay Shirky who once said that it was a mistake to try to solve social problems with technological fixes.

I think Wikipedia's big problems are social problems.

(Disclaimer: it's still an amazing site, and I'm a big fan...but it could be better.)

When I went back to Wikipedia after an absense of about 3 years or so, I found that there were some remarkably intolerant and rude editors who seemed to be able to do anything with impunity. Luckily, the main editor in question has since left, but it was ridiculous.

Regrettably my depression flaired up while dealing with the issues, and I left the project (again), but this time most definitely for the very last time.

Also regrettably, the Admin's Noticeboard - which I founded - seems to be a hotbed of intrigue and oftentimes is more a hindrence to the project than a help. It's still necessary, but I suspect that it has centralized power far too much. It may have changed as it's more than a year since my last foray on the project, but that was my experience.

I hear stories like this, but have not run into it myself, with the exception of areas that are inherently contentious anywhere, such as Israel-Palestine or current elections. What kinds of areas is it in? I most often run into the opposite problem of just not enough people around. I'll ask for comments on a talk page and nobody will answer for months, if ever; and even some of the WikiProjects are ghost towns if I ask a question over there. The people who do answer are usually polite and helpful. I mostly edit in history- and geography-related areas, along with a little bit of computer science. Are those unusually polite areas?

If anything, I personally find Wikipedia-writing a bit of a relaxing break from the often-rancor-filled tone of other online activities (even HN can have an edge to it). Digging up some books from the library and using them to write a new article or two, along with adding a few citations to other articles, etc., passes a pleasant afternoon.

Well, I started editing the Logic article in 2004 and "semi-retired" in 2009 for the kind of reason you see at


And since then I've occasionally looked at the main article and seen how the text has gradually become less readable and more error-ridden over time. For instance, following collaborative research discussing about 20 sources, I changed the first sentence to 'Logic, from the Greek λογική (logiké)[1] is defined by the Penguin Encyclopedia to be "The formal systematic study of the principles of valid inference and correct reasoning".[2]' in


However, people have their own ideas about what the topic is, and so the lead sentence has become the longer, vaguer, less cogent, unsourced, and, as a recent commenter has pointed out, incoherent two sentences 'Logic (from the Greek λογική logikē)[1] refers to both the study of modes of reasoning (which are valid and which are fallacious)[2] and the use of valid reasoning. In the latter sense, logic is used in most intellectual activities, including philosophy and science, but in the first sense is studied primarily in the disciplines of philosophy, mathematics, semantics, and computer science.'


(note that this [2] supports the old sentence, not the current two sentences)

All this in an article that the Wikipedia 1.0 project counts as amongst its 100 most important.

I liked editing maths articles - it seemed to attract a nice, well-informed, and result-focussed crowd. I value Wikipedia as a resource for a range of topics, such as network protocols. I liked editing philosophy articles in 2004, but I stopped enjoying it in 2006, and stopped believing that Wikipedia was a useful resource for such topics around 2009.

Good point. General articles like "logic" are contentious, which I had forgotten as I tend to avoid them. I feel Wikipedia is not at its strongest there, since it tends to require synthesis that is almost inevitably opinionated and subjective. Most good review articles in journals are opinionated critical surveys that make novel syntheses (e.g. propose a new taxonomy), which is not really Wikipedia's role.

Though, having participated in some attempts in academia to come up with consensus summaries on subjects, I think the problem may be inherently hard rather than Wikipedia-specific. If you get 10 prominent AI researchers, from several traditions, in a room and ask them to write a 5-page general introduction to AI, there will almost certainly be sparks flying. You quickly run into the trouble (as in the logic article) of jockeying over what sub-areas deserve prominent mention, how to conceive of and define the subject, etc. And the results inevitably sound "written by committee", since the intro has to include a nod to everyone's area.

I've lately taken a more systemic view of Wikipedia coverage, where my goal is to improve Wikipedia's network of articles in an area rather than mainly the top-level overview article. So I care less about what goes in [[History of Greece]], and more about the thousands of articles (most of which don't yet exist) on historical Greek people, places, events, etc. I think that network-of-coverage aspect is one of Wikipedia's strengths as a hypertext encyclopedia, whereas the overview articles are a strength of more traditional linear-text encyclopedias like Britannica. It also feels easier to make incremental progress in that approach: I sometimes just pick up a random history book from a library shelf, and start asking what it covers that Wikipedia doesn't yet cover, then begin adding that.

>I think that network-of-coverage aspect is one of Wikipedia's strengths as a hypertext encyclopedia, whereas the overview articles are a strength of more traditional linear-text encyclopedias like Britannica.

Hypertext for the win! May the power of the web be with you :)

I've had trouble with art-related articles.

After the most recent "well, ten minutes in Google would have been easier and more constructive than several days of telling people it's not Notable" incident, the deletionist kid involved at least had the good grace to be embarrassed at their actions.

I have authored many art related articles and corrected/augmented many others, and so far aside from the usual bots that clean up and fix typos and formatting errors, I have not seen any problems with deletions.

>I think it was Clay Shirky who once said that it was a mistake to try to solve social problems with technological fixes.

Was this http://shirky.com/writings/group_enemy.html#content what you were referring to?

It's utterly baffling to me to see them specifically call out editor gender disparity as an issue and then seriously suggest that a ui redesign will fix this.

The issue isn't the software, it's people. People being allowed to be terrible to other people. People pushing euro/american-centric, white-centric, male-centric viewpoints. People creating a feedback loop of intolerance by making the place only habitable by the intolerant.

You can't graphic design these problems away.

There is some evidence (I believe from a commissioned study) that the current interface disproportionately filters out non-technical would-be editors from contributing, because of its heavy use of markup in a plaintext editing field. Stuff like:

   <ref name="citekey">{{cite book | title = ... }}</ref>
is impenetrable to some would-be editors, especially when it starts getting more complex and nested. Even as a long-time editor with a CS background, I sometimes find it hard to skim past big blocks of markup to find what I want to edit, and it's easy to screw things up and end up with unclosed ref tags and that kind of thing. To someone who doesn't even know what an "unclosed tag" is (most people), I can only imagine it'd be many times more confusing.

Since people with technical backgrounds are disproportionately male, it could be a contributing factor to the gender disparity. Hardly the only one, but a WYSIWYG editor may at least broaden the population being drawn from.

If the site were more user-friendly, it would be easier to use for people who don't understand the initial complexities of Wikipedia. One of the contributors on that page wrote against a redesign because they don't want to have to re-learn how to use the site, insinuating that the initial learning curve was rough for someone who is an editor, not a web person.

The redesign should include better promotion for the site/make people aware of the contribution process just as much as it should focus on the content management and discussion portions.

Those -centric are problems, but there are others too; in a great many articles people with specific interests tend to form cliques that rigidly enforce certain viewpoints, not objective consensus or broader cultural consensus, just what those users insist should be seen.

Those interest cliques are very often nothing to do with being euro-centric or male-centric.

The real solution to 'deletionism' is to start your own wiki and impose your own policy. If your content is good enough for others, and licensed right, it can be copied to whatever other sites (wiki or not) want it; if it isn't, who cares? It will always be on your little wiki, available to the world.

Wikipedia has no obligation to host everything you want to put on it.

We're not talking about how difficult it is to get text onto the internet, we're talking about how Wikipedia, the community, could in our understanding benefit from certain changes.

The solution to a problem in a community isn't to go start your own community somewhere else.

Now this is how you go about redesigning a project like Wikipedia. All due respect to those other guys ( see http://news.ycombinator.com/item?id=4352290), but Wikipedia doesn't need a branding makeover. Everyone knows Wikipedia, but not everyone contributes. Why? It's a pain in the ass to do.

A better Wikipedia starts with streamlining the editing process (although something can be said in favour of a healthy barrier to keep the trolls out). Making Wikipedia look prettier is a nice extra to have in the process.

That branding makeover wasn't designed to be actually used. It was, as companies are prone to do, a case study/exercise with the intent of showing clients their though processes with a recognizable product. I wouldn't criticize those gents to strongly because they are a branding company and they were using Wikipedia to demonstrate the thought processes that go into a project. It's essentially useful as similar exercises that we got in college -- it isn't about actually implementing the changes, its an academic exercise. In the case of the design firm that did the Wiki remaining, it serves a similar purpose.

I think HN was especially hard on those guys despite the fact that it achieved its intended purpose: to start a conversation. Even bad ideas have a purpose in the design process (although I have to admit, I actually liked their work.) I'd love to see them do a similar case study on Craigslist or the Drudge Report, if just to see what they came up with.

I agree with most of this, except I'm a little skeptical of the claims in the "More bodies means better articles" section, such as "featured articles are so good precisely because they are edited by so many". Many featured articles are good precisely, to the contrary, because a small number of people did the library research to write a good article. This doesn't necessarily have a 1:1 correspondence with the views discussed in the article: When a history article includes coverage of 4 major historiographical views on a controversial subject, often all four of them were added and summarized by the same editor who is describing the debate, not by separate editors each representing one of the viewpoints.

That does seem to vary by area. The featured articles on major recent news subjects may be closer to fully crowdsourced, piecing together hundreds of sources, mostly newspapers/magazines and mostly available online. The good history articles tend to be more often written by smaller numbers of historians or history enthusiasts who have read and can cite proper books and journal articles on the subject. For example, one user I've run across is single-handledly responsible for hundreds of lengthy articles on classical antiquity, with >80% of the content and >95% of the references in each typically being hers: http://en.wikipedia.org/wiki/User:Cynwolfe

I think more and more diverse editors could have many positive benefits, but I think quality is more important than quantity. Some areas have very good editors. Mathematics comes to mind as probably the best in terms of having editors who are very knowledgeable in the field and also interested in writing Wikipedia articles; I was tickled to find out that mathematician David Eppstein is a prolific Wikipedian, among others. But some areas lack much in the way of editors who are knowledgeable in a subject and willing to do the library research needed to produce well-cited articles. Those areas could be greatly improved if they had the analog of folks like Eppstein (or Cynwolfe) writing articles.

The biggest problem for me is lack of clues of importance of the subject. Somehow I learned from printed encyclopedias that the longer the article the more important the subject is (in context of the subject of course). Take music. According to impression i got from Wikipedia, there's almost no important music made or recorded before, let's say 21st century :-) Or more specifically - http://en.wikipedia.org/wiki/4_(Beyoncé_Knowles_album) - seems to be one of the most important, influential record ever made?

Also, the most important things are ones which are mentioned in many different anime shows, all of which MUST be mentioned in the article, even if it takes pages to do so

Possibly of interest to readers of this link, Wikipedia is hosting an engineering meetup in San Francisco on August 15: http://www.meetup.com/Wikipedia-Engineering-Meetup/events/69...

Looks like a chance to talk about these projects and ideas with people who work on Wikipedia full-time, including on the mobile interfaces and visual editor.

I'm all for improving the functionality of Wikipedia (e.g., the visual editor and Flow), but I'm very much against changing the look of the site. Wikipedia's layout is simple and puts focus on the content instead of the design.

>> puts focus on the content ...

That's the purpose of good design. Design has but one purpose, to serve the UX. That's it. Anything else is just vanity.

When there's an obvious focus on "design", then the UX is harmed. I put design in quotes because we all know what I'm talking about, design that is more about artistry than utility. Real design is much different from art. While great design can be artistic, design is all about purpose and not that nebulous "What's the purpose of art" type of purpose, but actual utility. Getting something done. Solving a problem.

The Drudge Report is considered by many on the internets to be a terrible "design" yet it's exceptionally well designed for what it does: it allows people to get a very fast listing of headlines curated by someone with a specific point of view. The purpose is precisely clear -- it's a place to browse headlines and perhaps visit links. Going to the Yahoo page, by contrast, I'm left wondering what the heck the purpose of yahoo is. Presumably, it's a news homepage, but it's incredible badly designed because the purpose of the page isn't really clear. It looks like Yahoo hired rejected Naver (http://www.naver.com/) designers and told them to ruin the Yahoo homepage as much as possible while maximizing ad pixels.

Anyway, I know I'm rambling a bit, but I think the "look" of wikipedia could be improved slightly, but the design (at least on the consumer facing (not the editor) side is effective. I especially appreciate the mobile version of Wikipedia. It's a constant companion while I'm out and about and it's fast, crisp and gets me the information I need.

I think you've hit the nail on the head here - there's an enormous amount of pure visual-design influence in UIs these days, prioritizing aesthetics over usability. We can see it in this proposed Wikipedia redesign, in Microsoft's Windows UI overhauls, and in plenty of other websites.

There are major problems with what we see in the 'Athena' screenshots. Why are the navigational controls at the top of the page - ones that expose functionality applicable to the entire site, not the current article - skinned with a photograph taken from the current article? What does a picture of the Beatles convey regarding how to find information on Wikipedia?

The author says that the Athena skin "emphasizes content", but it causes article content to bleed through somewhat into the functional UI, compromising both. This isn't a good thing.

This redesign, at least as shown in the screenshots, is rather underwhelming. The new UI elements just seem to add a lot of clutter and distract attention from what is important.

Wikipedia has two core problems. At the level of merely an encyclopedia they've been able to cope with them but as wikipedia becomes more relied upon they will increasingly cause difficulties.

First, there is a fundamental misalignment between the people who are actually responsible for contributing the bulk of the value of wikipedia and the people who "run" the site. To the degree such that the people who are in charge are generally clueless where the value comes from. Almost all of the value comes from content edits, which come from the long tail of contributors, many of them anonymous. The admins and mods who run wikipedia generally contribute only formatting and metadata changes. Rearranging, fixing the layout, hooking up keywords and so on. However, there seems to be the impression within the wikipedia org that the 2nd group is responsible for the majority of value. So they structure guidelines and so forth with that in mind, sometimes to the detriment of the content.

Second, there are very flimsy standards of "evidence" for wikipedia entries. Overall this isn't such a bad thing because in reality most encyclopedia authoring groups are no better at ensuring they print the truth, but in wikipedia's case more readers will apply a critical eye and double check sources, which is all for the better ironically. However, as wikipedia becomes more and more trusted the need to rely on truly reliable sources becomes more and more important. And, as the internet becomes the primary method we communicate and store knowledge the need to store and catalog evidentiary sources becomes a more pressing concern than having a convenient encyclopedia at your fingertips.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact