Scripting News for 3/6/2006

Essay: What is an unconference? “If you swapped the people on stage with an equal number chosen at random from the audience, the new panelists would effectively be smarter, because they didn’t have the time to get nervous, to prepare PowerPoint slides, to make lists of things they must remember to say, or have overly grandiose ideas about how much recognition they are getting.” 

Mark Evans, a reporter in Toronto, suggests that we actually put four random people from the former audience on stage. That’s not exactly an unconference, but it is an interesting idea. πŸ™‚ 

A picture that was taken at a non-unconference. Another

I got locked out of Gmail again. Oy. Just checking email, sending some email. I’ll fill out the form again. Sigh. 

Rakesh Agrawal’s report from BarCamp in New Dehli. 

Question: What icon to use for an “include” nodetype? 

AOL: “The Open AIM program enables companies, communities and developers to sell advertising and access, and to facilitate up to 250,000 log-ins per day or two million log-ins per month, without additional licensing requirements.” 

Late last week I got an email from Ray Ozzie at Microsoft, and like the discussion we had about SSE late last year, this one promises to result in a new idea entering the web app space. In about 15 minutes (at 8:45AM Pacific), Ray will start Tues morning Ray will give a talk at a conference in San Diego, at which time I will be free to discuss the idea here. 

New header graphic, the Mississippi River at Cass County, MN. 

Wired: “I just hope when they’re done remodeling our living rooms, we’ll still be able to use our legally purchased content the way we want to.” 

N.Y.T.i.m.e.s article about R.S.S. 

A subtle change 

It was a subtle change at Google, but a profound one. It happened some time ago, without any fanfare, and at first confused me, so I haven’t written about it until now. But it’s major. Here’s the deal.

On some subjects, where there is a DMOZ editor, the editor’s descriptions of each page, right or wrong, biased or not, are included with each entry. I’m not going to give examples, because if I do, people will focus on the examples, not the practice.

DMOZ has a bad rep for having editors with conflicts of interest. And it’s exclusive, unlike Wikipedia which at least has battles (never thought I’d say that) of people with conflicts, DMOZ doesn’t even have dissent among conflicted people, only one point of view exists, because there’s only one editor for each category.

Because Google accepts DMOZ as authoritative, it also accepts the conflicts of the editors as authoritative. Bzzzt, that’s a bug. The opportunities for payoffs and bribery are incredible, so incredible it must be happening. No one supervises the editors. That Google is willing to give them editorial control over what Google says is surprising, to say the least. It’s almost unforgiveable.

Basically Google has become an, but with much more power, and without the scrutiny. They know they’re doing it, and so do people who are watching Google, like me. How do we have a discussion about this?

26 responses to this post.

  1. RE: DMoz

    Dave, you and I must be thinking of two different things with the same name and URL, ’cause the DMoz that I worked on was the site of major edit battles, frequent, harsh peer reviews, and so on.


  2. Posted by Elle on March 6, 2006 at 10:03 am


    Your information on DMOZ is incomplete. There can be several editors per category, many categories aren’t actively edited, and there is a certain amount of discussion and dissent, both productive and unproductive, that goes on.

    That said, your basic point, that DMOZ is unreliable and overrated by Google, is right.

    You’ve made the same point about Wikipedia, previously. The same has to be said about blog content, too. Much of it is junk, and Google overrates a lot of popular junk

    There hasn’t been enough discussion about the fact that Google’s results have become more of a link popularity contest than a measure of content’s value. In an ideal world, the sites with the best content would get the most links. In reality, it’s common for insubstantial content at popular sites to get massively linked and overrated by Google, at the expense of a lot of good content that may not be sensational, trendy or snarky.

    So – you can’t criticize Google’s handling of DMOZ or Wikipedia without including Google’s handling of blog content. Our sometimes utopian vision of the blogosphere needs to be tempered with the reality that Google is overrating blog output, and using bloggers to undermine the value of Web content creation.

    It’s been said that if you give an infinite number of monkeys enough time, they’ll produce the works of Shakespeare. Google the output of these monkeys, and you’re likely to find a few classic works in the results. You’ll also find a lot of junk, though, that’s popular with the monkeys.

    Google’s approach is throw AdSense peanuts to millions of bloggers, overrate blog output, and create a lot of happy monkeys. Given enough bloggers, Google’s approach reduces the value of content to almost nothing, making it difficult for traditional publishing models to survive.

    Google’s overrating of DMOZ, Wikipedia and blog content promotes a Web where Google makes billions, content producers make peanuts and spending the time and money to create high-quality content is riskier than ever.


  3. Elle, there are search terms, important ones, where one conflicted editor gets to position every hit on the page. I know of one category where the guy with the conflict sells links on his blog, so it seems pretty likely he’d do the same for the Google result page.

    I don’t place any more or less faith in bloggers than I do professional journalists. I have a career worth of experience to back that up. And in one of the categories, the conflicted DMOZ editor *is* a professional journalist, for what you would almost certainly consider a reputable pub. We live in a very dirty world, and to single out bloggers is to miss the bigger picture. Yes, bloggers are bad. I think at times they are even worse than the pros. But the pros are pretty bad too.

    I still think Google made a big mistake by taking DMOZ data and integrating it into search results.


  4. A friend of mine saw the same issue with DMOZ and started up an “improved DMOZ” site at The site is an implementation of DMOZ on a Wiki, making the end result more democratic.

    Of course, the limitations of a Wiki do apply. It’s a significant improvement in process. The end result at the site is organically (read:slow) growing and waiting to see what happens.


  5. Dave:

    As one of the founders of the Open Direcotry Project, I share some of your frustrations with the current state of — Couple of things to point out.

    – The Google Directory has always used dmoz data (it’s the basis of the directory)
    – Search results have used dmoz descriptions since the launch of the directory, should they exist for a site — since around 2000
    – The Google directory (and the dmoz listings) have been demoted in rankings and authority aftger being taken off of the Google home page
    – As noted, there is a vast system of peer review, multiple editors per category (such as with multiple editors on the bottom of the page) and a hierarchy of control up the chain. There is a forum and discussion ssytem within dmoz that gets thousands of posts per day discussing the directory

    So — the search results containing the dmoz data has been the case for several years, and often results in a really nice summary (and, admittedly, some biased site spam, as well)

    The real issue here is that dmoz is down to a very very very small staff at the moment, and has been on life support at AOL for some time.

    How do you start a discussion?

    Well, you could volunteer and become and editor in the directory. Should you do so, let me know, and as a meta-editor, I’ll log in and persoanlly review your application.

    With 60,000 editors signed up, it’s a crying shame that things aren’t working better. The system is simply overrun with spammers, and there has been no upgrade to the core system since the last dmoz full time coder left a few years ago (to my knowledge). It’s a credit to the volunteers who really run the place that it runs at all.

    Chris Tolles
    ODP nee NewHoo nee GnuHoo


  6. RE: Google Lock Out

    Dave, do you ever get the feeling that Big Brother may be watching you? You comment questioning Google’s motivations for something and then suddenly Gmail isn’t working. Its subtle, but effective. Eventually your subconscious starts working out that every time you cast aspersions on Google you get punished by losing access to Gmail. Finally,your subconscious starts trying to use reverse psychology on Google and only says good things in an effort to maximise your uptime.



  7. The idea of the un-conference made me remember those 7-Up commercials for the “un-cola”. They didn’t work, the market share didn’t budge.

    How would you promote an un-conference? “Turn up and see what happens?” I just think it would be hard to motivate people to come to something without getting some form of expected return. Unless of course you take away the concept of “going” and “time”. Maybe an “un-conference” is a free-flowing online discussion where everyone is free to participate and take from the conversation whatever they like.


  8. Chris, thanks for checking in here. I’m going to continue the thread tomorrow morning on Scripting News, but for now, I think the answer is to decentralize.

    Here’s a scenario. I don’t think the editor for Park City, Utah is being fair, because he comes from Snowbird, and has a reason to want Park City to wither and die. So I start my own directory. I don’t apply to anyone for the right to do this, any more than I have to apply to anyone for the right to put up a web page about skiing in Park City. Then I do a great job, and promote my directory to ski directories, and ask them to link to me instead of the bad guy, and if that doesn’t work, ask them to amalgamate us, to include both of us, in aggregate.


  9. Dave:

    What you describe *is* one way to do it….

    There is another way, within the ODP structure — which is to go and join the fray..become an editor and go in that same category and have at it…utilize the huge society that’s grown up within the dmoz community and make the directory better and use the arbitration facilities built up to resolve disputes.

    The social system that the ODP is pretty interesting, and while crufty and spam filled — the community dwarfs the number of poeple working on most other open projects.

    The idea is that dmoz is like a standards body — so all the various parties get to fight it out within the directory and come to rough consensus and a running directory. We even had the scientologists and the anti-scientologists more or less in equilibrium. You could propose your own set of sub directories, if you wanted the Park City sucks sub directory…

    Now, practically speaking, there are certainly issues with the dmoz, and it would be pitifully easy to enumerate the mess of things wrong with it — but the output *is* open for use, so you could also just glean what you’d like out of it (with the proper attribution), dropping the rest.

    Alternatively, you could build a parser sesigned to identify biased language and filter the descriptions — As what Google is doing here is benefiting off of the 4M site descriptions (many of them pretty good).. They don’t really care about the petty rivalries within the ODP, and as long as most of them are reasonable, they’ll likley keep using them.

    However bad it is, I’ll point out that dmoz managed to create the largest human edited directory of the web, built by 71k editors, and create the largest RDF, open content play out there today…

    So, while you can (easily) take (more than fair) pot shots, I’d encourage you to try and play within the rules that are set up, or at the very least, check out the society and culture of the thing.



  10. It would be a waste for me to try to fight with the people who own the categories I care about. I’ve tried, in other contexts, they’re better fighters than I am, and they win because they’re willing to keep at it when I get tired of fighting. That’s the problem with DMOZ-like approaches, the more negative you are the more likely your point of view will prevail. That’s why Google, for a while, was the best approach, it tended to reward people who provide a good service to people who care about a subject. it still is the best thing we have for that, but it’s getting diluted. I want to try the same approach with directories.


  11. […] An interesting discussion about DMOZ has started in the comments on yesterday’s Scripting News.  […]


  12. I’m not sure I’d agree that the more negative viewpoint wins out — Perhaps the more strongly held viewpoint, or the folks who have more at stake (either financially or personally).

    In other words, just like any other political situation.

    Mind you, you’ll find a lot of folks that will agree with you that dmoz has issues (which is, as one of the founders a sad thing) that are not being addressed. Finding a solution that doesn’t simply bring all the problems with it would be a neat trick.

    Mind you, the issues you bring up here are some of the big reasons that the team that built dmoz went on to do things with algorithms in, rather than building something that allowed people to vote up their favorite stories (a la Digg). Giving people a voice to talk back is important – but don’t let the tragedy of the commons actually be the centerpiece of your business model πŸ™‚



  13. Chris

    As an ex summarily removed editor of the DMOZ I will tell you straight up you have editall’s over their with an agenda, when I was a editior I single handedly built the US Military resoruce section from scratch to over 3000 listings. Being active duty Navy I knew where the resources where that mattered. I also edited several other categories overall having nearly 10,000 links credited in adding

    For 2 years I edited in peace until a editall started coming in and deleting links that all were related to the history of WWII, it was obvious that the person did not like some of the sites that were legitimately being added as historical resources. Then it got worse full sub categories were deleted, appealing to the editall to stop was not effective.

    Not long after that a new editor was assigned to the category, it was someone residing outside the United States and a few weeks after that I was locked out no warning nothing. The new editor then went to work and destroyed much of the work I had done. What pissed me off the most is you had someone from outside the United States ediing a category he had no business in, it was obvious the editor had the same political views as the new editor thus you had a editall who made it her business to destroy two years of work.

    This is one example of thousands of editors your jacked up system has quite literally crapped on. I don’t trust the listings, I dont trust the politics of the editall. I have no longer visit the directory and obviously have very bad feelings about the whole experience. Here is a post I put up a couple of years ago on the subject if you care to read.

    Plain and simple from my perspective and I am going to be harsh, it only takes a few peaches to spoil the whole batch and you have some very stinking corrupt rotten peaches over their.


  14. Yahoo also uses Open Directory descriptions, and I think MSN may at times, as well. Also, Google doesn\’t always use them, each and every time. Nor is the use new — it\’s gone on for years. What is new is the descriptions — and sometimes titles — are being used more frequently. And yep, that\’s a concern.

    Here\’s a proposed standard I put out about this last year:

    Here\’s a discussion we had with search reps about it after that:

    I still need to do my own final recap of this and the longer push, but the consensus was I felt that site owners wanted a meta tag/method to indicate to the search engines that in no circumstances should they use a third party description of their pages. You can\’t just go with the logic of \”if you see a meta description tag, only use that,\” because some meta descriptions are blanks (the authoring tool does that by default, for example) and sometimes grabbing a snippet of text from the page itself is more useful than the meta description tag. But fair to say having your page be described — sometimes inaccurately — by a third party with no recourse isn\’t acceptable.


  15. Danny, I gotta say it\’s good to see we finally agree on something! πŸ™‚


  16. Where are these friggin slashes coming from!

    Matt must be monkeying with the WordPress code again.

    😦 😦 😦 😦 😦


  17. Hey Dave, we agree on lots πŸ™‚

    I think the slashes are related to apostrophes, in some way. For example:



  18. Posted by Elle on March 7, 2006 at 11:38 am


    Now that you\’re posting to a WordPress blog and allowing comments, using daily posts vs. individual posts seems to work against developing threaded discussions.

    When there are a bunch of comments about DMOZ with an unconference comment thrown into the middle, it doesn\’t seem to do justice to either topic. It\’s off-topic to people following the DMOZ discussion, and the unconference comment may get lost.

    Would there be a better way to map your daily topical posts into WordPress items?


  19. The challenges at Wikipedia and DMOZ are ones of structure, I think. Hugely decentralized operations can have these problems. No one is really in control. Infighting breaks out. Decision-making gets paralyzed. Etc.

    Jo Freeman wrote the classic on this in 1971 as the women\’s movement was moving out of living rooms into political organizing and discovered their previous\”consensus\” decision making process no longer worked.

    Her article is titled\”The Tyranny of Structurelessness\” and has been called a \”classic article on the manipulative tendencies of unstructured groups.\”

    Her thoughts applied then and now, and to any number of political groups – as well as, I think, to organizations like DMOZ and Wikipedia.


  20. I often wonder how DMOZ picks their editors and sites???

    Must be a cartel of some kind πŸ™‚

    I have submitted several times to no avail. Maybe it is because we are a search engine.



  21. Todd:

    I’m very sorry that you had a bad experience with dmoz — you’re pointing out some of the worst issues, and I can’t say that I disagree with your conclusions…

    Personally, I was hoping that AOL was going to either staff up or sell dmoz to someone who would put more effort into it. There are a lot of things that could be done technologically to make the whole thing work a lot better.



  22. The assertion is that having a _listed_ editor = descriptions and titles printed _verbatim_. Is that right? If so, Google ought to stop that. But, I have a hard time believing Google is doing it. Can you give examples that prove this?

    Some of the busiest editors deliberately pull their names off of categories to remove the perception of “ownership” of certain little categories that they can edit anyhow. Perhaps Google could/should give some weight to nearly top-level domains, but little piddly categories ought to not get such treatment. That’s where the bias can really creep in, in the tiny “leaf” categories.


  23. Posted by nothanks on March 14, 2006 at 1:09 pm

    did you really have to make it so difficult to read your website? light blue, white and light gray? i mean really, what a horrible decision. and the font is extremely small to boot!


  24. Posted by chris2001 on March 14, 2006 at 11:56 pm

    > On some subjects, where there is a DMOZ editor, the editor’s descriptions of each page, right or wrong, biased or not, are included with each entry.

    As KC, I doubt that Google would set up an algorithm that depends on a ODP category having an editor or not.

    > Yahoo also uses Open Directory descriptions, and I think MSN may at times, as well.

    Yahoo has been among the ODP data users in the past (~2004 if I remember correctly), but it doesn’t use ODP data today afaik. If you think they still do, Danny, would you mind pointing me to an example?
    MSN is using ODP descriptions since beginning of the year, see the attribution at

    Some updates related to what Chris Tolles said:

    > The system is simply overrun with spammers,

    Not more and not less then the overall web and any other system living in it, especially systems where people hope to get an advantage for free. ODP has faced this situation for years, and has by now collected lots of spam- and abuse-related knowledge, tricks and tools – several of them pretty recent developments (Chris, you may not know them yet – I’d be happy to give you a tour around the place!). So, whatever new spam trick comes up next, chances are good that the ODP editor community will find a way to address it.
    Many younger social search / open content / information sharing projects still have to make that experience of being spammed and abused – how many of them will ever get big enough to be hit by the problem fully, and how many of these will be able to learn enough to survive, I wonΒ΄t try to predict…

    > The real issue here is that dmoz is down to a very very very small staff at the moment, and has been on life support at AOL for some time.
    > Personally, I was hoping that AOL was going to either staff up or sell dmoz to someone who would put more effort into it.

    Considering the available evidence, AOL seems to have chosen the first option (they stepped up technical support during the last year) combined with handing more responsibility to volunteers where it makes sense to let volunteers and not staff manage affairs.
    ODP certainly hasn’t the technical, financial or PR support – or the press attention – of let’s say, Flickr or Wikipedia. But such advantages may bring in new problems which ODP doesn’t have, or not to the same degree… πŸ˜‰

    Regular information on ODPΒ΄s status is published monthly at

    ODP volunteer editor


  25. Hi Dave,

    I stumbled accross this following backlinks to my site ( – the one Tom Ligda mentioned) – first, if you haven’t already solved it, I think the slashes are probably due to “magic quotes” in php – note sure, but a likely hypothesis.

    Second, I would love to hear your thoughts on a directory in a wiki. I’m having pretty decent success, although I had to hard code a few filters to prevent spam and help people post objective descriptions. I think the big test will be when it’s large enough to attract browsers, not just webmasters looking for links. At 6,000 listings, it’s beginning to be usable πŸ™‚

    Warm Regards,


  26. I wonder how DMOZ picks their editors and sites???


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: