Difference between revisions of "User:Datagrok"

Revision as of 10:24, 27 August 2007

User:Datagrok is the user page for Michael F. Lamb. My personal website is datagrok.org.

I have a page here because I'm casually interested in the UniversalWikiEditButton effort.

Here's the AboutUs meta-page for my website.

Thoughts on AboutUs.org

I'm interested in AboutUs because their community goals are noble, and there's some good discussion to be found within. Also, Ward Cunningham, inventor of the Wiki, is the CTO, which lends them some credibility.

At the same time, I disagree with some of their methods, especially so-called "domain pages" and their local cache of WHOIS database information. Others have vocally and sometimes bitterly criticized them as well. I'm amazed that they actually make money, but they've received at least $1 million in investment capital, and can afford to keep offices and pay employees! I hope this translates into more success with their community-building goals. It's also encouraging personally, as I've got a pile of ideas for web-based community-oriented software I'd like to try to turn into a business.

Some of the concerns I raise below have already been raised by others, and are discussed on AboutUs.org itself: Shifty Scavengers, AboutUs:Concerns, ConcernsFAQ, ConcernsPage, CommonConcerns, Thoughts On Vision. I think it's noble that AboutUs tries to be transparent, and have public discussions with their critics.

Technology

• TopSoil is an interesting idea, but seems very far away from this MediaWiki free-for-all.
• I think traditional wikis like MediaWiki are crap for threaded discussion.
  • However, wiki editing semantics should be great for discussions. (TheWikiWay)
    • TheWikiWay may be applied to other types of content than wiki pages. I think there should be more web-tools with wiki-style editing policies, tuned to support different tasks. (database, ontology-building, debate, discussion)
    • AboutUs in its current state seems to be trying to use MediaWiki to solve every problem involving community contribution. Granted, MediaWiki does have some tools for categorization and it might be twisted about to solve some of the issues mentioned above, but I fear it's not the right tool for some of these jobs.

Focus/Purpose

What is the focus of AboutUs?

• A wikified version of a category-browsable website catalog like dmoz.org?
• A wikified website annotation engine like shiftspace.org, twiddla.com, sharedcopy.com?
• A community wiki like infogami.com?
• A knowledge base with "more freedom" than wikipedia.org?

The aparrent answer seems to be "all of the above." I think many of these concepts, while useful to each other, should be managed within different domains, using tools specific to the task.

In each of those use cases, what does AboutUs bring to the community that differs from the existing sites?

AboutUs as a domain catalog

• I like the idea of applying wiki editing semantics to a website catalog. Seven years and numerous requests to dmoz.org for wording changes resulted in my listing being simply deleted. That's not a valuable service. I'm glad AboutUs allows me to change my own website description.
• I dislike the idea of mass-importing metadata for any junk domain with a bot. dmoz.org's value is directly related to the amount of time and consideration spent by the site's moderators to approve websites for addition and maintain the index by hand. I think that employing a bot to populate a wiki in this way goes against the "community" spirit of a wiki.
• While claiming to be a site for the community, AboutUs might be setting themselves up with "domain pages" as an easily-exploitable SEO spam tool. Quality, community-created wiki content will boost the pagerank for AboutUs, but then bot- or SEO-created outgoing links to junk websites may lend them unwarranted credibility.
• Compared to dmoz.org or del.icio.us, AboutUs seems to have less emphasis on categorization and ranking, and no cooperation with existing categorization and ranking tools. (del.icio.us, pagerank, etc.)
  • How might this work, if they could work together? TedErnst
    • (Placeholder: I want to reply to this)
• What value is there in meta-pages for websites? I prefer to describe my website on my website. The real value in a website directory comes from peer review, ranking, tagging, categorization, etc. MediaWiki has few of these features.

Scalability

• Will it scale?
  • There's 28,550 subdomains of infogami.com alone, and most of them are just spam. (according to a nonscientific google search)
  • How does AboutUs prevent spam in the domain listings? The answer thus far seems to be patrolled edits, in other words, a team of people to clean up by hand. I don't think this will scale.
    • It seems to scale at Wikipedia, though this is likely to be much, much bigger, so it's an open question. TedErnst
  • Why dedicate a whole Wiki page to every junk domain and subdomain out there?
    • Good point. Any ideas for criteria for inclusion? TedErnst
      • Community interest. I think dmoz.org has a good model with the exception that there's no way to make changes if the privledged folks doing the equivalent of "patrolled edits" there go away, get bored, or dislike you. If AboutUs is about community, and they really want to mass-import content, I say ask the community do it. The ideal website directory I'm imagining is rather similar to dmoz.org, with the addition of some policy based on TheWikiWay for permissive community management of an ontology/tag set and additions/updates. I think personal recommendations from real people would be far more valuable to a website catalog than bot-spidered wholesale importing of domains, even if it makes for a smaller catalog than what might be built with a robot. Going down this road though, what do you bring to the community that they don't already get from http://del.icio.us?
      • Furthermore, another point I'd make is that there are so very many sites out there, I think a whole page dedicated to each with no good (that I can see) way to collate them nicely is unhelpful. As a user, I'm not interested in meta-descriptions of websites whcn I can visit the websites themselves. If I want address information, I'll ask WHOIS because the information statically stored on some wiki is bound to be out of date. I'd be happier to see a list of what sites fit into a particular category, what sites are similar, and in particular, comments from people who have used and added those sites to the catalog.— Mike^talk 01:54, 12 June 2007 (PDT)
  • Why attempt to replicate the entire DNS namespace within a wiki structure?
    • When domains appear or die, the wiki must be kept in sync. I don't think this will scale.
• I dislike the fact that AboutUs links 'datagrok.org' to their own pages, not to my domain. This dilutes my pagerank, and makes Google send people looking for my site to the AboutUs meta-page instead.
• AboutUs seems to be a reaction to restrictive policies at Wikipedia. They wish to be a "more open" "everything" wiki.
  • How will they handle the issues with abuse that wikipedia avoids with a restrictive policy?
• How can one browse through the "community content" separately from the huge number of pages which are simply pointers to domains?
  • A. Using MediaWiki's categories feature may be the current answer. But I don't think it's up to the task, faced with this much information, and gobs of "website metapages" cluttering the "community wiki" namespace.
• I think it is rude to answer many people's criticisms of data-scraping errors with "you should fix it yourself."
  • Logging in isn't required. TedErnst
    • Ah, I was incorrect about that. I misinterpreted FixName, "The most common type of name fixing here at AboutUs is fixing capitalization... If you are logged in then you can move a page," and missed the context.
  • Are the errors with the AboutUsBot or with the underlying data, I wonder. TedErnst
    • There is FixName, the domain capitalization issue.
    • There are addresses which become outdated or get scraped incorrectly. Many people who maintain information in the WHOIS database do not consider that data to be authoritative, do not take care to keep it updated, or are not aware that it is public.
    • Many people wish for their domains to be completely excluded completely from this site. While I do not whatsoever disagree with the legality of hyperlinking and fair-use snippeting, if AboutUs espouses a pro-community stance, It seems rude and belligerent to place the onus of fixing a poor scrape on a user who does not wish to participate, and to favor a "stub page" (Example: FullHouseSale.com) over no page at all. AboutUs' argument seems to be that "other people in the community should be able to comment about a website." (CommonConcerns#Opt_Out_Concerns) True enough, but if there has been no interest within the community to do so, why keep the stub page around? Perhaps it's a technical obstacle or something, but I think it smells to most like mal-SEO trickery or Googlejacking. — Mike^talk 01:54, 12 June 2007 (PDT)
• I think that it is not constructive to provide a poorly-updated mirror of data scraped from dynamically-changing sources like WHOIS and others. Which users need this service? Which users cannot discover it already by using WHOIS tools? At most, bot-scraped pages should contain a link to perform a WHOIS lookup.