Interesting view on Blogger Effect on Google

Ray [Ozzie] writes "don't blogs actually increase the value of Sponsored Links?"  Unless I misinterpreted his argument, Google makes money by selling control over keywords, control which becomes more important as noise (i.e. blogs) gets louder.  As blogs get more popular, everyone will have to pay protection money to stay above the blogging flood.

He has a valid point.  But, while the value of Sponsored Links rises, the value of Google as a useful service will go down.  Google will have to find a sustainable balance between the two values.  I think the solution I outlined yesterday provides that balance because, while most bloggers will choose to implement it, the solution is not under Google's control and allow intentional abuse by, er, evil bloggers.  Evil bloggers?  Aiya!  Mr. Bush has some powerful memes.

BlogShares

Just in case you haven't noticed the BlogShares banner at the bottom, I claimed my blog at BlogShares.  While not as helpful as Technorati, its an interesting way to keep track of my blog's popularity.  Looking at the trade history, looks like there are BlogShares daytraders.  Derek Wilis seems to be a good trader playing long and accumulating as my blog moves up.

Neuralizing Blogger Effect on Google

Scoble writes about advertisers pressuring Google to devalue webloggers, a Must-Read according to Dave.  I agree.  My own post (in March something) about the effects of blogs on Google PageRanking algorithm is here.

While I am a blogger, I want Google to either provide separate search service for blogs (maybe combine it with News) or reduce the weights for blogs.  PageRank bias for ndividual item or date pages don't seem to be too bad.  It's the main page that cause most of the problem because it:

  • changes too fast – by the time you find it on Google, previously indexed content is gone.
  • heavy with links unrelated to content – blogrolls, etc.
  • covers wide range of topics

One simple solution is for blog tool developers and services to voluntarily 'mark' main blog pages by adding a META tag with NOINDEX attribute by default.  This solution:

  • removes the need for Google to identify blog pages
  • is not Google-specific
  • works with Google as it is now
  • requires bloggers to intentionally unbalance PageRanking

I wish there was a standard way to do the same at block level to hide blogrolls on item or date pages, but this solution will do until W3C provides us with a way to specify NOINDEX at block-level.

Skyscrapers in Blogland

Lets face it, Blogland is as flat as LA and sprawls like it too.  Hotspots, quiet neighborhoods, noisy industrial areas, linked together by freeways and byways.  If today's Blogland is LA, tommorrow's Blogland will look like NY with skyscapers reaching for the sky.  By a skyscraper, I mean a nested group of blogs forming a hierarchical structure.  From little two story buildings for startups to equivalents of Empire State building for mega-corporations.

I started with my observation that, while hierarchical structures are common in the real world, this is not the case in the Blogland.  The real world is populated by individuals, but groups tend to form a hierarchy.  To project these groups into Blogland, we need hierarchical blogs.  I chose skyscrapers as a metaphor to deliver the idea more vividly.

As a company gets larger, flow and storage of information becomes crucial.  Blogging technologies can help in this area, but typically chaotic network of blogs could do more harm than help.  A CEO is not likely to know about about, let alone subscribe to, a lowly QA engineer's blog.  Noises generated in a large network of employee blogs also needs to be managed.

Introducing hierarchy to a blog network will help us solve these problems.  Exactly how this can be done remains to be solved, but I am absolutely convinced that this is where blogging technology is heading.  If we can raise productivity with blogging technologies, companies will pay.  I believe introduction of hierarchy is the first step.

Update: My zealot-style of writing can sometime work against me, so I thought I should clarify that I am not proposing absolute hierarchy for blogs, just adding hierarchy to what is already there to reflect real world structures.  This helps orient bloggers in the group and eases understanding of built-in information flow (up, down, sideway, etc.).

GraphViz and Technorati

While my son is watching Saturday morning flood of cartoon, I am playing with GraphViz, a neat tool for visualizing graphs.  Visio is nice, but it can be tedious and writing out complex structures in dot is much easier than the usual Visio routine.

Given its popularity, its lack of ready-to-use Python binding is weird (please don't mention Perl, I am allergic to it).  It should also support Flash and Visio output.  Still, a very cool tool and a joy to use.  I could sit here for hours generating pretty graphs.  Wouldn't it be really cool to hookup Technorati with GraphViz?  Wooo.  RDF and GraphViz also go really well together.

More Comments on Misgivings about Social Software

Interesting string of comments to Ming the Mechanic's post referring to my Misgivings about Social Software post [via Ray Ozzie's post about superconductive relationships].  I particularly liked this comment made by Ming.

"Maybe it is exactly what we need. That we're able to group together into drastically different realities. As opposed to us all arriving at some kind of uniform agreement about everything. I am pretty sure that we need lively, creative diversity, as opposed to bland uniformity.

But the fears that it brings up is if it enables people to walk around together in what might be perceived as more negative or dysfunctional realities. Such as violent, hate-based, mis-informed communities who start taking action based on their shared beliefs.

So, maybe better social software would allow the ku-klux klan to organize better, and to feed itself with self-supporting information. Maybe it will enable gangs of delusional teens who think the world is a dungeon where you shoot Nazi's with bigger and bigger guns. But I think I'll lean towards believing that the larger effects of better communication will more than balance out the potentially dangerous aspects.

Powerful social software might allow a small group of people to work closely together on some nefarious plot. But it might also allow millions of other people to work together on rendering such a plot useless. "

BTW, Blogland sure is a bewildering place, but services like Technorati helps quite a bit.  Using Technorati, I found people discussing my post around the globe .  Too bad, I can only read in English and Korean.

Emergent Markup Languages

Believe it or not, we live in a technologically backward civilization.  Everyday, hundreds of millions of people turn on their computer and enter a long sequence of characters into it, sequence which computers store, networks transport, and search engines index without understanding what they mean.

Markup languages such as XML can add meaning to those sequences of characters so computers can process them more intelligently, like differentiating prices from order numbers or Don Park from Donner Park.  But XML has, so far, failed to deliver on its promises.

Two great pitfalls of XML are:

  1. need for centralized control over creation and evolution of schemas
  2. high cost of developement, standardization, and education in time and resources

For some piece of data to be marked up, some one has to first define, document, publish, and publicize the schema.  After that the fun part starts: dealing with competing schemas and standardization process.  After standardization, millions of users have to learn how to use the new schema.  Whole lot of work and wait, for all parties involved, just to see a glimpse of XML Heaven.

What if we can skip all that?  What if people markuped content using their own names and structures, not those dictated by the central committee?  Will the resulting chaos be unsurmountable?  I believe not.  I believe that constraints and mechanisms inherent within human languages and social structures lead to what I call Emergent Markup Languages, common tags and structures that emerge from natural behaviors of individuals following simple rules like "call it what it is" and interacting with their immediate surroundings and neighbors.

Initially, Emergent Markup Languages will be most useful in marking-up salient parts of free-form textual content: phone numbers, addresses, numbers, names, etc.  Doing so will lead to abundance of fine-grained data we can harvest and process.  Its not quite Semantic Web yet, but we'll be much further along than before.