Four Pillars: Thinking about search

Doc Searls pointed me to a thought-provoking post by Tristan Louis called Where Virtual and Physical meet. Amongst other things, Tristan makes the point that as genres like SecondLife evolve with their own community standards, the question of how these standards operate in a court of law is moot.

The question per se is not new. Larry Lessig (in Code and Other Laws of Cyberspace) had this wonderful story about two neighbours, one growing beautiful flowers that killed all that touched them, the other growing pedigree canines that exhibited powerful and sensitive emotions. Petal from flower person floats into pedigree chum’s garden. Dog dies. Painfully. Lawsuit. Prosecution says “make the flowers not kill”. Defence says “but this mix of beauty and death is what sells, tell neighbour not to create dogs that feel so much pain”. Judge rules that flowers can kill only when paid for. Cyberlaw. All my paraphrase, I don’t have the book to hand and I read it maybe seven years ago. Apologies for any misinterpretation or misquoting.
What is new is that genres like SecondLife exist. What’s this got to do with search, you ask? Patience, patience.

Separate thread. Joel On Software gets me tangentially to a place where a community formed off Joel is discussing Microsoft and AJAX. Arguments about whether anything on AJAX can be found on MSDN or not. Turns out that MSDN is “localised” and search results relate to a specific location.

Separate thread. In the March issue of Release 1.0 (which I’m still reading) Esther Dyson makes the point that “now that it has found a business model, search is likely to evolve rapidly in two directions that empower the user to filter rather than find….”

Separate thread.  A recent conversation in the blogosphere where I work looks at the use of microformats such as hCard to try and resolve person/role conflicts between different applications when looked at from a workflow perspective.

The Tristan Louis post that took me on my random walk started that process by talking about the community rules in SecondLife and how they would stand up in court. And that made me think how easy it was for virtual rules to be come bricks in the walled gardens.

Somewhere in my head all these threads are the same. Once you move away from request-response 100%-accurate-lies models of database queries, you move into a new world.

Non-deterministic “probabilistic” search with relevance and ranking and heuristics-based feedback loops is this new world. Metadata and tags and microformats are just routes to improving how we improve. Improve search and find and archive and retrieve. And even improve mix and mash and cook and come up with something completely different.
And search means different things in this world.

  • To some it means resolve conflict, deduplicate, cleanse information.
  • To some it means improve information by empowering the creator/consumer.
  • To some it still means provide an answer I can use.
  • To all it means improve when successful

It is when the locus of the search is (a) consciously constrained by someone other than the searcher or (b) unconsciously constrained by someone other than the searcher, and this happens without the searcher’s knowledge or consent, that we hit a major problem.

So whether it is Google or Yahoo in China or MSDN re AJAX or community standards in SecondLife, we have to look at the garden wall and who’s doing the building.

When there are no walls, search can be refine or filter or cleanse or repair or fix or even good old find. Where there are walls of omission or commission, search is a corrupt and poor reflection of what it could be. 

Learning about blogging

I received a comment from 2036AD about my last post, and it was good constructive feedback. I guess I fell into my own trap, and rather than be Thinker or Linker, I wandered into Stinker territory.

The rules for creating value when linking appear simple:

  • 1. Tell people enough about what you are linking to, enough to make them choose to find out further or not, as the case may be.
  • 2. Tell people why you are doing the linking and pointing out, provide some opinion of your own. Is this good? Is this bad? Does this matter? Why?
  • 3. Make sure you explain any jargon or specialist terms within your post.
  • 4. The reader must have enough information (as a result of your post) to understand what is being linked to and to decide whether or not to acquire more information by following the link.

These four criteria make the difference between Linker and Stinker. With my EULALyzer post I went into Stinker territory. Thank you 2036AD for pulling me up.

And so as not to make it worse, here’s what I should have said. Someone has come up with a tool to analyse the terms and conditions in an End User Licence Agreement. This is interesting and useful. But it currently only works on Windows. Which makes it less interesting and useful.

I went on to add that the end-user agreements in the ISP, cableco and telco space are also proving challenging, and that having a tool to analyse them would be useful.

EULAlyzer

LifeHacker  pointed me towards  this while I was travelling. Interesting concept, once it works on multiple platforms.

Now if someone somewhere could come up with the equivalent for analysing the end-user agreements offered by every ISP, cableco or telco, in the run-up to WalledGardenia, we could all benefit.

“You can unleash the players and be confident that you will like the results”

That’s a quote from a wonderful interview of Pierre Omidyar in Release 1.0, March 2006.

In the article, he says that the challenge is to create the right environment where you can “unleash the players and be confident that you will like the results”. In his view, such an environment has three aspects:

  • open access to all
  • open connection (as in ability to communicate and to transact)
  • skin in the game (ownership and accountability)

What he says about markets and environments is as true for institutions. Every firm needs to know how to attract people, retain them, develop them and extend their potential. Omidyar’s rules rule. You should read the whole article.

I’d saved up the March issue to read en route to PC Forum, and to use to annotate the sessions. Unfortunately circumstances were such I could not go. So I got to read it on holiday.

It is probably the single most important print journal I subscribe to, in terms of things that challenge and educate me. Esther Dyson can be relied upon to bring together the people that matter on subjects that matter, and summarise things her unique way. Opensource. Blogs and wikis. Collaboration in general. Tags. Trust. Attention and intention. Identity. Consumerisation. In each case, things I was experimenting with became clearer in my mind as a result of reading the journal.

Sometimes people come up to me and claim that I have A-list-blogger syndrome, that I like having conversations with the bloggerati. And the way it’s said, it’s a bit like being accused of being studious and a teacher’s pet while at school. My reaction is the same as it was when I was at school. What matters to me is that I learn.

Stop, children, what’s that sound?

…. There’s something happening here.
What it is ain’t exactly clear.
There’s a man with a gun over there,
Telling me I got to beware.
I think it’s time we stop, children, what’s that sound?
Everybody look what’s going down

Stephen Stills, For What It’s Worth 

We live in interesting times.

A couple of nonfiction writers (Baigent and Leigh) sue a writer of fiction (Dan Brown), apparently about copyright. Weird that one of the antagonists in the fiction book is called Sir Leigh Teabing (an anagram of Baigent).

Apple Corp and Apple Inc go to court fighting about something to do with their names. And play a Coldplay song to prove some point or the other. I guess I should be glad that they didn’t ask Chris Martin what his daughter’s called, since it could be argued she has some connection with music.

And eBay. And RIM. And Chinese manufacturers suing US ones. And more and more patents being taken out everywhere, just at the point that even legal eagles seem to be agreeing that the system is wrong. Defensive patents. Frivolous patents. Expensive battles.
Everybody look what’s going down.