Saturday, March 18, 2006

Doom, something we can all agree on

We expect this sort of thing from environmentalists and leftists, so we don't really take it all that seriously. But when prominent Republicans write doom-slinging books about immanent theocracy and economic collapse, I'd say it's time to kick up worry level a tad.

Friday, March 17, 2006

Math is hard!

I was complaining before about semantic web stuff being heavyweight? Here's a section from the SPARQL (RDF query language) specification:

2.5 Basic Graph Patterns

A basic graph patterns is a set of triple patterns and forms the basis of SPARQL query matching. Matching a basic graph pattern is defined in terms of generic entailment to allow for future extension of the language.

Definition: Basic Graph Pattern

A Basic Graph Pattern is a set of Triple Patterns.

Definition: E-entailment Regime

An E-entailment regime is a binary relation between subsets of RDF graphs.

A graph in the range of an E-entailment is called well-formed for the E-entailment.

This specification covers only simple entailment [RDF-MT] as E-entailment. Examples of other E-entailment regimes are RDF entailment [RDF-MT], RDFS entailment [RDF-MT], OWL entailment [OWL-Semantics].

Definition: Basic Graph Pattern equivalence

Two basic graph patterns are equivalent if there is a bijection M between the terms of the triple patterns that maps blank nodes to blank nodes and maps variables, literals and IRIs to themselves, such that a triple ( s, p, o ) is in the first pattern if and only if the triple ( M(s), M(p) M(o) ) is in the second.

This definition extends that for RDF graph-equivalence to basic graph patterns by preserving variables names across equivalent graphs

Now, this is not all that abstruse, but then I have a math degree and am used to it. Way back then I studied mathematical logic, thought it was fun but a lousy way to describe the world and a lousy model for computation. It seems to have triumphed, however, and now people are expected to speak that language (the RDF semantics spec is even worse). I have a hard time believing that working programmers (a set of which I myself am a member of, nowadays) are going to be gleefully soaking up all the model-theoretic semantic theory behind the semantic web.

Maybe I'm wrong, there are people willing to boil things down into understandability. This well-written article doesn't cover any sort of non-trivial queries however.

There seems to be a big gap between RDF-as-first-order-predicate-logic-with-model-theoretic-semantics and RDF-as-useful-flexible-emerging-data-standard. Looked at positively, it's a two-pronged thrust, from academia and hackerdom acting together.

Wednesday, March 15, 2006

Boring is worse than wrong

The real reason I find myself vaguely defending some religious ideas against village-atheism is the same reason I used to go around getting into arguments with libertarians. It's that overly simple belief systems really annoy me, even if (especially if) I mostly agree with them. Libertarians used to make me crazy, because I always knew exactly what they would say in any situation (the spur for this post is this column (TNR subscribtion required) which makes the same observation about the New York Times op-ed libertarian, John Tierney). Same for the atheists, like PZ Myers. I can pretty much predict his reaction to any piece of news involving religion, which makes his posts on those subjects information-free (they can be entertainingly nasty though).

I guess I don't like fundamentalisms of any sort. Libertarianism and atheism both tend to be all-explanatory and ignorant of nuance. Libertarianism in its fascination with the distributed market model ignores anything which doesn't fit its framework, such as public goods, distortion of the market by the overly powerful (who can manipulate it unfairly) and the poor (who may have no motivation to respect property rights). Atheism tends to ignore the nature of religious belief, treating as something like bad science, rather than trying to understand what it might really be about. Let's just say God is the referent of the word "God", and while it may not exist the way a chair exists, it has a real conceptual role in thought which it might be useful to understand.

Tuesday, March 14, 2006

Semantic Con

I don't really mean the pun in the title; the Semantic Technologies conference was actually pretty interesting, in a hypeish sort of way. It represnts the attempt to diffuse a set of academic technologies and standards into industry. Note that while the standards (RDF, OWL, etc) are promoted and developed under the rubric "Semantic Web", the web is notably absent from the conference title. That's because these technologies are not being taken up so much in by the web proper, but are being sold to huge organizations with massive and complex data integration needs. These include defense (aero), intelligence, finance, and biotech. At the grassroots, it is being pushed by working life scientists trying to solve the same sorts of problems. Here's a presentation by Carole Goble that summarizes the hype pretty nicely, and without much technical detail.

The Semantic Web standards are complex, verbose, and hard to understand. They are backed by Web programmers prefer simpler standards: REST over SOAP, folksonomies over ontologies. Web guys do mashups and go to ETech; semantic technologists propose complex architectures and require large funders to get them realized.

It surprises me but it looks like the pressure of hard problems with good funding will overcome the complexity barrier of the Semantic Web. It surprises me not even so much because the syntaxes are complex and verbose, but the results are fairly inexpressive and inflexible in certain ways (for instance, OWL, the ontology standard, is based on description logic which makes it hard to do something as simple as default reasoning, something any simple old-fashioned frame system could do). Having been out of industrial grade AI for awhile, I will reserve judgement.

Sunday, March 05, 2006

Semantics in a San Jose Hotel

I'm not much of a conference-goer, but this week I'll be at Semantic Technologies. This amusing site of recommended meetings lists it as a "Chess Club Conference", ie, geeky, which is only partly right -- it's actually part geeky and part marketing, where architecture astronautics meets the hype machine. I should get my cyncial slogan printed up as a t-shirt.

There is actually some pretty interesting stuff going in this space, especially in life science, and an awful lot of coporate hype. I will be attempting to separate out the two.