Friday, October 22, 2004

Instance editing with Giblet

Announcing Giblet Features include:
* Class selection and resource creation
* Appropriate property selection, based on the domain of the class and its super-classes
* Appropriate value input, whether as a string literal or an internal resource reference (based on the range of the property)
* Saving and loading of simple graphs to and from local storage, via cookies
* Export of graphs as RDF/XML, and a submission to the W3C's online image generator for RDF graphs
* Availability of templates (initially FOAFNet only) to help a user get started

Demo.

I also noticed RDF Templates which seems to have similarities with Kowari's Descriptors.

Thursday, October 21, 2004

Pychinko

"Pychinko is a Python implementation of the classic Rete algorithm (see Forgy82 for original report.) Rete (and its since improved variants) has shown to be, in many cases, the most efficient way to apply rules to a set of facts--the basic functionality of an expert system. Pychinko employs an optimized implemention of the algorithm to handle facts, expressed as triples, and process them using a set of N3 rules. We've tried to closely mimic the features available in CWM, as it is one of the most widely used rule engines in the RDF community. Several benchmarks have shown our Rete-based Pychinko to be upto 5x faster than the naive rule application used in CWM (see presentation below for preliminary results.) A typical use case for Pychinko might be applying the RDFS inference rules, available in N3, to a document. Similar rules are available for XSD and a dialect of OWL."

Gibson vs Stephenson

Who would win? "In a fight between you and William Gibson, who would win?

You don't have to settle for mere idle speculation. Let me tell you how it came out on the three occasions when we did fight.

The first time was a year or two after SNOW CRASH came out. I was doing a reading/signing at White Dwarf Books in Vancouver. Gibson stopped by to say hello and extended his hand as if to shake. But I remembered something Bruce Sterling had told me...During the regeneration process, telescoping Carbonite stilettos had been incorporated into Gibson's arms...Gibson and I dueled among blazing stacks of books for a while. Slowly I gained the upper hand, for, on defense, his Praying Mantis style was no match for my Flying Cloud technique. But I lost him behind a cloud of smoke."

Ethnoclassification

Metadata for the Masses "We’re beginning to see ethnoclassification in action on the social bookmarks site Del.icio.us, and the photo sharing site Flickr. Both services encourage users to apply their own freely listed tags to content — tags that others can then employ when looking for content. See a web page that looks interesting, but don’t have time to read it? Post it to Del.icio.us with a tag that will help you find it again."

"The primary benefit of free tagging is that we know the classification makes sense to users. It can also reveal terms that “experts” might have overlooked. “Cameraphone” and “moblog” are newborn words that are already among Flickr’s most popular; such adoption speed is unheard of in typical classifications. For a content creator who is uploading information into such a system, being able to freely list subjects, instead of choosing from a pre-approved “pick list,” makes tagging content much easier. This, in turn, makes it more likely that users will take time to classify their contributions."

Wednesday, October 20, 2004

Unit of Knowledge

CBD - Concise Bounded Description "A concise bounded description of a resource is a body of knowledge about that resource which does not include any explicit knowledge about any other resource which can be obtained separately from the same source.

Concise bounded descriptions of resources can be considered to be a form of representation, however they are a highly specialized form and not the most usual or obvious form in a web primarily intended for human consumption. They are, however, a key form of representation which semantic web agents need in order to reason about such resources and adjust their behavior accordingly."

Deploying Metadata

To Metadata or Not To Metadata "Given the rather pathetic record that many metadata efforts have racked up, it is little wonder that organizations have begun to question the entire value of adding metadata. However, there is another side to the story. First, the cost of adding metadata can be reduced in several ways. For example, the $200K for a metadata initiative performed by outside consultants can be greatly reduced by not starting from scratch in each case, but rather starting with existing metadata standards and controlled vocabularies and taxonomies. The cost of a unique custom job will always be higher than one that at least starts with predefined components.

In addition, the cost of doing metadata has to be weighed against the cost of not doing metadata. Assume for the moment that adding metadata would solve all of the problems associated with search. One estimate from IDC puts the cost of bad search at $6 million for a 1,000 person company. Now it is unlikely that adding metadata will solve all search problems, but even if it only solves half, that is still a savings of $3 million per year. In this context, $200,000 for metadata doesn't seem so exorbitant."

"You wouldn't think of running a company without organizing your employees, why do you think you can create access to information without organizing that information?"

Tuesday, October 19, 2004

Googling your way to freedom

Australian journalist tells of capture in Iraq "Initially they thought Martinkus was a CIA agent or a contractor working for the US. Then they entered his name into an internet search engine to check his story. Convinced he had no links to the United States, his captors let him go."

The BBC story confirms that it was Google.

tolog

One topic map query language, tolog. In comparison to SPARQL, behold the magical abilities of projection (no distinct required), count and sorting. Also, non-failing queries (optional), negation, limit and offset, and built-in inferencing predicates.

Their negation looks very similar to iTQL's exclude.

Monday, October 18, 2004

Date and Pascal on RDF

ON METADATA, RDF AND RELATIONAL REPRESENTATION and ON RELATIONAL BINARY DATABASE DESIGN.

From the later:
"Research papers abound from the Semantic, Topic Map and RDF camps that claim relational binary database design is the superset of n-ary or the true relational model.

The only reference to relational binary that I could find on your site was mentioned by C.J. Date in passing. He mentioned that Codd showed that n-ary and even 0-ary relations have unique and important properties that would in fact make n-ary the superset."

"The foregoing point notwithstanding, there are actually some pretty strong arguments in favor of making all base relations binary. Well, not binary exactly, but irreducible, rather; an n-ary relation is irreducible if it can’t be non-loss decomposed into two or more projections each of degree less than n. (In practice, irreducible relations often are binary, a fact that might account for part of the confusion I mentioned; however, some irreducible relations are not binary and some binary relations are not irreducible.) But the question of whether base relations should be irreducible is a database design question, not a relational model question."

From Good recent writings on data.

Another one, ON THE “SEMANTIC WEB”. I was going to say that in iTQL we don't return NULL but we do if the value is unconstrained. I was going to say that in SPARQL it doesn't return NULL but it does. I've been playing with REL as well, just to see how it performs certain queries too.

Breaking Classloader Hierarchy

Find a way out of the ClassLoader maze "Thread context classloaders were introduced in Java 2 Platform, Standard Edition (J2SE). Every Thread has a context classloader associated with it (unless it was created by native code). It is set via the Thread.setContextClassLoader() method. If you don't invoke this method following a Thread's construction, the thread will inherit its context classloader from its parent Thread. If you don't do anything at all in the entire application, all Threads will end up with the system classloader as their context classloader. It is important to understand that nowadays this is rarely the case since Web and Java 2 Platform, Enterprise Edition (J2EE) application servers utilize sophisticated classloader hierarchies for features like Java Naming and Directory Interface (JNDI), thread pooling, component hot redeployment, and so on."

The 3rd Age

The end of data? "There used to be a difference between data and metadata. Data was the suitcase and metadata was the name tag on it. Data was the folder and metadata was its label. Data was the contents of the book and metadata was the Dewey Decimal number on its spine. But, in the Third Age of Order (see the previous issue), everything is becoming metadata."

"So, in the Third Age of Order, all data is metadata. Contents are labels. Data is all surface and no insides...Why does this matter? It changes the primary job of information architects. It makes stores of information more useful to users. It enables research that otherwise would be difficult, thus making our culture smarter overall...But, now that everything is metadata, no particular way of understanding something is any more inherently valuable than any other; it all depends on what you're trying to do. The old framework of knowledge — and authority — are getting a pretty good shake."

Saturday, October 16, 2004

SWAP Highlights

From the Semantic Web Applications and Perspectives program:
* Peer-to-Peer Semantic Coordination which considers "...the problem of coordinating hierarchical classifications". A paper on the algorithm listed, CtxMatch, is "Improving CtxMatch by means of grammatical and ontological knowledge – in order to handle attributes".
* RDFGrowth and the Dbin Project. Another paper was published here. Dbin is "..a p2p semantic web algorithm designed so that each single peer is only allowed to cause a minimal, bounded, remote computational burden".
* Semantic Web Tool Evaluation "We have tested the following systems: Jena, Sesame, RdfSuite, Triple, RdfStore, 4Suite, RdfDB....RDF(S) is unable to express basic representation needs typical in simple ontologies...few tools are unable to correctly handle RDF(S) ontologies...no semantic web tool is able to correctly handle even OWL-Lite...the only tool able to correctly handle OWL-Lite ontologies is the description logic inference engine Racer." I'm not sure quite how correct these statements are, maybe with respect to the projects listed, I know we intend for Kowari to support RDFS and OWL-Lite.
* Querying the Semantic Web: a new approach Also talks about using CtxMatch, "...we introduce in the paper a first set of semantic parameters we consider relevant: (i) the type of relation, i.e., the semantic relation that holds between a concept of the source schemas and concepts in the target schemas (e.g., equivalence, greater or lesser generality); (ii) the ontological distance, namely the distance between a source and a target concept with respect to some reference ontology; and (iii) the lexical distance, that is the distance between the formulation of semantically related concepts in different schemas."
* A Framework for Unified Information Browsing Which discusses the RDFX Project.

Friday, October 15, 2004

Google is the platform

Search your own computer "Google Desktop Search is how our brains would work if we had photographic memories. It's a desktop search application that provides full text search over your email, computer files, chats, and the web pages you've viewed. By making your computer searchable, Google Desktop Search puts your information easily within your reach and frees you from having to manually organize your files, emails, and bookmarks."

Screenshots.

Google Desktop Search Launched "Indexing is fast and only happens when your computer is idle...Get a new email? Visit a new web page? All this information is automatically recorded and made searchable within seconds."

"Each time you view something, a snapshot of what you've seen is created. Did you visit the same web page several times in a month? A copy of the page each time you visited is made. The "1 cached" link will change to reflect the number of copies recorded."

While support is incomplete, this article mentions that other file types, "Images, Acrobat, Windows Media and MP3" are indexed. This probably suggests future features.

The most important missing feature seems to be, failure to capture metadata or allow you to modify it. What a wasted opportunity. There's no advertising but hopefully this can stay as a way to integrate people's desktops with the web site and not have to become its own money spinner.

Another article with screenshots, "Google Your Desktop".

AOL is going to do the same, curiously "...the desktop search tool is being tested alongside the AOL Browser but declined to elaborate further. She said the AOL Browser will launch as early as November."

Tuesday, October 12, 2004

Resolvers

Masala That's what Kowari's resolvers should do - it will be able to either provide a standard interface to live data (like LDAP queries, SQL queries, etc.) or to extract the metadata/data (from files, email, etc.) and index them in order to perform full text or normal RDF queries on it (through iTQL, RDQL, etc.).

Democratization of Radio

podcasting "Podcasting is a term used to describe the recent phenomenon of pointcasting audio programs over the Internet to people's iPod for later listening. We're all familiar with the iPod; it's become an icon of the times: hip, fun, and progressive. Most people use the iPod, and other MP3 players, to listen to music. But, combine an iPod with the Internet and innovative things start to happen."

"iPodder and similar applications have sprung up over the last few months to solve that problem. iPodder checks a list of RSS feeds on a scheduled basis and downloads any attachments right into iTunes. Anytime I hook my iPod up to my PowerBook, new MP3s are downloaded to my iPod. Very convenient."

Masala

IBM delivers Masala ""What IBM will be able to do [with Masala] is offer a federated data model that brings together a number of disparate sources in one place [to let users] search, index, and retrieve data without writing to individual data sources as you might have had to in the past," said Stephen O'Grady, senior analyst at Redmonk. "It is a fairly significant step up," he added."

"Masala has been at the heart of that effort to help "redefine" business intelligence. It is designed to allow corporate users to create a "virtual database" by collecting information seamlessly across the enterprise from things such as customer service records, e-mails, tables of numbers, photos, and other forms of information and view them all as if they were in one location.

With the new offering, IBM will also be able to compete against traditional enterprise information integration vendors such as Composite Software, MetaMatrix, and BEA's Liquid Data along with enterprise search vendors such as Verity, Endeca, and Autonomy."

Sunday, October 10, 2004

Bi-Directional Interoperability

Metatomix Delivers the mtx ERI PLATFORM for Enterprise Resource Interoperability "Metatomix(TM), Inc...announced the release of the industry's first bi-directional Java platform for integration, correlation, and policy-based automation...Based on the Semantic Web's Resource Description Framework (RDF) and a central ontology framework -- two technologies recommended by the W3C for commercial-grade data integration -- the mtx ERI Platform is the only interoperability platform that enables enterprises and government agencies to leverage node-based data integration, distributable across an entire network, for unprecedented data interoperability. The mtx ERI Platform combines scalability, reliability, and extensibility with the power of the Semantic Web to create complete resource interoperability."

Saturday, October 09, 2004

Jumping to Java 5

* Java 5 - "final" is not final anymore "It makes sense to allow updates to final fields. For example, we could relax the requirement to have fields non-final in JDO. If we read section 9.1.1 carefully, we see that we should only modify final fields as part of our construction process. The use case is where we deserialize an object, and then once we have constructed the object, we initialise the final fields, before passing it on. Once we have made the object available to another thread, we should not change final fields using reflection. The result would not be predictable."
* Annotations in Tiger, Part 1: Add metadata to Java code "Annotations are modifiers you can add to your code and apply to package declarations, type declarations, constructors, methods, fields, parameters, and variables. Tiger includes built-in annotations and also supports custom annotations you can write yourself."
* AbstractStringBuilder "AbstractStringBuilder and its subclasses demonstrate another new feature of Java 5.0: covariant returns. This is the ability to narrow the return type of a method when you override it in a subclass. It got added to the language along with generics." See also, Generics Tutorial.
* alt.lang.jre: Twice as Nice "Java 5.0 incorporates generic types as an effective means of addressing ClassCastExceptions, but this adaptation will do little for many corporations still using Java 1.3. Nice, on the other hand, was developed with ClassCastExceptions and NullPointerExceptions in mind. As such, the language supports two features, parametric classes and optional types, that go a long way toward keeping applications from throwing these exceptions. What's more, with Nice you can employ these features today, on any Java platform 1.2 or higher."
* Tiger's Corner - Displaying array contents java.util.Arrays has a toString or deepToString for multi-dimensional arrays which will print out the contents of an array.
* Why Mac Developers are Concerned About the J2SE 5.0 Wait Highlights the historic lag of porting Java to Apple OSes.

Mistakes Made by OO Databases

Whatever happened to OODBMSes, anyway? The 5 reasons given are:
1. They were seen as nonscalable and nonperformant.
2. It was "new" technology in a time when only the tried-and-true was being used. 3. It had no compelling features for anybody but the developer.
4. They just didn't perform or scale well at first.
5. OODBMS doesn't store the same way as the RDBMS. This may be the kicker--people building OODBMS systems as if they were RDBMS systems were bound to be hurt by worlds of disappointment, and vice versa (which is where we are today).

Friday, October 08, 2004

iCite

Why can’t I manage academic papers like MP3s? The evolution and intent of Metadata standards "Two key differences are identified: Firstly, digital music metadata is standardized and moves with the content file, while academic metadata is not and does not. Secondly digital music metadata lookup services are collaborative and automate the movement from a digital file to the appropriate metadata, while academic metadata services do not."

"Academic metadata is obtained in a variety of ways, yet even the most sophisticated methods stop short of the level of automation and convenience reached by digital music. For those not using citation management software the process is a manual one of re-entering citations in the required format. This is often done through “reference list raiding”, is subject to copying errors and is extremely tedious."

"Thankfully systems like the CDDB or the one proposed above have the characteristic that only one upload is required to make the metadata available to all. It is established that online availability increases citations to one’s work so it does not seem a stretch that easy availability of accurate metadata would also increase citations...We propose that two tasks are required to move towards more effective practices: using XMP to store academic metadata with the file and using content-based hashing to map between the digital files and a user contributed metadata database."

There's also the recent Wanted: Cheap Metadata: "What about new incentives for adding judgment-call metadata? Stephen Cayzer's work at HP Labs (see his XML Europe paper), which demonstrates how better user interfaces can make the entry of metadata less trouble for the user, will hopefully inspire others to think more about acquiring good metadata and postpone some of their ideas about what to do with that metadata."