notes on Web Evolution

Javascript: an introduction

ajbadd — Thu, 06 Jan 2011 07:21:30 +0000

I recently stumbled upon a presentation by Douglas Crockford regarding Javascript. This talk was a good introduction to the language. There are many misconceptions about the language and the biggest contribution of this session is that it sheds some light on the history and nature of Javascript. I have borrowed heavily from this presentation.

Javascript is a misnomer as it is not a subset of Java, nor is it an incomplete “web toy”. It is completely independent of Java and is a complete programming language. While its is a very popular language its is not without its blemishes. It has design errors, some of which have been rectified while others have not been. Some of the negativity associated with Javascript is due to the fact that many early implementations were less than stellar and the DOM environment inherently is not an easy environment to program in. he confirms that good resources on Javascript are hard to come by and the O’rielly book Javascript: The definitive guide, 5th ed was probably the best book on the subject.

Next was a mini history lesson about the origins of Javascript. Here Crockford sheds some light on some confusing nomenclature and history of the language. What was supposed to be called C++++ was called livescript and then finally Javascript. The naming was a by product of Sun and Netscape collaborating to counter Microsoft. Jscript was created by Microsoft by reverse engineering javascript.

Some salient features of the language:

1) load and go delivery. Code is delivered as text not compiled or classes, designed to be embedded in web pages.

2) loose typing – more expressive, any variable can receive any type of object

3) objects as general containers

4) prototypal inheritance – different from classical inheritance, obis derive from other objects not classes.

5) Lambda – functions as first class objects

6) Linkage through global namespaces

The session ended with some information about the datatypes and operators.

———————————————————–

Example Javascript program

Filed under: Uncategorized

Show me the data!!

ajbadd — Sat, 27 Jun 2009 00:07:41 +0000

The core of the “Semantic Web” phenomenon is rich, machine readable data. In my last post I had mentioned that some data on the web is annotated. Viewed by itself, this body of data is vast. A large part of Wikipedia is available in an annotated format at DBpedia. Freebase is another huge, user driven data repository. There is a plethora of smaller sites like bioontology.org. This still is only a fraction of the data available on the Internet. A large amount of content, and hence knowledge, is published every day on the Internet and a very small proportion of this is produced in the annotated format today. An increasing number of people and companies are are publishing their data in an annotated format.

How can you use this data? You would have utilize these data sources in semantically aware applications to fully realize the potential of such data. There are also several semantic search engines that look for annotated data. Some of these are – swoogle, sindice, falcons and watson. You can type in a search term at these sites to get results of annotated data, but to fully leverage these sources you would have to access them programatically and leverage computing power to draw inferences from vast amounts of interlinked data.

The vision of the “Semantic Web” has a tenet that any object or concept is recognized by its own URI (uniform resource identifier). An example of a URI is:

http://dbpedia.org/resource/Cricket

This is the identifier for the English game of cricket. It cannot represent you, Mike Tyson, Christmas or even the insect cricket. It may or may not be an actual link that opens up in a web page. (When a URI opens up in a browser it is dereferenceable). As annotated data proliferates on the Internet there is bound to be a lot of duplication and and dispute involving URIs. Whose definition of cricket is valid ? which one should I use? This is a stumbling block for the semantic applications. There are several initiatives that address this issue. Sameas.org is one such site. It lists possible equivalent URIs for a given URI. It would be great if sameas exposed an API but does not look like they do as yet.

Another initiative is Faviki. Faviki is a bookmarking tool that allows you to tag pages that you like. Now there are many services that do this, what’s great about Faviki is that you can tag pages with concepts and objects that already have URIs in DBpedia (or other sources). This is a great alternative to using arbitrary text tags whose functionality is limited. Adding little text tags to content is a whole another initiative that would better be discussed under Microformats (more about that later).

Posted in linked data, semantic annotation, semantic web Tagged: Database, DBpedia, freebase, linked data, semanticweb, Site Management, Wikipedia, World Wide Web

The “Semantic Web”: a low-fat, no-nonsense introduction

ajbadd — Tue, 09 Jun 2009 23:41:47 +0000

So what is all the buzz about ? I’ll get right to the point -

The “Semantic Web” is a broad term that denotes the annotation of data on the Internet and the operations these annotations will allow. The annotations better describe an object and relationships between objects.

1) Idealistic, pie in the sky take

All data on the Internet will be annotated. Computer programs will be able to trawl the net, drawing inferences and producing new knowledge.

Why is that a big deal ? Consider the case where you have a web page about Sam; it says Sam has a son Max and Max has a daughter called Zoe. All computers know to do is display this information, so you can read it. A traditional search engine will not display any information from this page if you queried ” Sam’s grand-daughter”. Now consider the case when the same page is semantically enabled. In addition to displaying information, computers know more about the content and can make inferences from the page. Like both Sam and Max have children and Sam has a grand-child.

2) Pragmatic, “working today” situation

Some data on the net is annotated. People are working on annotating more data. There is a good possibility of exploiting this annotation in the Web as it exists today. These benefits are not necessarily from technically spectacular acts of drawing inferences.

I try not to use the phrase “Semantic Web”, capitalization and all..such usage can get you associated with the idealistic, flag waving bunch (who may be proved right in years to come). I personally prefer the middle path – there are benefits today and there is a possibility of a lot more, but there exist sizable stumbling blocks (more about these later).

_____________________________________________________________

How is data annotated ?

One popular method of annotating data in the Internet is to use RDFs. RDFs is an XML based standard that allows you to describe data.

What is an ontology?

In this context an ontology is the representation of knowledge in a domain. OWL is a popular standard for depict ontologies. Ontologies provide a vocabulary for annotation of data on the web. In the example above, computers could use, say a human relationships ontology, to infer that Sam has a grand-daughter.

This is an attempt at a very basic introduction to the subject. There are several great resources on-line for those who wish to dig deeper.

Posted in semantic annotation, semantic web Tagged: Knowledge Representation, ontology, owl, rdf, semantic web, Semantics, semanticweb, web 3.0