Daniel’s Blog

Web 2.0 was all about making the web more socially aware, however, was it really a good idea to start a versioning system for the Web?

When Web 2.0 was coined, I got quite excited about it, it was “Yes! this is a new way of thinking, this makes the web more personal”… but I did not thoroughly think through the versioning of the web. Applying 2.0 to the end of the term Web, meant that this is somehow new, somehow better and somehow a superset of what we had before… but it wasn’t, it just clumped pre-existing technologies together under one buzzword which businesses could use.

I fear that this is now happening with the term Web 3.0, which is seen as a swing towards the data. Unfortunately, the term Web 3.0 is being applied as a buzzword already (see the recent post by Paul Krill at InfoWorld titled “Salesforce touts Web 3.0 as platform as a service“, and the post by Simon Wardley titled “3 is the new 2…“).

I say that the Web is the Web, and it will always be the Web… whether it is a Document Web, a Social Web, a Data Web, a Platform Web or an Intelligent Agent Web. When it comes down to it, it will still be the Web, and documents will never be fully replaced by data objects, data objects will never be fully replaced by documents, the social apps won’t replace documents, and intelligent agents won’t replace search engines… things might change slightly, or become slightly more efficient… but essentially it is still the same.

Therefore, I don’t think it is clear when people start versioning the Web… which is why I have started categorising sub-webs, like I have done above. So these Sub-Webs are:

• Document Web: A Web of Documents with hyperlinks between. This Sub-Web provides a subjective view of information from the context perspective of the author.
• Social Web: A Web of Socially Aware Applications. This is done in either, or both Document and Linked Data forms. This Sub-Web provides a subjective view of information from the context perspective of the user or group profiled.
• Linked Data Web: A Web of Data Objects with relationship links between to enhance meaning. This Sub-Web provides an Objective View of information, ready for a contextual perspective of the user.
• Platform Web: A Web of Services in which you can run Web Applications on - and/or - against. This Sub-Web does not provide a view of information, but is an attempt to provide a distributed network of services capable of providing multiple subjective and objective perspectives.
• Intelligent Agent Web: A Web of Intelligent Software Agents. It is a Sub-Web because software agents will talk to each other in order to find out information. This Sub-Web communicates internally objectivity and subjectivity depending on the subjective desires of the user.

The important thing is that each of these Sub-Webs does not replace any other kind of Sub-Web. They can be used all together on the Web. No versioning required or desired.

However, this doesn’t mean that new technologies shouldn’t be looked into and implemented. They definitely should be looked into! I am just stating that I don’t believe versioning is the best way of labeling technologies, as it is pretty meaningless when something is not better or improved.

Any thoughts?…

The nice guys at True Knowledge Ltd (who are based in Cambridge, UK) have given me some beta invites to give out. So if you are interested in beta testing trueknowledge.com then please send me an email or comment in this blog and I’ll send you your invite.

True Knowledge Ltd is a company founded by AI sage William Tunstall-Pedoe, and their product (in beta stage at the time of writing) is a search engine which answers questions. It answers questions using natural language processing (NLP) and a highly-structured wikipedia-style question and answer system.

Unfortunately the True Knowledge system is not yet exposing Linked Data, but I am sure that they see a benefit in that area once the engine is released fully and publicly.

I haven’t got many invites, so be sure to let me know if you want one soon. It will be on a first come first serve basis.

Thanks,

Daniel

OK. It’s a funny title for a blog post, but it is an acronym which I don’t want to start up as a buzzword, but it where I see things going.

A lot of people know about Universal Plug and Play (UPnP), its a set of protocols and formats to allow you to just plug a hardware device into another hardware device and its guaranteed to “just work”. UPnP is all quite low level stuff, which is to be expected because we’re talking about hardware in this case.

The ease of UPnP is starting to appear on the web too, and what I shall (temporarily) call Web Universal Plug and Play (WUPnP). What it’s plugging together is various web applications components.

This started off in the form of open web services which could be “mashed-up” with a nice little GUI (e.g. Nestoria ).

We are now starting to see “mesh-ups”, which are fusions of data from different Linked Data sources to give context-sensitive results to queries.

Mesh-ups and Mash-ups are part of a bigger concept of WUPnP, which is not just about data or how to view it… but everything becomes a little more agnostic, and therefore creating a plug and play architecture.

To do this, we have:

At the Data Level:

• RDF (the framework, with its various socially-friendly vocabularies such as FOAF and SIOC)

At the Machine Interface Level:

• OpenID and OAuth
• OpenSocial
• SPARQL, SPARQL/Update and the SPARQL Protocol

At the User Interface Level:

• Rich Internet Application (RIA) Frameworks (e.g. Flash/Flex/AIR and Yahoo! Browser Plus)
• AJAX / AJAR

Network Level:

• HTTP (in particular REST)
• Cloud Computing (e.g. Amazon EC2)

And the key to this all working is the unique data source name:

The Uniform Resource Identifier (URI)

Without it nothing would be linkable or usable.

I’m not trying to standardise things, I’ll leave that up to the lovely people at the W3C. I’m not even recommending anything, thats up to the W3C and the DataPortability Group. I’m just saying that we have the technologies to be able to plug X into Y into Z, and hey presto we have an awesome app “voila web-scale plug and play”.

An example is a user interface designer could make a user interface in such a way, that it would be able to work with any component that a user may choose. Just for some details, this could be an Adobe Flex based user interface, which could then hook into any Linked Data based encyclopedia (e.g. DBpedia) and provide data in context.

You should be aware by now, from various previous posts that:

• OpenLink Virtuoso provides the Machine Interface Level, the Data Level and the Network Level
• OpenLink Data Spaces provides some User Interface Level functionality and some more functionality on the Machine Interface Level
• OpenLink AJAX ToolKit provides JavaScript, AJAX/AJAR and some funky bits of functionality and eye candy ready for User Interface Level development.

(Yes, there are a lot of acronyms in the computer science field, and particularly the web…. it’s because if we said the long names for everything then we really wouldn’t get anything done because conversations would be about 10 times longer! Unfortunately we have to deal with it)

(If I’ve forgotten anything useful in the WUPnP stack then let me know)

The Tribes of the Semantic Web

There are actually tribes within the Semantic Web:

These include:

• The Linked Data Purist
• The Semantic Annotator
• The Semantics Purist
• The Semantic Insider

Generally this is because they prefer a certain architecture over another, but I am not saying that each is exclusive. There are people in the community which straddle two or more of the tribes.

The Linked Data Purist

The Linked Data Purist will have links out to various open data objects, and will provide content negotiation in a Human-Friendly format (e.g. HTML) and in a Linked Data</a