The Burgeoning Openly Owned Web

Gotta write up on this for a peer-to-peer architecture document. Cool and interesting stuff
To digest…

• Hierarchy Versus Facets (tags)
• Folksonomies a year later
• Taxonomy and Folksonomy

Some nice links found today…

• Learn to live the 80/20 way
• Common REST mistakes

Was talking here in the lab and here’s more ideas:

• Service composition through the portal
• Integration of external non-DBE services using dynamic invocation ActiveBPEL handlers
• Automatic generation of atomic and composed services
• Demo services a DBE photo printing service using flickr
• Demo DBE property rental service using google maps

Nice read on ambiguities of SOA on Martin Fowler’s blog.

WIP… I need a user interface!

Problem:
How to present a rich user, expressive interface to a consumer in a user-friendly manner, which allows invocations of X services on a P2P network via a web interface.

Requirements:

• Run in any web browser and hence minimal dependency.
• Needs to be embedable within a web browser (the universal application!).
• That what is executed to present the UI needs to be verifiable e.g. via digital signatures.
• No remote code downloading once the UI has loaded. This presents security concerns and proxying issues.
• Standard based and compliant to that standard.
• Rich set of controls to express a business’ service offering
• A low developer learning curve and preferably an eclipse GUI builder, if not an XML schema.
• It’s got to look good - I don’t know about you but any application is sold within the first 2 minutes of it been used by a consumer. If it looks clunky, it’s not intuitive, not entertaining…. forget it!

Characteristics:
Reliable, Continuous, Portable, Versatile, Freedom, Internationalised (i18n), Pervasive.
Declarative & Imperative Offerings:

• XUL
• RCP/SWT
• XForms
• OpenLaszlo
• Java applets (Swing/AWT)
• Thinlets
• XHTML
• Flash
• XAML
• Rebol
• RIB
• And all those mentioned on the Open XUL alliance sourceforge project

A WIP…
Here’s some notes detailing the a service-oriented, folksonomic peer-to-peer architecture we have in mind. It will be:

• Folksonomy rather than ontology, collaborative categorization
• REST-style web based portal to access services, will allow interaction with service via web browser
• Lightweight Directory Service & Replicated Peers
• P2P Architecture, DHT, robustness, availability
• Service tagging of service implementation using annotations
• On-demand Distributed Service UIs, dynamism
• Service implementation language neutrality
• Service protocol agnostic
• Transparency, simplicity
• Oh… open source naturally!

Stay tuned for more on this; including high-level waffly diagrams that only highly paid system architects can appreciate… “Oooooo more boxes!”.

I’ve been looking at the required components to get this up and running (not too interested in writing the paper(s) yet). I estimate that we’ve about 50-60% implemented already.
Now naturally I gotta expand on all of these bullet points; docs, links and further text to follow.