Subscribe
RSS news feed
Add to Technorati FavoritesRecent Posts
- Cuil bursts onto the semantic search scene
- iPhone 3G, and Mozilla show user adoption to new levels of Power
- A Business Model or Architecture for Web 2.0
- Innovation Brief
- Socio-Technical Systems not IT Systems (‘The Dreyfus Moment’)
- Business Technology not Information Technology
- Exciting IT? Come to the Dark Side.
- ASAP is dead! Long live ASAP!
- New Web 2.0 services; Data Portability
- Innovation Brief
Navigate
Search the blog
« From Benchmark to Mashmark | Main | iPodification leads to Nothing »
Sandy
Many of us believe we ain’t seen nothing yet in terms of the usefulness of the Web. Or perhaps from the other perspective, many see massive political, business and societal risk in the evolution of the Web.
But with thanks to Peter Evans-Greenwood for sharing with me a fascinating link over the weekend, perhaps we are starting to see how this usefulness will manifest itself – and how the inherent risks can be managed.
The semantic Web, or Web 3.0, is so called because it represents the evolution whereby machines can read and interact with the mass of Web content and services, as well as people. It’s really worth reading that last sentence a couple of times because if one considers what this could really mean, it is very interesting indeed. Or put another way, think machines, not just people, working in the Web 2.0 way.
And perhaps Sandy sheds some light on what we’ll soon be seeing in our homes and offices.
To borrow words from this recent and extremely well written Economist.com arcticle ‘When Rael Dornfest e-mails his business partners about meetings and interesting titbits worth archiving, he copies the e-mail to his assistant, Sandy. Though Sandy cannot yet organise his evenings in foreign cities, Sandy can run his diary. Sandy also runs his address book and forwards reminders from his family to his mobile phone without being asked. Sandy, of course, is not a person. Sandy is a piece of software that Mr Dornfest and his colleagues are developing and whom, once Sandy is thoroughly tested, he hopes to sell to a wider world…’
And so what might this mean to the here and now of management strategy?
Intriguingly Sandy starts to blur the boundaries between people and IT – or in one organisational management sense – between human resource and IT asset management.
To some, Sandy would be described as a software agent and Rael would be described as an employee.
But in a sense both are in fact agents.
Both Sandy and Rael have Values (goals and constraints), act on Policy (mandates, laws and rules that govern how things get done), have a focus on Events (real world things that happen – often the stimulus for business activity), work with Content, and work in the context of Trust (the trust relationships between the participants in a system of value).
For sure Rael will undoubtedly be able to outsmart Sandy but Sandy can do things in seconds that would take Rael hours if not days – and probable do a much better job of it – and perhaps more importantly can do things with the universe of Web content and services Rael could never do.
So, today, we talk about IT in terms of applications and databases. Tomorrow, we might start to talk about IT more in terms of agents and information.
Sandy is of course an IT resource, and Rael is a ‘human resource’ – but both are ‘agents’ and I think understanding this will not only help us get really useful services from IT, but will also help us get along just fine. Oh, and a possible evolution for the HR function of the future? AR – the ‘agent resources’ function – I just hope Sandy gives me positive 360 degree feedback.
Posted by Carl Bate on September 5, 2007 | 
Digg this | 
Seed Newsvine
TrackBacks
TrackBack URL for this entry: http://www.capgemini.com/cgi-bin/blog/mt-tb.cgi/201
Comments
# on September 5, 2007 4:50 PM, sig said:
Andy, got a "Helloo, what's this?" when the article went "group objects and their features... into meaningful relationships and hierarchies"
Hierarchies? That's about two generations ago when it comes to organising methodology... being two dimensional (nifty for a paper based life), now gone three or more dimensional (tags for example, better when IT can crunch the data).
Triples have the inherent promise of going even further, smashing the object-atoms and packing relationships (aka knowledge) directly into the objects... even replace properties and other fun.
Kind-of-lost interest in the article after that glitch, but it may be only me not "getting it" :)
# on September 5, 2007 5:01 PM, sig said:
Carl! Many apologies!
Feeling pretty stupid not reading the signature...
Hope you're not too offended, Andy's a fine chap to be mistaken for is he not? :)
# on September 5, 2007 6:10 PM, andy mulholland said:
hi sig
flattery, it is said, will get you everywhere, however after working together for so many years i doubt that Carl will fall for it!
however back to the comment which i recognise as a good challenge. in discussing Web 3.0 the so called Semantic Web, or machine web, with Carl some time back i stated that if this meant a new approach to an understandable issue then i was in favour, but if it meant an opportunity for re processing old arguements then i was not in favour.
Carl then proceeded to convince me of the difference in the approach, i suggest that he needs to air this in more detail with you, may be off line through e mail.
regards andy
# on September 6, 2007 8:47 AM, sig said:
Andy, yeah, wasn't that a good stab at flattery?
Semantic Web - too much weight on the "web" part these days, easy to forget that it's really about "knowledge" and that I hear is quite an enterprisey are of interest ;)
Will see if Carl is up for some conversation in that space...
# on September 6, 2007 10:54 AM, Carl Bate said:
Guys, thanks for the comments - though not too sure whether I should be flattered, concerned or perhaps a little of both about the mix-up with Andy!
Couple of specific things I would like to add:
- relationships - and therefore hierarchies - are of for sure important to convey knowledge (and thanks to my colleague David Hunt last night for interesting email exchange on this topic!)
- to my mind the underlying trend here is 'externalisation' - i.e. in the ultimate sense being able to share your information so it can be used by anyone and any machine for any purpose; in effect, so any consumer and any agent can usefully use it!
So, I think some really interesting developments are:
- the 'externalisation' agenda can be served by both 'traditional' and 'semantic' technologies - the key is to ask the question about broadest possible information use first, rather than specific use and open it up; and bake this into management strategy agenda
- more small and large enterprises than is generally known are already using 'semantic' approaches and enabling standards (such as RDF, OWL and SPARQL) to meet business goals in more effective ways than more 'internalised' approaches to information sharing (e.g. applications and databases)
- for sure the semantic Web is only the next continuation of the journey; but the face that standards are already being adopted that can enable more of the mass of information shared on the Web to be processed in useful ways by machines I think is truly interesting
And of these three, the externalisation thinking is really helping at the coal face to extract business value and minimise risk from these developments. New possibilities for information strategy do exist. And we don't need every latest gizmo going to realise them :)
# on September 6, 2007 2:50 PM, sig said:
Carl,
beg to differ regarding "relationships - and therefore hierarchies".
If you've said "hierarchies - and therefore relationships" I would have no problems.
Hierarchies was a nifty way, probably the only one to express relationships when stuck in two dimensional mode with pen and paper. That's why Comte de Buffon's multiple relationships taxonomy lost out to Carl von Linnae - they had nothing but paper in those days.
But today, with IT? Nah, much easier to model any real relationship - you can add metadata to any data-object representing real objects. Tags, owner, place, links, most anything that all adds knowledge (as in how object relate to each other, Plato).
For that you would need a multilayer grid of different hierarchies - not very efficient...
It's a bit more to it as well - hierarchies requires a common semantic, rules and training:
Would I find a Marmotte in a taxonomy, a classic hierarchy/tree-structure? Nope, not being trained I would never look for it in the squirrel family (funny enough, that's where marmotta marmotta belongs) as such furries usually climb trees and that the Marmotte definitely do not look like it's about to do anytime!
Triples are even cooler still as the need to add extra data to the objects disappears - now it's baked in. In fact if you extend that you can even break up the objects and build them from Triples.
If all our information was in that form doing a query of "what friends of Carl has the same bike gearset as Sig's bike?" would be as easy as "where does Sig live"!
# on September 6, 2007 3:10 PM, Carl Bate said:
Hi Sig - the mild irony isn't lost on me that we appear to be discussing the meaning of hierarchies!
Btw, I love Triples too.
# on September 6, 2007 3:47 PM, sig said:
Hi Carl!
And I didn't even utter my usual "hierarchies are useless!" :-D
Think I have post brewing re non-web semantics... hehe
# on September 27, 2007 2:00 PM, sig said:
Carl,
did not have your mail so...
A bit more on the Semantic Web in practice here http://thingamy.typepad.com/sigs_blog/2007/09/no-more-tags.html
As you'll see between the lines I'll humbly suggest that the current information formats will not allow very successful use of the W3C ideas, but alas, there are ways... etc. :)