Monday Mar 31, 2008

A word from our sponsor

Ms Viviane Reding, the sponsoring commissioner spoke to the conference via a video cast. This isn't yet available, but I'll post a link when I can, hopefully to streaming video.


What's the EU doing?

The conference is sub-titled, or has the tag line of "Perspectives emerging from R&D in Europe". A new web site has been created which promotes and reports on the conference, together with other initiatives. A document "The Future of the Internet, A Compendium of European Projects on ICT Research......" was distributed to delegates documenting the EU funded projects which were invited to attend.

I can't find it on cordis, and I used the catalog number, KK-30-08-142-EN-C

tags: "

Future of the Internet, the 2nd information revolution

The conference, "The Future of the Internet" was opened by Dr. Ziga Turk. He is the Minister for Growth of Slovenia, which holds the rotating EU presidency at the moment. He opened by talking about Slovenia's adoption of the internet, which was prior to independence and stated that the internet was an important tool for the campaigners in pre-independence Slovenia. After my experiences in trying to get connected in Italy, I have been pleasantly surprised. Easy connection for both phone and laptop.

He then, cleverly (well, I thought so), compared the development of the internet and its opportunities with the discovery of cheap paper and the renaissance. I was particularly interested in his assertion that while the invention of paper came from China, it was the european's letter based writing that enabled the first knowledge based revolution since printing was easier. He also pointed out that the first global knowledge revolution, the "p-revolution" while global, was led in Europe, but today's revolution, the "i-revolution" is not. The european response to this needs to be two fold. The simplest is to continue with EU enlargement, the other political responses are within the EU's "Lisbon Strategy". This is aimed at creating and stimulating jobs and growth in europe, and places innovation and research at the heart of this effort. It is also about dynamism and entrepreneurialism in the context of caring for people and environment. The driving economies of the US and Japan are being challenged by China and India, but by placing knowledge as a 5th freedom, the EU can hopefully harness the creativity and entrepreneurialism of its citizenry.

The first four freedoms in the EU are a bit different from Roosevelt's declared at the time of the founding of the United Nations.

We've been promised the slides, but they're [still] not available yet.

This article was written from notes taken at the time, posted the following week and back dated to the approximate time the speech was given.


Wednesday Dec 12, 2007

So what is wrong with the Internet?

On Day 2, of the NESSI AGM, we broke into seminar groups. The first session I attended was called the ' Future of the Internet', it was led by Mike Fisher of BT [Google him], who presented about the forces for change on the internet, both historic constraints and changes being brought about by technology innovation, and demand. Again a key view of the future is the the internet evolves from a network of computers, beyond a network of things to a network of services. Since Mike comes from a network company, and a large one at that, and so understands how poorly IT is ready to manage the challenge of scale raised by these factors.

In the afternoon, I attended the 'Service Orientated Infrastructure' session. Some aspects of the problem domain are very broad and interesting, but the discussions seemed focused around today's grid solutions in academia and commerce, although I arrived late. This working group's documents are also available on the NESSI web site SOI work group page, and their own web site. The GRID Strategic Research Agenda is available from the NESSI Site [.pdf].


Tuesday Dec 11, 2007

NESSI's Research & Projects

There then followed a series of presentations about the current approach to research and most interestingly presentations from the leading strategic projects. These can be found on the NESSI's AGM page on their web site.

This was followed by cocktails. Very nice!


Coming soon...servicenet

A EU funded attempt to create an academic Network of Excellence, led by University of Dusberg-Essen, called S\*Cube. There are 16 partner universities who will all participate equally, the UK partner is City University, London.

Dr. (Klaus) Pohl, predicted that the nature of the Internet was going to change so radically, that its name should be changed. His vision is of a network of services. Will the transition from a network of computers to a network of things require new network paradigms and protocols? Will it challenge the Atomic locking single write-ahead log database?

Dr. Pohl exposed a research framework, that analysed Service Technology as the existence and interfaces between business processes, Service Components and Service engineering. These need to be created which requires engineering knowledge and science and monitoring and adapting, which are classified as Service Engineering. It is felt that interfaces between these domains can also be developed and the S-Cube research is looking at developing knowledge from current intellectual property around BPM, grid, systems engineering and service management.

S-Cube have a web site coming on line with a pre-registration feature.


The ambition of Open Systems

Prof. Carlo Ghezzi, of the Politecnico di Milano, presented on Academic/Corporate collaboration and among other things examined the drivers of macro-change in open world. He argued that inter-operability is not enough, and that both a series of what he called self-\* qualities are required such as self-healing, self-configuration etc. He also again identified the self-advertisement as a new problem to allow services to be discovered and used.

I wonder if these ambitions are contrary to the classic inspection and vote that takes place in today's clusters.


Software, economics and society

We have just heard from Dr. Frans De Bruïne and Ken Ducatel, the former talked about the need for Security, to guard against global warming and the demographic time bomb; Europe is not just interested in health care because of the socialists. The ageing population is a jeopardy to the wealth engine of work and the various governments and commission all have different responses (Oh Boy!).

He stated that for instance in Holland, they're playing with a government 'Facebook' page. Will this lead to you having to document your car insurance, child support liabilities and private pension provision on-line in a government portal. The latter might help keep track of what the insurance companies owe you, but do you really want this hackable, or publishable at the will of politicians and civil servants? Despite these fears it is a possible first step to a real EU Web 2.0 and user created content, I am not sure what value one can create through communities in such a portal. It would also need some serious investment in reaching all the EU's citizens, both in the network and server infrastructure to reach everyone, but also in client access ubiquity. Not everyone has access to a computer, although most have phones and the ipod touch with its wi-fi is an interesting and probably popular innovation of the internet hand held device. Wi-fi is neither as ubiquitous, nor as cheap as in the US yet, and I suspect it varies massively with the EU.

He also spoke about how in Germany, networked medical care systems in the home allowed patients to be discharged earlier, thus saving money. Presumably the IT reduces the number of relapses.

Ducatel argued that the US uses its its (minimal, except for defense) public requirement to seed ICT innovation. I wonder if this is because US business has a greater appetite to build its own code. The flip side of this is that "Europe under uses Software". Its an opportunity for growth, and an opportunity for supply, but commercial stove piping inhibits the growth opportunity.


Driving change on the internet

The first key note was from Dr. Joao Schwarz Da Silva, a Director from the Commission's ICT. He envisioned a network of services driven by trends easily observable today. These are,

  • Social Networks
  • Digital Production
  • Virtual Worlds
  • Internet of Things

Much of the consideration around social networks seems around how to monetise the size of the network. The value created by cooperation seems always to be under valued. Dr. Da Silva predicted that the growth of social networks and user created content would lead to the growth of what he calls Digital Production. At its most simple, this will be just allowing mashups on a home page, however more complex models such as the tools for machinima or audio manipulation are clearly here today, it'll be interesting to see where this goes.

I am more questioning that virtual worlds will become ubiquitous and powerful problem solving tools. It is clear that World of Warcraft is a hugely popular both social network and digital world, but we have spent 1000's of years devising two dimensional representations of most of the problems we seek to solve. We need new representational metaphors before 3D rendering and virtual worlds become serious problem solving devices. I mentioned this earlier in the year. These criticisms are before considering that a Social Network needs to leverage the wisdom of crowds, or at least the wisdom of huddles. Facebook's visual {book/DVD} shelf works because you can see what both your close friends and strangers say about the books and films you're interested in. You can see what everyone, or at least your friends recommend. An interesting counterpoint though is that if you consider electronic gaming to be a social network, then sharding reduces the wisdom of crowds; you can only learn  from the wisdom of a shard. There's lots of work to do before 3D and/or Virtual Worlds truly take off.

He then looked at how in a network of services, one discovers anything useful. So this is partly how does one discover any content, such as images (tags), houses (attributes) etc., but for services we expect a directory solution. There isn't yet a directory of internet services.


Thursday Feb 01, 2007

semweb & FOAF

So one thing I have done inspired by last week's workshop has been to set up a personal FOAF file.

I used the foaf-a-matic to create it, and it can be browsed using the tabulator.


Reading my FOAF file using the tabulator


I shall probably use Henry Story's foaf file (who wrote about his contribution to the workshop here...) as a template some time soon; it has a richer information syntax. In fact Henry's write up is pretty good and offers a number of key links to progress one's understanding.I put him on my blog roll a couple of weeks ago.


Wednesday Jan 24, 2007

The shape of the internet, inside and outside the corporate firewall

I have been discussing the efficacy of our internal search tools and how hard it is to find stuff, and to be honest, assumed that it was the crapness that most users accuse their IT colleagues of. However a colleague, Bernard Horan recommended that I read "Searching the Workplace Web", which suggests a different answer.

Searching the Workplace Web argues that intranet's are different from the internet and that more flexible, and different search algorithms are required to search an intranet; the most successful internet search algorithms are not necessarily going to work well on an intranet.

The author's made four observations.

The first assumption is that content in the intranet is often created for the purposes of dissemination of an authorised opinion or fact, or a statement of policy. There is no design intent to attract readership. One observation on which this is based is the fact that often content in the intranet is very light of additional hyperlinks to suggest further reading or to quote sources. Why suggest further reading when your authoring policy? There is no further reading to be done. Why quote a source when the answer is "because my boss said so!" For example, check out your companies expenses or travel policy. The authors argue that a corollary of this is that "in list" based algorithms such as PageRank may be less effective on intranet searches. Interestingly, I ran this past Chris Gerhard, who said that he'd been looking for the original text of the Road Traffic Act, but that Google (an in-list based search engine) had difficulty finding it as it preferred commentaries on the law, because they were more referenced by web page authors.

These two examples take us to assumption two. In the intranet, we're often not looking for the "Wisdom of Crowds"; there are often very small result sets for a given query, often the correct result set is only one entry. This will occur when you are looking for a policy, or an officer's represented opinion. Will expenses pay this journey cost? Is this a supported configuration? It occurs when the researcher is looking for authority not opinion.

Observation three, is that there is (likely) to be less spam inside the firewall. (I wonder if this is an aging observation, with the growth of blogs and the opening of mail archives to search, it may be that this is weakening in strength, but its unlikely (but not unheard of) that large porn collections will be found be accident in an intranet search). The corollary to this observation is that some ranking algorithms that are unsafe on the Internet, become useful inside the firewall.

Observation four is that Intranets are less friendly to search. The authors observe that much content is held inside databases, or document servers, portals, directories and other specialised interfaces

While reading Benkler's "Wealth of Networks", I first came across the concept of, a shape of the internet. Obviously we all know that some sites are very influential and highly read, but the internet's hyperlinks have a topography that can be described and measured using graph theory. This was as far as I can tell first explored by Broder, Kumar, Maghoul and others in their paper "Graph Structure in the Web". These topographies were discovered during the ascent of the dynamic search engine, which won out to the detriment of the directory based references. These two papers are contemporaries and it'd be interesting to see if these topographies remain useful as insight today.

IBM discovered their intranet topology was different to the Internet, with a smaller "core" and a larger periphery. The core is a bunch of sites that meet formal graph theory definition as strongly connected. (See Graph Structure in the Web). The size of the OUT segment, pages that can be reached from the core, but do not return is larger than in the internet, and is much exacerbated by domino document repositories. The periphery is also much larger than in the internet, they can be found from the crawl seed pages (which must be in the IN segment) but not from the core. They measured the frequency distribution of the probability that an in-list based sort algorithm would place on a page's relevance and discovered a difference in shape between the intranet and internet results, with a lower proportion of high scoring pages in the intranet.

Another interesting innovation was that the research team created three indices (most solutions used only one) for determining relevance, these were content, title and anchor text. (Anchor text is the text between the anchor tags, and thus chosen by the author to represent the link in the original document). They then build a flexible ranking engine that had a number of input parameters. (I might write about this another day, but if you want more now go to the original document).

It's three years later and its almost certain that with the changes in user content authoring tools and the fact that there is more spam and more opinion, that the topology will have changed. The improved content creation tools represented by blogs and wikis also weakens the assumption that intranet content has low link counts. Sun is very permissive about blogs, as are as far as I can tell IBM, but the introduction of blog and wiki technology has both strengthened and weakened the firewall and hence the boundary to the intranet. Company staff are better informed and make better judgements whether to publish their material internally or publicly and can do so more easily both politically and technically, but the fact that sometimes/often the authoritative statement by a colleague is on a public blog, means that intranet search needs to pass through the firewall and "join" intranet and internet resources.

One very obvious example, illustrating the difference in intranet content can be discovered by examining tag clouds. If one were to compare my tag cloud with my internal delirious tag cloud, there are huge differences, I have no picture gallery inside Sun, none of my food, gardening and culture bookmarks are stored, internally I have a bunch of "How to" and "Do not do", repository links and applications home pages. (Let's face it a lot of the Technical documentation is on the internet now, and the secret R&D stuff I don't get to see anyway!) Also the clouds have very different shape, partly because I have over 1200 bookmarks on my public site and considerably less on internally. Tag clouds may also be another way of overcoming some of the four observations and corollaries

In order to "see" my true tag cloud, I need to "add" my private and public bookmark lists, which I organise using delirious & It needs a form of federation.

It would be hoped that tags might be part of the answer, but the different shape of the intranet, may make the development of disciminating tags very hard. My experience at the momement is that htis is true. I had given up, but I have been inspired to have another go.

Sites like Digg and where user generated content creates huge numbers of hyperlinks because of the number of users, will also distort the shape of the internet; as they become part of the core , the size of the "out degree" segment will become larger. Hyperlinks become the votes of web readers, not authors. Although its possible that since these sites are designed to be read by people, that there will be a more limited reference to them on other sites and they will remain part of the "In" segment. The XML feed services will however be referenced by many sites, and the linkroll gadgets mean that they are referenced..

So intranet search queries require a different approach to internet search, but is it getting closer or travelling in different directions.


Thursday Aug 24, 2006

Using to find Flickr pictures

Did you know that if you bookmark a flickr picture at, that it posts a thumbnail into the HTML page of the feed.


screenshot graphic


I originally wrote, if you check my delicious HTML page out now, you can see what I mean, but I decided to upload the screenshot. I have however checked the RSS feed using Mozilla, RSS Owl and Bloglines, and it would seem that the picture is not propagated inside the RSS feed.


Thursday Jun 01, 2006

Sun's SPOTs

There were a couple of demos of Sun's SPOT (Small Programmable Object Technology). This is seen as really important and is possibly the best piece of research from the labs hitting my sweet spot (no pun intended ) of proximity, wireless and database. They have their own web site The researchers are looking to innovate the platform to enable new applications and new developer productivity models. Check out the SPOT docs page, which also contains highlight arguments about the problems they're trying to solve. They also seem to have a view about collaboration, potentially missing from many of the demos and prototypes I have seen over the last two months at Sun & BT.


Thursday May 18, 2006

Food for Thought

Just looking back on Tuesday reminded me of various science fiction films and books. From William Gibson who in particular wrote about the Network, bio-feedback systems and AI, and obviously influenced the Wachowski Brother's Matrix, to James Cameron in Terminator 2: Judgement Day also spoke about how the network came alive. Marginally more benignly, Heinlen wrote about another computer that woke up in "The Moon is a harsh mistress". Some of John Ames presentation brought "Minority Report" to mind, at least the scene where John Anderton (Tom Cruise's character) is using the computers, but not anything to do with precognition.

It just goes to show, some of this is a lot closer than you might think.


Tuesday May 16, 2006

An iconoclast's glimpse into the future

The day ended with another glimpse into the future from Graham Whitehead of BT. He is a passionate and enthusiastic speaker and while I've neither met him before or heard him speak he is well known. He started with

"There will be more change in the IT in the next 10 years than in the previous 100"

and then asked who agreed. After the day we'd had we were all ready to do that, so he contradicted himself and stated that actually it'd only take eight years. (For the aficionados of the history of the Soviet Union, very reminiscent of the Five Year plans.)

Graham mentioned the revolutionary nature of BT's 21st Century Network (see also here...) as an enabler of the internet of things. He obviously thinks this is important, the Register reported a speech of his at the Irish Internet Association's Congress last year

"that the anarchic and hazardous nature of the public internet meant that companies were now constructing supervised private IP networks. These private networks would be able to handle the amount of traffic that would be generated when broadband was ubiquitous, phone networks were IP-based, and common household objects had their own IP addresses. "The internet is dead, or dying; it's full of viruses, worms and porn, you have to wear a kevlar suit before you go online," he said. "BT is creating a private network, which will be joined to other private networks, to which we will add voice over IP.""

He described the new paradigm (my word not his), as AORTA, "Always on Real Time", which goes down a storm in the healthcare industry.

One of the things he examined was around customer care, "there's never a queue in virtual reality", I bet he didn't try and file his income tax online in the last week last year. He suggests that machines don't say "Sod Off! I'm busy.", but actually they do. While he argued that people will deal with machines if they're better than people, (Bank ATMs are a proof point), if you want to speak to a person, that's it! People with machine generated scripts are as helpful as the machine, as Graham says, while no machine has ever passed the Turing test, many people have failed it. This is actually quite rude (what me worry!) and the problems with customer care call centre staff is often that their systems aren't good enough to help and the automated telephone menus absolutely infuriating. However the headline he offered that people will use machines if they're better than people is true enough and we'll probably get better at trusting and delegating stuff to them, which leaves the question as to what people are going to do?

He took this to a travel agency and asked them where he thought they'd be in 2015 and posed a not desperately unreal scenario that flight (well, aviation fuel) will be very expensive and rationed (anyone watch Dr. Who last week, the return of the airship) but that today's text/messenger kids will (or did he say may) be prepared to take virtual reality holidays. It wasn't at this point that he talked about a scent generator but it is an example of the extension of virtual reality to all five senses. (This is also a device that was not demonstrated during the day.) I'm really unsure about VR holidays, but some people go to Murder Mystery weekends today!

He was also quite keen on Robots, which put the demo of Sony's robot dog in a context in the Home 2.0 showcase. For him, robots extend the network's capability, and while the dog is not very useful, the programming packed into the AI is pretty amazing.





« July 2016