IN MY LAST BLOG ENTRY I described being at the BarCamp5 Spillover event held at the BCS London meeting rooms. On 9th October it will be my turn (with colleagues) to run an event in the same location: the MetaKnowledge Mash-up 2.0 gathering [But see NOTE below], organised by KIDMM and ISKO-UK and with the focus Making and Organising Knowledge in Communities.

(A word of explanation: KIDMM is a discussion community around issues of Knowledge, Information, Data and Metadata Management — its roots are in the British Computer Society — and ISKO-UK is the UK chapter of the International Society for Knowledge Organization. I co-ordinate the activities of KIDMM on a voluntary basis.)

Plastic poker chips being used to ‘rate’ the importance of topics generated on cards by group members, in a discussion about computers and climate change.

Plastic poker chips being used to ‘rate’ the importance of topics generated on cards by group members, in a discussion about computers and climate change.

The instructions for BarCamp said ‘Come with something you want to talk about,’ and I brought along with me a concern about the best way to conduct small group exercises on the afternoon of the 9th. People there will be seated in groups of up to eight around tables, an arrangement which affords a surface for a card-sorting exercise.

In the BarCamp discussion which I led, I explained how cardsorting works and how it can be used to understand how people group ‘things’, be they ideas, features of a product, whatever. In information architecture and product design enterprises, letting a representative group of users sort cards can help clarify what people expect of or hope from a product or service. There’s a variant called ‘divide the dollar’ which I read about through Uzanto’s Mindcanvas site, which offers these ‘Game-like Elicitation Methods’ (GEMs) on-line. In this, each participant is given a number of coins or chips and asked to distribute them across a set of topic cards to indicate in quite a subtle way the degree of value attached by that person to each.

What has this to do with a KIDMM/ISKO conference? Well, in considering ‘making and organising knowledge in communities’ we will be hearing case-studies from a number of organisations who are using electronic tools to enable long-term and long-distance discussion and collaboration within their communities, with the aim of accumulating and organising knowledge. Tools like email lists, bulletin boards, blogs, wikis, social networking sites, and online media repositories for documents, audio files or video — in addition of course to face to face gatherings and designed-for-print publications. Which of these approaches seem to hold out most advantages, and how might they be combined? The workshop session is when we can deliberate on this, and a cardsort is my proposed mechanism for this.

Hopefully, all participants will appreciate the opportunity this exercise will afford to sift through the approaches we’ve heard of through the day, and brainstorm a few more. And for the KIDMM community, this foray into ‘requirements elicitation’ has another benefit — with the assistance of a final-year software engineering student, Susan Payne of De Montfort University, we are embarking on our own venture into online community. In the next eight months, Susan will be writing a custom web application based on a CMS (content management system), designed to match the needs of knowledge communities such as KIDMM.

If you’re interested in learning more about cardsorting and other user requirement elicitation strategies, I have written up a five-page account of my BarCamp5 session — how I presented the topic and how people responded to it — which is available for download as a PDF file. [But see NOTE below]

Also available for download is my preconference paper: a literature and concept review on knowledge management, communities of practice and approaches to the evaluation of online community toolkits. [But see NOTE below]

NOTE: The domain has been kidnapped! Following KIDMM-hosted links will currently take to a Web site about baby clothes and things. Apologies while this gets sorted out. It is partly due to the incompetence of my ISP, Adept (should be Inept?) Hosting, who are also currently ignoring calls for support. [11 Oct 2008]


I’M NEW TO BARCAMPS, and my first one was yesterday — in fact, a spillover from the main London BarCamp event, which was booked solid. (Or was it, I wonder? This spillover event organised at the BCS London meeting rooms allegedly had over fifty people coming, but half were no-shows, and it was two hours before enough people assembled to actually do anything…)

For an explanation of what a BarCamp is supposed to be, try this Wikipedia page. BarCamping shares some roots with Open Space methodology: there is no agenda in advance. Everyone attending is encouraged to make a presentation or run a workshop, with spaces and timeslots allocated dynamically.

Does it work? My verdict on that was also ‘open’. There were three sessions which I got something out of (not including my own). The first one was convened by my friend and KIDMM-colleague, Cher Devey. She asked, ‘How do we explain to people the products that we [i.e. developers] create?’ Four sub-topics emerged: Training in IT applications; Documentation of them; Interface design; and (slightly deviating from the question, but relevant without a doubt), whether the application was actually designed to meet users’ needs and expectations.

Technology journalist Arjun Jassal described an Indian government project to put an Internet kiosk in every village. This is not meeting its goals, said Arjun, because these big unfamiliar machines which whirr and blink are intimidating, and villagers have mostly never used a keyboard before. In contrast, some non-government initiatives using mobile phones have taken off in rural India. As the expression has it, ‘What has that to do with the price of fish?’ Well, in Kerala fishermen who have landed a good catch phone in to several ports to see where they can get the best price for their catch (see here).

Arjun's Hindi-enabled phone (left), and Isriya's Thai-enabled one.

Arjun's Hindi-enabled phone (left), and Isriya's Thai-enabled one. Click to enlarge.

This raised the question of which languages are used for Asian SMS and how the keypad relates to that. Yes, said Arjun, of course people write SMS in their own local language, and that language is on the keypad. At which, he pulled his Hindi-enabled mobile from his pocket, and Isriya Paireepairit brought forth his Thai-enabled one. I took a photo of the phones side by side.

Both these languages have more letterforms than English, so while every English letter is listed, the Hindi and Thai keypad annotations take the form X-Z instead: a range. Since I’m more familiar with Thai, I can explain that the No. 2 key is used for five consonants from kaw to ngaw, the No 3 key for six consonants from jaw to yaw and so on. The star key evidently accesses the inline vowels, the zero key gives the superscript and subscript vowels, and the hash key gives the most common tone markers mai ek and mai to.

I find it interesting that this imposes a further requirement on the user: to know the order of their particular alphabet. Thus maw, raw and law are all accessed by multiple presses of key 7, but none of them are marked on the keytop.

GLOBALLY, COMPUTERS USE A LOT OF ENERGY; and given the dominant role of fossil fuels in electricity generation, computing is therefore responsible for a lot of the increase in carbon dioxide in the atmosphere. The ‘carbon footprint’ of computing is thought to equal to that of aviation (about 3% of global energy use); some put its burden as higher, and it is clearly increasing fast.

Targeting data centres and the desktop

Attention falls in particular on the impact of data centres, where thousands of machines are racked up to store data, host Web pages, process transactions etc. The density at which these machines are co-located means that for every watt spent on computation and data access, another watt is spent extracting heat, so they are sinners twice over. There has also been a tendency at data centres to ensure operation 24/7 by running the machines full time regardless of the computational load, by over-provisioning, and by adding a layer of uninteruptable power supplies to the infrastructure.

In March this year, at the Royal Society conference on ubiquitous computing, I heard an interesting presentation by Professor Andy Hopper of the University of Cambridge, on the subject of Computing for the Future of the Planet. More recently, Andy reprised this topic as the inaugural lecture of the UKCRC. I recommend visiting Andy’s site where you can find his presentation slides, and the paper he has submitted to the Royal Society for publication.

I found myself in agreement with Andy’s ideas about locating data centres close to sources of renewable energy and ‘moving the bits’ rather than moving electrical power long distance and suffering transmission losses. I also applaud the work his team is doing on virtualisation, moving jobs around the data centre so as to shut down as much of the system as possible when it isn’t needed. But I confess I baulked at his urging that the personal workstation, the PC as we have come to know it in the last quarter century, should be abolished in favour of network-centric storage and services accessed from ‘thin client’ machines.

Defra purges the desktop

I recently attended a meeting of the Carbon Footprint Working Group set up by the British Computer Society. CFWG harnesses the energies and expresses the concerns of the BCS’s Ethics Forum, the Data Centres Specialist Group and the Communications Management Association. Of particular note is the work being done on a voluntary code of conduct for data centre operators, and a project part-funded by the Carbon Trust to develop software to help data centre management model the energy use of their systems and play ‘what-if’ experiments to find ways of being more energy efficient.

However, CFWG’s concern doesn’t end with data centres. At that meeting, we heard an interesting pair of linked presentation by Bob Crooks of the British government’s Department of Environment, Food and Rural Affairs (Defra), and by Richard Lanyon Hogg of IBM UK which provides Defra with extensive IT support.

IBM helped Defra conduct an audit of the energy cost of its ICT systems, using meters and thermal imaging cameras. The results showed inefficencies in many unexpected locations. A frequent culprit was departmental print servers. The survey has led to many reforms, including the abolition of many desktop systems and their replacement by laptops: Richard showed off the Lenovo ThinkPad which is now his sole machine for portable, home office, desktop and hot-desk use.

A plea for the peripheral

Laptops are great things. The need for them to run on battery power has been a great driver in the direction of energy efficiency. I wrote most of this blog text on an Apple PowerBook G4 laptop, on buses and in hospital waiting rooms. But I wonder — what kind of compromises does one accept by trying to use a laptop for everything? Energy efficiency is a good thing; so is human efficiency.

There are three kinds of work I do that are more efficiently accomplished by being done on my ‘desktop’ machine (actually, my Apple Macintosh G5, the largest computer I’ve ever had, is a tower system that lives in a trolley beside my desk, not on it).

  • Media publishing work in Adobe Photoshop, Illustrator and InDesign. It is a great benefit to have as much display space as possible, for two reasons — to see the document being worked on, in both extent and detail; also to have rapid access to the horde of control palettes. Otherwise one wastes much time calling up and dismissing palettes, and scrolling around the document surface: not very efficient!
  • Video editing work. Here the actual products being edited do not need so much display area, but two 720 x 576 video previews need to be accomodated size by side, and editing efficiency is enhanced by seeing a big horizontal slice of the ‘timeline’ at reasonable magnification. Plus, video editing can require the simultaneous use of many peripherals: in my case, at least one large and fast external drive, and a FireWire link either to my camcorder or my Sony DSR-20P digital VTR. Desktop machines can support more simultaneous connections to peripherals.

Now, arguably these publishing applications are one for which even Andy Hopper would make an exception. Editing video on a thin client system is certainly out of the question. But of course, these are specialisms which it would be rare to find in an office environment.

However, consider my third scenario, which can hardly be unusual among knowledge workers:

  • Working across many windows — It’s not uncommon for me to be writing a paper or a contribution to an online discussion (or this blog), composing text in one window while referring to several Web sites, PDFs and emails, each in their own window. I arrange the windows so I can tell enough by the bits poking out which contains what. At the moment I find I have 15 applications running, which isn’t unusual for me; and 8 open windows plus 7 minimised to the Dock — again, not unusual for me. This behaviour, which I find efficient in terms of research and writing productivity, is supported by the ‘affordances’ of a 1680 x 1050 pixel Apple Cinema Display.
NEC-8201a laptop

NEC-8201a laptop, circa 1983 – 32k of storage and a comms link to my Mac

But, hey! I have at times to remind myself that in 1986 I was working on the 512 x 342 pixel monochrome display of a Mac Plus. Many dialogue boxes in modern applications are bigger than that!

And what about my first laptop? The NEC PC-8201a (see right) had just six lines of forty-column type display. That required something like an orator’s sense of what point you had reached in the argument evolving just to the north of your fingers.

Ironically, you don’t get much more space when composing text for WordPress…

As we may think

One thing I felt about the Defra/IBM response to the challenge of controlling energy consumption by IT, as expressed in the CFWG meeting, was that their choices about ‘End User Devices’ were effectively limited by what technology is currently on the market. What if we allowed ourselves the luxury of imagining the EUD of the future? Would it look like a Lenovo laptop? Would it look like a PDA?

I hope it would be more modular than a laptop, for the sake of the environment. On my bookshelf, for I can’t bring myself to get rid of it, is a nice white Apple G3 iBook whose motherboard got fried a few months ago. To repair it would be expensive. Yet destined for the grave along with it is a perfectly good XVGA colour LCD display, CD-writer etc. Seems a waste. (I guess I could prise out the RAM chips and put the hard disk in an external drive box.)

I’d like the EUD of the future to be a small core device that can be extended like crazy to suit the task and environment at hand. Maybe it would be about the form factor of the Asus Eee PC (see Flash presentation), with small tolerable keyboard and trackpad, wireless networking and Ethernet, a daylight/backlit energy-saving screen somewhat like in the One Laptop Per Child XO machine, and about 16 Gb of flash memory. I envisage a clip-on base that provides bulk storage, mains power, a secondary battery or fuel cell and more expansion ports. For efficient desktop use, a better keyboard and mouse could be attached.

But for me, the real breakthrough would be how my envisaged device would work with external displays. Some models of Apple PowerBooks already show the way in their ability to hook up to large external displays and run a fully interactive desktop and applications over two displays at once. Let’s explore this further. What about being able wirelessly to hook up to a number of displays, some of them forming part of the Desktop, some perhaps temporary repositories to which a document window could be copied for viewing and possibly some touch-screen interactivity?

In response to Andy Hopper, I guess what I am indicating is:

  • I want an environmentally responsible computer, or ‘End User Device’, and one the energy requirement of which scales according to how much work, and what kind of work, I am doing with it.
  • I don’t like the idea of entrusting my bulk storage to the Internet, and needing access to the Internet to do any serious work, which is what the ‘thin client’ model suggests to me.
  • I would like to look beyond the ‘one size [laptop] fits all’ approach that to me is implied by the choices Defra has made for its staff.

I guess I’m asking for a lot, but there’s nothing new there!


How is Defra tackling climate change? — page links to science notes.

Computing & Information Processing for the Future of the Planet — four-page op-ed paper by Conrad Taylor in PDF form, prepared for a meeting of the BCS Geospatial Specialist Group on 10 July 2008

On the radical energy-saving architecture of the XO laptop, see A Conversation with Mary Lou Jepson (hardware designer of the XO) in ACM Queue.