Tom Hume

dConstruct 2008: Tantek Celik: Social Network Portability

Why does every social site make you

1. re-enter your personal information?
2. re-add all your friends?
3. turn off notifications?
4. re-specify privacy preferences?
5. re-block negative people?

Keeping multiple sources of info (social networks etc.) up-to-date is a maintenance problem.

The goal should be giving users complete control over their data. Portable data + consistent URL = data syndicatability.

Had an interesting chat with Mr Thorpe after this one, who raised the point that in many cases, you don't want to share your identity between sites: you want to maintain many identities and quite specifically compartmentalise them.

Leaving aside whether we can enable people to express very instinctual and deep-seated behaviour explicitly... will they want to? Could we take the N different inconsistent versions of ourselves that we project in the real world (all the time lying to those we project to that yes, they're seeing the genuine article), XML-encode them and set them in concrete online?

dConstruct 2008: Joshua Porter, Leveraging Cognitive Bias in Social Design

Web design is now about psychology; web designers need to learn about it to create decent experiences.

"Bandwagon effect": people often do and believe things because many other people do and believe things.

Heuristics: rules of thumb which prevent us from needing to gather all information required to make a judgement. Sometimes these don't work all that well - this is cognitive bias. e.g. "Not Invented Here", prediction biases which lead to underestimates of time to work.

Seminal paper on this is "Judgement under Uncertainty: Heuristics and Biases (1974)"

Talks about a few biases:

1. Representation Bias.
2. Loss Aversion.
3. Ownership Bias.

These sometimes combine: in the sign-up problem, loss avertsion and ownership bias overvalue what they use by a factor of 3. Because of ownership and prediction biases, software manufacturers overvalue their products by about 3x - so there's a combined 9x dissonance between what they want to charge and what users want to pay.

Demonstrates Freshbooks, which emphasises number of users through a worldwide map showing them (encouraging bandwagon effect).

Q: isn't this evil?
A: It's more about business ethics.

(Sounds like a "guns don't kill people, people do" type argument).

See also Duncan's talk at XPDay last year. Alan Cooper also touched on this at Agile2008.

dConstruct 2008: Aleks Krotowksi: Playing the Web

The web industry and the games industry DO NOT MEET. Strange, given that web people don't talk about games.

Games are sticky. Some people die playing games; many people lose their lives playing them. Stickiness is important because of advertising. (Shows a Wordle of business plans from Seedcamp to demonstrate the important of advertising to business models).

What do game designers do to create this social web and stickiness?

Graphics? Games have great graphics, but some games (e.g. VIb Ribbon) have deliberately poor graphics and are still compelling. So it can't be graphics.

Story? Many games have strong stories. But traditionally the story is the last thing to be stuck onto a game system.

No, it's the stickiness of play. "The experience economy": a very boring term for the word "fun".

Games designers and developers use three systems to bring social elements into games:

1. Controlled systems: what designers explicitly build, deliberately giving reward and encouraging repeat play. The web does this - encouraging an investment of personal data in order to see more value. Also consider openness: creating spaces to play in, or sandboxes. Look at Grand Theft Auto world. Sometimes this backfires, in games which are too open and too large (e.g. Tomb Raider 3). But the web is enormously open, vast space. The challenge is to create a funnel that feels wide enough that you have freedom, whilst directing them towards an "ending goal"

2. Enabling systems: social phenomena emerge based on the design decisions made by developers. On the web we have community; in games there may have been some, but not a great deal until games met the web (with Everquest, WoW, etc.) (Not sure I agree with this: what about LambdaMOO etc?). Look at real-money transfers on ebay arising from virtual goods in online games: a community rallying around a virtual object with real social value. Game walkthroughs or FAQs might fit into this category. There's no need to create an economic model around your site to do this: look at PacManhattan, amillionpenguins.com, PerplexCity, or ludic visualisations.

3. Psychological systems: e.g. the relationship between avatar and reality (shows lovely slide of people photographed next to their avatars). But yet most of the web is personalised: MySpace, Facebook, but even before that pseudonyms/tags/avatar photos. Web developers see points-earning systems as a means of bringing gaming principles to the web. Look at PMOG as a game where you earn points for on-web behaviour (e.g. "don't use google for a week"). Game developers create beautifully efficient feedback systems to encourage repeat play, should they not engage with these types of things? Games developers and designers don't tend to use formal HCI, they tend to be instinctual by nature - and by and large do a good job of it, partially because games developers are making games for people like themselves. In contrast, the web industry tends to be applying their skils to create things for other audiences.
   

Why is there such little games representation at web events?

Ends with a call for a group hug between games and web industry, then questions.

Online games involving community in the form of MUDs/MOOs predated the web by quite a long time. How has the arrival of the web (as opposed to the improvements in graphics and UI) changed the way these communities operate

dConstruct 2008: Steven Johnson: The Urban Web

Wants to start with "a rousing speech about intestinal disease".

London, 1854 - mired in its own filth. A Victorian city with an Elizabethan health structure, creating a smelly environment regularly swept by cholera, particularly in the summer. The smell was seen to be the cause: the miasma theory "all smell was the disease".

A public watering hole in Broad St gets contaminated. August 28, 1854 the first victim dies, and following on 10% of the neighbourhood dies in the next 2 weeks.

The story: John Snow works out that cholera is caused by filthy water and creates a map visualising deaths and their relation to contaminated pumps. Snow was a local physician who saw the concentrated outbreak in his community as an opportunity to identify the source of water and thereby prove his theory. He produced a diagram showing location of deaths plotting on a street map - though this itself was nothing new. But he also plotted on the map the area within which local residents would walk to get water - i.e. showing who would be affected by this pump. This disproved the link between miasma (smell) and disease.

Another significant individual: the Reverend Henry Whitehead (local vicar, 25-26yo) was well known in the neighbourhood as a "connector" figure. The pump was well known for its water quality (!) so Whitehead set out to disprove the theory through interviews with local residents, in the process often uncovering data supporting a link back to the infected pump and identifying individuals who had left the area. Whitehead eventually discovered "patient zero".

Snow & Whitehead has access to archives of open data, created in the previous century by William Far (sp?) and in a standardised format allowing consumers to identify deaths from cholera by geography. The idea behind this was that third parties might do interesting things with this data: Snow & Whitehead's first mashup!

Snow's incredible intellectual skill was the ability to move between levels - individuals up to the water system of all London - and draw conclusions. His map was "a social network of dead people" - those united by their disease.

Cholera never returned to London after 1866.

So, to the geographic web. The initial web kicked off in part because of having a standardised means of locating pages: the URI. Stacks can be built of top of this information only because you know where it is. We're now starting to get standardised geographic formats for data online (e.g. Google/Yahoo mapping APIs).

We have local expertise (knowledgeable sources of local information, spread via self-publishing in blogs), open standards for information, and visualisation/mapping tools.

We should be able to filer queries and provide results by "what people near me are saying". Yet in real life things that are said near to us matter more than things said further awau.

Demos "outside.in GeoToolkit" - to help authors Geotag their content properly, then examine it.

Another product: Radar. Takes a location, shows you what's happening in 1000ft from you, your neightbourhood, your city...

Amongst startups doing location products, there's a disproportionate emphasis on finding restaurants, local businesses, etc. Whilst this is valuable, there's a lot more to geography in everyday life.

Outside.in seems to provide a twitter-like feeling of connection to a location.

Geoweb could provide "eyes on the street" (quotes Jane Jacobs, "The Death and Life of American Cities" - sure I remember Adam Greenfield mentioning this at LIFT or PICNIC last year). These eyes, and the intelligence behind them, are what make cities great.

- snow go ethno

- what geo formats?

After last years event, where I did a panel session discussing mobile user experience with Marek Pawlowski, I've been invited back to talk again at the Future of Mobile 2008.

I'll be drawing on the experiences we've had at FP of delivering mobile services across a whole range of technologies, and talking about to decide where to put in effort when building and launching your product.

The quality of other speakers is, I'll confess, a little intimidating: many faces that'll be familiar from the London mobile scene (plus a few folks I've not had the pleasure of meeting yet), ending with what I'm sure will be a heated panel fight^H^H^H^H^Hdiscussion chaired by Mike Butcher.

Slightly lower-key, it's Brighton Barcamp 3 next weekend (following closely on the heels of dConstruct 2008, which has been keeping Soph out of mischief for the last few months). Last year I was running on empty after travelling back from Black Rock City. This year I hope to actually stay awake long enough to do a talk... which looks like it'll either involve presenting our experiences adopting Scrum over the last year, or something about software development, martial arts and craftmanship (presuming I can OTA mind-meld with Mr Whiteland between now and then).

If you're coming and have any preference, do comment! From discussions with Ms Cottrell in the office and Mr Silver at the Tuttle club on Saturday, I think there's going to be a range of really interesting stuff going on...

I've been meaning to post up a little summary of the news on LocoMatrix. There's been quite a lot happening here over the last few months:

• We've been selected to receive a grant from the Technology Strategy Board's Creative Industries programme. It's a joint proposal with Brighton University, all themed around creating tools to assist with authoring games: a really important part of Loco. With a couple of radically different game formats already out there, we're keen to get some third parties using the platform and seeing how their perception of what they need differs from or matches up our thinking so far.
• Whilst we're on the topic of education, we're working with Portsmouth University on a collaborative project starting before the end of this year, the aim being to work with MSc students there on game authoring.
• A nice double-page spread appeared in London Lite, accompanying a story about location-based gaming which featured Loco quite prominently.
• New Scientist also ran a story on real-world gaming, in which Loco MD Richard Vahrman was interviewed.

We're living in an interesting time for anyone selling mobile applications and services. This particular slice of the telecomms value chain is getting a little shaken up, in ways which I believe will ultimately be good for both consumers and the industry.

Today in Europe we have 2 significant platforms on which to develop apps and services and 2 worthwhile means of distribution. The platforms are mobile web and J2ME, and the means of distribution are via portal or by text-message response (text in a keyword to a shortcode, get a link to the application back).

Yes, there are others; Flash, Python, BREW, widget runtimes and native applications all have their place - but none have the audience that J2ME and mobile web deliver. And distribution is possible via Bluetooth and sideloading of applications from PCs, but at Future Platforms we're not seeing this happen in great numbers (with the exception of GetJar, who snuck under my own personal radar and seem to be doing very well, judging from their figures).

I have a suspicion that this is about to change - and it pains me to join the chorus of folks bleating that the iPhone seems to be leading the way. When I first got hold of a (jailbroken) device last year, I was impressed by the third-party installer for applications, which at the time offered the best experience I'd seen for distributing mobile content. Apple have improved on it significantly with their own version: the iTunes App Store.

At a time when much of the mobile industry in Europe seems content to whine endlessly about the twin bugbears of device fragmentation and operator revenue shares, it's refreshing to see a new entrant arrive, do things differently, and publish numbers demonstrating success. We don't tend to see figures for application downloads from operator portals or third parties here in the UK, I assume probably because the numbers aren't good enough to impress. In contrast, Apple have announced 60 million downloads by August 2008. A few smaller developers are also revealing their figures.

Now there's a slight difficulty comparing numbers here because the App Store aggregates upwards of a million potential customers from across all the operators selling iPhones (one per territory right now, I believe); so we can't just compare these figures to, say, those for Orange UK (were Orange UK to release figures). But they still indicate there's a significant marketplace here; remember that with Apple taking 30% of revenue rather than the traditional 50% common in Europe, and providing all distribution (instead of app providers having to do their own or go via an aggregator) the margins are higher. And I'm honour-bound to point out that this is for a single device - so no porting costs (though higher costs for initial development, with Cocoa/C++ apps being more technically complex than J2ME - another post on that coming soon).

I'm frankly astonished that Apple have managed to enter an industry, partner with the incumbents and smoothly route around them in this way - though I'm not privy to commercial agreements between operators and Apple, which might provide a revenue stream back to them I suppose. It looks like classic "divide and conquer" to me, and leaves Apple owning the customer's mobile data experience (if not traffic), whilst their operator partners keep control over voice calls and text messaging.

If you work in mobile right now, you'll be as sick as me of everyone bleating about the importance of mobile user experience - as though it suddenly became an issue the day that Apple launched their Big Shiny. But obviously it plays a role. And Google seem to be following suit withe their Android "marketplace" (so much less materialistic than calling it a "store", doncha know) - though with the existing mobile industry customising Android to its hearts content, how watered down this is by the time it reaches consumers is a completely different matter.

My best-case outcome for Android: the mobile industry wakes up, sees Apple, craps itself and starts unashamedly lifting the best bits of the iPhone model in customising Android devices. It does a reasonable job, producing handsets 75% as good as iPhone and selling them well.

My worst-case outcome for Android: operators and handset vendors buy it, lobotomise it to fit into the way they've always done things, and it ends up in the same cupboard as SavaJe.

But what a fun time to be working in mobile applications (which, plug plug, is where my company sits): good commercial results, an industry being shaken up, and the largest handset vendor in the world (Nokia) moving in a similar direction with Ovi. After seeing a few presentations over the last year where accepted wisdom was that there's no future in applications (and knowing full well this isn't the case - I have some fantastic figures to post here when I get permission), it's good to see a sacred cow or two get slaughtered in Cupertino.

Is it just me, or is it slightly strange that:

1. Twitter is a business built on lifting an artificial constraint from another medium (140-character posts, suspiciously close to the 160-character limit we know and love in SMS);

2. Meanwhile, there's a vast amount of whingeing going on about how it's deeply unfair that Twitter is turning off mobile updates... all because it's constrained by a different aspect of SMS: cost-per-message-sent;

3. This latter aspect of SMS has probably let to a better ecosystem of profitable businesses (i.e. pretty well the whole mobile content industry from ringtones onwards, let along the direct person-to-person messaging revenues) than most Web 2.0 startups will ever hope to;
   

See also Fred Wilsons lovely post on constraints and why businesses ought to embrace them:

"I believe constraints are key to building great web apps. I am not sure about rules that are dictated by the market or government. But the reality of the place we are in is that we have to deal with them. And the best entrepreneurs will figure out how to play these rules to their advantage."

There's a very interesting presentation from Alan Cooper at Agile2008 here - I'd encourage you to go and read it. In the past (particularly when reading The Inmates Are Taking Over The Asylum) I've been *really annoyed* by Alan - in particular with his view that developers are inherently incapable of undertaking interaction design, in quotes like

"[Programmers] struggle with this idea of making computers behave more like humans, because they see humans as weak and imperfect computing devices."

But I'm finding a lot I like in this presentation; maybe I've misunderstood him in the past, maybe he's mellowed or changed his mind:

"While interaction designers are pretty good at inventing user interfaces, lots of programmers and product managers are good at that, too."

I need to read it a few times and mull it over before I can give an impression of the whole, but there's some bits which get my bulb percolating:

1. The approach around slide 37, breaking product development into 4 stages, with the middle two (design/engineering) as agile, others as not. This implies iteration and change occurs in defined periods, not all the way through product development;
2. Slides on cognitive bias (as a driver for observing users, not just asking them what they want) reminded me of Duncan Pierce's skit on the subject at last years XP Day;
3. And the engineering phase sounds eerily like a development equivalent to the sketching which Mr Buxton advocates:

"You are going to write it twice anyway, whether the discarded first one is in tiny parts or one big chunk, so you might as well make the first time count for the max.   Brooks says that we will do things twice. I say we should acknowledge this truth and maximize the first time for understanding, and maximize the second time for efficiency."

Quote of the day:

Like watching an old man curse and scream at the weather, watching webheads complain about the idiosyncrasies of mobile as if they were only that - mere irrelevant idiosyncrasies - can be a test of patience.

See also Eternal September 2.0