While we’ve gained so much since the dial-up BBS days and pre-search-engine USENET days, we’ve really lost a lot with that collapse of context. We lose so much potential when we arbitrarily destroy the boundaries of every online venue/situation.

And there’s no TECHNICAL reason why we can’t bring back rich contextual boundaries.

But I have such incredible difficulty communicating the benefits of those boundaries, and what those boundaries even are, to business folk and especially to techies. Seriously, I’ve tried for years and just lost patience with it. And a big reason for that difficulty lies in the term “wall” and how it’s used here.

Talk about a loaded term. There’s the confusion of the concept “wall” as in “walled garden” with all other sorts of contextual walls. Try to discuss this, and techies immediately start thinking “walls are bad, they’re old-fashioned outmoded artificial Microsoft-style evil;” you can just watch that internal dialogue get going within that rigid frame and the ears are shut to other ideas. Similarly, business types start thinking in the narrow realm of “walls are for holding on to value and keeping it in our database/product/network.”

If I talk to someone far outside the tech biz though, it’s easy. So it goes.

Very interesting post and discussion. It ties in with my belief that not only are walls desirable, but that we need finer granularity – local walls. Issues of privacy come a poor second to hype and buzz in the web2.0 world, but without providing the infrastructure for private lives to be conducted on the web we’re sleepwalking into a post-celebrity culture of universal and unrelenting exposure.

In Defence of Walls

Before I do, I should say first that it’s absolutely true that erotica drives representation technologies, and it always has. Second, it’s absolutely true that people are far more forgiving of visual weirdness in games than other types of applications, and that’s partly due to the adrenaline but it is even more due to the relentless design discipline of game designers. They are far better than designers of applications that produce deliverable outputs, in part because simulations are their own ends, and if you have only one goal, you’re far more likely to reach it. Third, dating services predate any and all of these technologies, and they used to be called clan grandmothers or political advisors to rulers telling them who to marry their sons and daughters to. So given all that, it would be strange if dating services were not the next frontier, and if visual weirdness were not tolerated there too, as long as you could see the blemishes on your blind date.

So here’s the story. In 1996-7 I led a design team working on a massive overhaul to a major dating service, webpersonals. A major focus of this effort was to radically simplify glossary, vocabulary and internal taxonomy/ontology, so that promises to the users could be more easily understood and expressed as new features. And to simplify support calls. And to make it easy to translate the work into other languages. And so that there would be fewer misunderstandings of user expectations when the design evolved. The application was released in 1998 (it’s now called lavalife.com and the basic design hasn’t changed), and took off wildly. My name was on it as the chief designer or something like that, but I wasn’t involved in its evolution after the release. And perhaps I should have been, because a mistake I made at the database layer caused some users serious distress.

About a year and a half after the release, in mid 1999, I got a few irate (actually very angry and disturbed) calls from women who were absolutely livid about a problem with the service. That being, whereas they’d been told initially that their profiles in the three different silos (long term romance, dating, and casual encounters) would remain separate and that no one would know if they had profiles in the others, the company had broken that understanding and suddenly made it possible to see which other silos a user was registered in.

Now, it’s common both to have profiles in both long term life time romance seeking and short term overnight encounter silos, maybe for the simple reason that no one wants to wait forever to get laid, and have different standards for one night stands than for “the one”. People often/usually use different names for these different goals, and don’t really want to tell their overnight dates why they aren’t going to get married. Fine. I knew this, and it’s one of the main reasons for the three-silo design. So this misfeature the programmers had added, to see the other profiles, on the rather thin excuse that some users “wanted” it, had been a very major breach of privacy and a violation of the contracts and expectations under which people revealed things like, oh, their list of sexual fetishes. This isn’t necessarily what you want easily visible and linked for your potential romance partners. It opens the possibility of being pursued not for your values, but for your kinks, which was absolutely why we asked utterly different questions in each silo and kept them separate.

What was my mistake? In the database I had let them call the user’s name the “real” name, on the grounds this was short. I should have had them call it the “secret” name or “hidden” name or “dangerous” name or “what we are being paid not to use as a pivot in the database”. Any of those would have reduced the probability that the misfeature would have been added and reduced the angry calls I got. Which I did feel I deserved as it really was my job to prevent exactly this kind of error by the less ethical, less thoughtful and less talented people I knew would take over once I and the rest of my handpicked top flight design team was gone. That’s what software architects have to do – allow for the chaos and temptation and greed that will cause the system operators to do harm to the users and to compromise their privacy. It’s our job to make that very hard to do, and to seed the entire application at every level from database field names to command verbs to “about” pages, with a short list of clearly-defined terms that prevent such errors from even being considered. That’s what an “ontology” really is.

So, having told the story of a single wall that fell down to great anger and dismay, I should try to answer the questions:

I agree that “walls provide context and allow us to have some control over the distribution of our expressions. Walls should be appreciated, even if they are near impossible to construct.” Obviously that’s exactly what was at issue in the webperonals incident. And, I did appreciate the value of the walls and how difficult they were to construct, I just wasn’t careful enough in one detail to keep a particular wall up.

“If robots can run around grabbing the content”, it’s always a problem, even if the content combined is from affiliated or associated services that everyone knows cooperate in other ways. Even if no one outside is “caching the data”.

“If the walls come crashing down, what are we actually losing?” Privacy. The right to choose how we present ourselves on that vital first date or job interview. The ability to choose the clothing we wear for the purpose we pursue. And, more important than any of that, the right to disappear, and to cease to be discussed by people using a service that we no longer patronize. That’s an awful lot. We might, in some political contexts, also be losing human rights. Consider the plight of gay activists in Iran or Chinese human rights advocates or Russian anti-Putin journalists. Those who would do us harm gain more power every day to sift through the chaos and find grounds for suspicion and intimidation. The dissidents won’t have the edge forever.

Go look up the use of Canadian libel suits by bad guys to stop investigative journalism by good guys, and the way they have started to go after the online services that host the articles. The bad guys are learning.

So I agree completely that “walls provide context, context is critical for individuals to properly express themselves in a socially appropriate way. I fear that our loss of walls is resulting in a very confused public space with far more visibility than anyone can actually handle.” That is true.

That said, if the user is not themselves in charge of all this wall raising and falling, and not far better informed about the consequences of choices that should require more than a mouse click to authorize, then we’re creating a Panopticon.

And, if the users cannot organize in self-protective groups as factions and interest groups, and receive hearing, redress and before-the-fact notice of changes that affect their privacy and possibly public image, then relying on well-meaning gurus is futile. Regardless of how brilliant we are, we have bad days, and we have no moral right to substitute our judgement for the user’s consent. Outing real names is so noxious, and so liable to be wrong or misguided, that it needs handling on a level that looks more like Scooter Libby’s trial than like a chewing-out at the office. There have to be real consequences for these failures, and the only way to guarantee that is for users to band together and collectively bargain. And, when a legal or social problem arises, to have at least the right to have one’s friends handle the situation, before going public.

So, for instance, rather than assign someone a label that may stick to them for life, even one that they put on a pseudonym in the fetish room, it’s essential (not optional, not just a good idea) to have peers identified who can intervene, ask a question about intent or choice of words, and de-escalate any inter-user conflict or conflict between users and the system. The net is littered with dead and half-dead social networks now and that’s mostly because they ignored those long lists of friends when the time came to consider whether to throw someone off. Or because they arbitrarily changed the rules. Or because they created a “data jail” that users realized was a poor format in which to keep their increasingly useful data.

So yeah, walls are useful, but only if they’re my walls, or the walls put up by my friends and I voluntarily. Another thing games do well is let us choose our own teams or ladders. Even if all we want to do is grenade each other…

In meatspace i can move my stuff from one apartment to another without losing the company with the people i like. I can take a friend i know from one context (this pub i visit regularly) to another one (this gallery whose artists i adore). I can roam between contexts and interact with people no matter where they are coming from.

Take photo sharing communities, for example, it is not easy to roam between Flickr, Fotolog, and Zooomr. To move my “stuff” from one provider to another i loose basically all my mediated social relations. Sure there are always memories, but i am not in control of my own social relations. I have to go through a rather cumbersome process to interact in multiple communities. I can look at photos on all communities, but the simple feature of adding a comment forces me to register. Sharing photos with people from other communities forces me to upload photos twice.

Context is a great thing, if users are entitled to participate in the construction of walls or fences. But if the contexts are a result of profit-driven community providers, i would rather call it fragmentation and social data silo.

“Yet, most people don’t care that much” – Hmm ! that makes some sense.

I was referring to data to include every bit/word about/around a person, those words that is used to describe self, the data that represents friends list, the data that are the scraps/comments that is exchanged with friends.

You refined what I said to data that makes meaning most to a person, specifically the interaction/socialiazaton and which distinction is certainly needed & very useful.

- Rajan

MySpace et al have different sort of walls…egress and entrance is fairly easy, but the core profile stays in situ. Thats the rule of playing there. Upsets some, many don’t give a monkeys.

Is there a value to web walls? Sure – privacy and security – and the original concept of Paradise is a walled pleasure garden, the controlled environment within allows more delicate flowers to bloom (read this allegorically….).

And I think teh case for walls is getting stronger – I have blogged before that I think 2007 is the year that Privacy and Trust (sides of a coin) will become major issues in Social Media, mainly because they are being ridden over roughshod right now by every vested and invested interest you can think of.

If we go back to “meatspace”, in the many Dark Ages when barbarians were at the gates people didn’t wait for the legions, they threw up walls around the towns pdq.

Except for Sparta of course – they gave every male that become a citizen a spear and shield and told him he was part of the Bronze Wall until he died.

I suspect, with all the abuse of privacy and trust now happening in online social media, the idea of walled gardens will start to become much more attractive over the next few years.

Or, should we be like Sparta and require every netizen to carry their own armour and be part of the web’s bronze walls?

From my POV, there’s a problem in the conversation because there’s a value to walls that most of the anti-walled garden folks don’t recognize. They want to tear down all walls. They also believe that people want their data, that they see it as a property right. Yet, most people don’t care that much. They want to be able to hang out with their friends. If it means giving up data, so what, no big loss. The problem in the mobile space is that people can’t do what they want, they can’t gather with their friends. It’s not a data issue there either – it’s a freedom to socialize issue.

I am long regular reader of your blog and everytime I read your post I have great takeaways from your writing. But this time I am not sure it is the same this time and I would venture to say that this post is totally out of context.

‘Walled Garden’ is a term which is used more in relation to mobile technology world (where I have my day job ) rather than the internet world. There it refers to access to network which is gatekeeped by network operators. The operators are natural monopolies and thus have no incentive to innovate. Keeping their gates closed (restricing access to network) they don’t allow smaller/innovative players to participate in the ecosystem and thus stifling a lot of potential innovation which is why an uproar against Walled gardens. I have never observed/seen anyone use the walled garden in the context of the internet but mobile world everybody up in arms against it, infact there is not one single of the MoMo event that I co-ordinate in bangalore that goes where people don’tt crib about walled garden. It is most favourite past time now

Coming to the example that you have used about data locked in silos of big online corporates I am not sure If I would use the notion/construct of walls to describe them, kind of does not give a good way of thinking about them. I would rather think of them as property rights issue. What are the rights that I enjoy about my own data. ( use/ exchange etc) and for that I think as O Rielly said some time last year we need a Richard Stallman for data and given criticality of issue involved I am sure one will arise soon.

Cheers,
Rajan