I was upset at Twitter for a simple reason: many (A-listers) claim that it’s fantastic to have many people answer your questions. I’ve tried: it’s not — at least for what seams to me to be a standard account: around hundred followers, and as many followees, one third of each being respectively stars and spamers, one half being physical friends. I barely ever have any relevant answer any to simple question, and I’ve tried most subject. I’ve also tried @-replying, links to blog posts, talking to perfect strangers: you get significant kudos — but insights are dimes a penny.  So far, the service is mostly a terrific link-source, better (complementary rather) then Hacker News. Therfore, a good service, disruptive for the press or stars mostly — like blogs were — but not a revolution on human interaction like Facebook is proving to be.

I was upset, until a friend of mine, who is helping VIPs to tweet (Nobody active yet, and I advised him against using any ghost Twitterer, or at least to reveal them if he does — this blog will keep its strict policy of not revealing any juicy news) asked me about hecklers. I tried to explain to him the rule about seeing @-replies (you only see the replies from people you follow to people that you follow) and it made me realize: Twitter managed to set up a fantastic tool for stars, a great filter for context-dependant elements. If you are well-known and someone comments your last tweet, then only people following him get to know what this is about: hecklers talk only loud enough to have their friends hear them. As a star, you can repeat what they say and make them famous if you need — but you don’t have your hard-earned audience hi-jacked by commenter like you kind-of have in blog comments. Asymmetric conversations are finally possible, and can happen in parallel, with little interference.

This structure (character limit excluded) does resolve the comment issue that I’ve mentioned earlier: a commenter cannot value easly his contribution between blogs with the same coherence that a blogger can. Because of that, that kind of rule (you only see replies from those your follow to someone you also follow) seems very important to me now; Facebook only partially does it, with comments on Fan pages (your friends have priority, but a couple of comments can always appear to encouarge you to go further). I’d love to know that @ev and @biz are aware of it and don’t kill it in another re-tweaking of their principles.

There is still one thing that would be great: a RT filter that would delete any non-commented RT of a message that you’ve seen already — but I do realize that this wouldn’t help partially attentive Twitterer notice important news. Being only almost exhaustive myself, I like to know that I can still skip the river from time to time.

I didn’t want to appear hostile to the idea of generativity, a notion that I found far more appealing to economics then what J. Zittrain seems to assume; hopefully, the most recent development of the Hulu/Boxee disagreement gives me the opportunity to call a quick-sanded fool when I see one.

Hulu is a fantastic free video-on-demand service that has all the features right: great quality & fast stream, all the right content, well structured; the distribution technology is a massive CDN and might benefit from some P2P support to be more affordable, but the ads are relevant enough to make the whole thing possibly profitable.  However, it has been fighting against people outside of the US accessing the service (nothing was done, in spite a clearly identified existing demand: that’s just lawyers’ unacceptable laziness) and is not preventing Boxee users to access it through a specific hardware (I guess because of the ad-filtering DVR larger screen); the latest technique seems to include a convoluted HTML encrypting.  Both battles seem to be loosing grounds.   The third axis of generativity that Hulu is killing with its artificial barriers are social aspects: if a hacker wants to develop a Hulu-based live video commenting system using Facebook Connect, well. . . Too bad.  Letting many test early adopter test such systems, suggest their own, even offer a sea of solutions would have been the profitable, generative way to go — presumably opening the way to far more relevant socially targeted ads.  I’m starting to see a pattern of executives killing good ideas early, and that worries me.

I’d love to say the similar things about the current debate in France on three-strike laws — but the talks are simply too depressing.

Most of you should be familiar with Prof. Zittrain, a legal scholar, cult-worthy for his wit and ability to resolve lurking conflicts by going meta.  He tries through a book and a blog to prevent the Internet from being reduced to closed platforms, controlled by central institutions who only let certain innovations flourish.  Legos, PCs, the Web, Open Source are example of generative technologies that have brought far more diverse ideas — while ready-to-assemble toys, AOL, Apple & Mac, iTunes & iPod/iPhone have been fantastic product, but with little room for radically new ideas from outside.  Please note his theories are not about good vs. bad as much as long vs. short term.

His latest trick is Herdict, a browser extension to let people around the world compare what sites are unavailable, and sort technical hiccups from political censorship — but the obvious initial object of his lust and his wrath is the iPhone and its open-source twin, the Google-inspired Android.  I just forwarded him my opinion on the latest event in that galaxy, and I thought you too might be interested in my say about that recent development of the War in our Pockets.

Tethering apps are applications on your OS-based mobile (iPhone, Android, BlackBerry, Palm Pre) that let you connect your laptop through your cellphone, and use your presumably unlimited mobile plan — abuse according to cellular-ISPs and the contracts that they have established with Google and Apple (who pulled all tethering apps from the iTunes App Store).  Google still controls Android Market with a licence that is binding enough to have to have them banned from the market.  As pointed out by Android Community, the openness of the platform allows users to get the tethering apps directly from the developers (and the ruckus around it will only make those apps better known).

However, I’d argue that tethering is hardly new: you might have some interesting way to make it work but large companies with a megalomaniac leader have a better track record at designing well such obvious features. Tethering is an known, identified service, and there is nothing generative per se in going against what must be the only legible paragraph in the user agreement.

Of course, Apple hasn’t done tethering specifically, but the service could, should exist: a little pressure on the cell-ISP would
help.  Being European, I’d even be in favor of a mandatory offer at a reasonable cost-based price — but anyone reasonable (and familiar with how saturated mobile Internet antennas could be) might agree that the added service should be billed on top.  I’d love to believe that executives at cell-ISP are, if not reasonable, interested in more revenue, and would agree with me — but the lack of offer leaves me worried.

Having two-tiered price helps the ISP to lower its monthly rate for the tethering-protected option — and more affordable, unlimited mobile internet for the masses is far more generative then free tethering against €5 or €15/month.  That’s mostly true because of the spectacular effects of combined congestion & price-sensitive demand: similar configuration have been studied for multiple-way roads (bridges, tunnel) where artificial price distinction ended up being beneficial for everyone.

Once intermediate solution appear, there might be a debate on what to consider tethering if they are ambiguous cases — but innovation would benefit from having a ISP-controlled app on both the iPhone and Android.

Many commenters have so much to say about how Google should have bought Twitter. The thing is, Twitter, or micro-blogging in general, is like macro-blogging, or RSS: it’s better opened, like a publicly available format for everyone to hack — as in, letting users refer to posts or users from a different service.  Imagine WordPress blog could not cite, comment, quotes posts from Blogger.  You might have needed a distinct account to comment until OpenID, but you still could include your blog URL in your signature.  The character limit is interesting, but the success of Twitter over Jaiku is simply a decision by Biz and Ev not to let Google’s Jaiku milk from their success.  Too bad, because, like for blog readers, micro-blogging clients have loads of features one should tweak.  Can you do what you want so far? With a rich variety of Twitter clients, you pretty much can so far — but don’t be fooled, while a Twitter monopoly makes sense, there is still another situation that we might reach soon, and from where you can’t get away: open-source and distributed micro-blogging.  More then paying through the nose for both an easy-to-replicate code base and a hard-to-get-away-from crowd, Google went for the longer term solution and both gives away hosting and opened source the code for any wanabee micro-blogging mogul.

I don’t understand why nobody noticed how important that move could have been: any disgrunted Twitter user can now get on the list of alternatives to Twitter. ‘Could,’ because like all things social, it will only makes sense if enough significant people do. ‘Could’ because unless someone finds anything missing from Twitter, why not use the free, nice, well-known, now-working solution? Any feature request (and I know plenty, from combined user-list and keyword- based filtering, to semantic understanding).

Google didn’t buy Twitter for the same reason that they haven’t bought the Web or Netscape: because neither it is for sale, nor it is cheap, nor it is necessary to make a decent business model out of it, nor should it be kept proprietary, nor Google’s interest is to have anything but vibrant competition in associated sectors.

Next up: my class, some news about my career, and maybe some explanation on how to define “associated sectors”.

Is Facebook a natural monopoly? In my previous post, I argue quite vehemently that, unless DiSo is implemented and massively used, it could become one — and I’d like to tamper that opinion with a distinction notion in competition economics: chalengeability.

The SNS market offers grounds a natural monopoly: everyone wants to use the same service as their friends, and because everyone is connected to everybody else (presumably by a chain of less then six steps) all want to be on the same SNS — the reality is more complex then that, but the latest opinions converge to say that, by considering the importance of weak links and availability of celebrities, you end up with having any decently put SNS a possible candidate for an imperial market share. And that’s the catch, a notion introduced by Microsoft: market share doesn’t necessarily mean control.  Some markets let any small, independent, innovative player with little means take control quite easily. The idea applies to SNS far more then OS, Office software or Flight Simulation, so it’s ironic Microsoft paid for its development, but it is a very relevant idea for SNS.  We saw many flips in recent history; LiveJournal was a step to Friendster; Friendster paved way to MySpace, whose issues made Facebook possible. And protests against Facebook’s features, so easily visible thanks to the NewsFeed could make a fifth generation service, like Twitter, gain traction overnight.  The question is: would that next SNS have to be better? i.e.: Does being a incumbent help ou or handicap you?

A great framework to start from is the one laid in that great article from RRW.  I particularly like that Bernard Lunn makes crystal clear the argument that I’ve been hammering for the past years:

In a social network, the value for existing users of a new user joining the network plateaus once users have most of their own contacts in that network.

but I would like to extend the analysis. Users are sensitive to change: the expression “same old” is often used in IT and caries well the meaning that reproducing existing winning features is easy.  Testing is hard, and Melissa Chang rightfully say that public complains against Facebook feaures is their main asset: they know what to change, in detail; it’s also their competitors main asset — so it’s both important, but doesn’t differientiate them, just like Nick Carr argued that IT was, some years ago.

The real catch is that people adopt not the service that is the best, but those that they expect will be used by their friends soon. The most impressive is relevant only through what you think of your friend’s eyes.  Because you anticipate they are blasé about all those old tech, you look forward to the newest, different kid — and behind early adopters, you have people who know the trade well and over-estimate how much their relatives know about technology.  With increase public scrutiny on SNS, and semi-decent coverage by mainstream outlet, one might expect that Facebook is not only diging it’s grave by publicly sharing it’s issues, but also by going mainstream and letting the next cool kid surf on their wave.  Making your service the center of attention is not changing who your users are, but how well they understand your business, and would value alternatives.

Many people are now trying to surf the wave of ‘open’ and that seems to annoy Chris Messina.  The idea is abused, no doubt about it — however, the argument that ‘Open’ is clearly defined, or fully beneficial isn’t true either.  Social networks are not a simple product: they are composed on many elements, and opening always focuses on a limited set of those features.  More generally, the proper angle to adress ‘opening’ is either from an CS point of view: what is feasible; or, rather from an economic point of view: what is made easier, cheaper to do?

Communication studies have described more interactions between platforms through three types of efforts: switching (changing provider) multi-homing (having several provider) and roaming (reaching someone on a different provider).  DataPortability encourages switching; most projects, especially OpenSocial, multi-homing; DiSo is the only effort that adressses most features of a social network and offers cheaper roaming. Facebook Connect still demands that you have an account and abide by Facebook TOS — and so should your friends who want to communicate with you.  I still have no way to share my Facebook-generated feeds with people who refuse to subscribe to the site.

I’m not saying the site doesn’t offer amazing features, on the contrary: they offer a great service — the only way to reach a monopoly with a free service.  I’m saying that, by making it easier to go from another SNS to Facebook (or any other monopoly) one is not helping struggling sites, but giving arguments to consider the largest site, the one likely to end up being the only one around, rather then the on with the features that match the niche needs or you and a few of your friends.  Same for multi-homing: language studies have repeatedy proven how by having everybody speak two languages: English and another, native one, you end up having all the conversation in English, and loose the useless one; similarly, if everyone is on Facebook and another SNS, say a professionally-oriented one, rapidly, the relevance of the other one fades — and, from apparently evenly shares, the market clicks into a monopoly.

Roaming — ie. befriending people outside of the walled-garden of a given SNS, without having to subscribe to each, and give both rights over your data — allow you to test your own version of any given open SNS, tweak it and innovate.  Without that feature, any open code for SNS will be useless: who’d subscribe to a SNS without friends? Who’d contribute code to a project with no rational users?

More importanly, being open, as in ‘letting do what you want’ neglects that my friends might want something else.  I don’t want an openness that let’s me in the dark around those disagreements: I want to know what I’m encouraged to do, I want to have socially-sharable authorisation rules.  And that openness, that Facebook has clearly identified, has nothing to do with other forms of Open.

I don’t want calls to Openness, I want clarity — and so far, I haven’t read anyone who was unambiguous about it.

Since I started this blog, I’ve been having issues with how to manage audience: most of what I want to say is reactions to someone else’s remarks; there is far less readers of this blog then people reading the comments thread of O’Reilly, Batelle or Carr’s blog — so I have always been tempted to comment and let this blog wait for my original inputs; but by leaving them custody, the consistency of my remarks might only appear to the empty over-lapping set or readers of all.  I could (and have) link-bait but it doesn’t work, anymore at least — and thank God for that.  There is no way for me to organise, value and use my comments the same way that blog can be used as a draft, a note pad.  For one, there is hardly any feed back to my remark — and whenever their is they are too hard to track.  Some blogs offer comment threads, but those a not convenient.

The inconsistency might come from a disonnance among the early or professional bloggers: all mention “conversation” but it’s never clear if it is done through comments or reply blog posts — most likely both, of course.  That however does allow them to neglect a detail: some do not have a blog well-known enough to come and comment with the strong identity needed to make constructive arguments.  I’m glad that people are judged by what they say, not where they are saying it — but this actually devalues commenters.

Companies like Disqus are trying to adress that by integrating all comments — but short of a compatible technologies, or even whort of showing a structured portrait of each commenter, neglecting who is behind corrections won’t turn out of habits.

I might get a larger audience by writing in French (even have a twin-blog for that): many more have reactions by translating ideas — but I still beleive not enough people make the effort to show to the world that some good internet research in done outside of the iLabs, in spite of meager means, if only to tell apart what results from there is true among American, Western or English-speaking users.  You can see a theme there: I’m not confortable with the room left for critics when publishing is made so easy; neither am I with an bling-folded enthusiasm about how opening a blog will make you famous, when attention is so concentrated.  There is a merit in sorting insights from intuition, and this work should be valued and made visible.

Don’t tell me that I’m wrong to be pickish and crave for validation from amazingly brillant people with an excessive number of followers: the largest audience of this blog by far (several orders of magnitude beyond what I’m used to if you want to ask) was not when I corrected one of the afore-mentionned pundit, but when he acknowledged it.

I make a parallel between posts and comments and papers and remarks made during a seminar: there are two levels, and the lower is not valued.

Spending some time in Oxford, I could access some papers that I’ve wanted to check for a long time: having though about the summary made the papers that referred to them, I gave those enough though-cycles to find significant concern — in peer-review articles, that is.  The net results is that I don’t know what to do about a (now dead) senior academic being a bit bold in one remark in a paper published in a second-rate paper (now disappeared) before I was born.  I’m really bothered by this general attitude that science has to be built, brick by brick, rather then carved with Okham’s razor, cut like a jewel into something where —to paraphrase St-Exupéry— not nothing more needs to be added, but nothing more need to be removed.

Do bloggers need to write posts, academics article — and I am arguing to reduce that? Yes. How will this could be as vivid, with less? By including the remarks and comments in the count, by encouraging people to specifically reply to what you do, and let third parties vote up the relevance of one comment; by letting readers decide if something is worth mentionning dynamically, not put an instituonal self-selected frontier.  Collaboration is not fostered if repect for an orinigal idea is at the price of not openly offering improvement.

I talked last night with friends about Oxford’s colleges — the basic life-unit here: those are not strictly academic discipline, most host under-grads, grads, fellows and seniors; all have an active cultural life and take part in frequent symbolic competition.  More importantly, there are the central area of scholarly discussion; as an homage, the set of scientists working on the same field without living in the same country, quoting each other, often meeting only at conference, is often labeled an ‘invisible college’ — and since the early academic Internet, we have countless amazing ‘virtual’ colleges.

Last night, the most knowledgeable of us mentioned the ancient city wall, and its still standing parts that one can see around town — he mentioned that a large chunk was still barring the way, in the middle of one of the most recent colleges, and it seemed a very inconvenient architectural feature.

Although most of my references are either from the invisible college around my adviser, or too early to be more hen pre-prints accessible on arXiv, but some are behind pay-walls too expensive, diverse and inconvenient for my university to trample them all.  I’ve been able to grab most of the references that I needed thanks to the far more comprehensive Library service here in Oxford — and I would like to thank the Web Science Research Initiative and the Oxford Internet Institute for inviting me if only for that — but to the many dwellers of the inspiring spires who, like ghosts, can cross those paywalls, I would like to point out how annoying those become when you realise that they are there.

Those limitations are not wrong because they prevent ‘anyone’ to read your work: they are wrong because they prevent your friends and critics, the few people who you know, or would like to hear from, the handful of those with whom you can actually talk shop, goof, be yourself and not just have to dumb down your hesitations. These walls are not and the edge of the academic city to protect it from hordes of ignorants—they are in your invisible college: they bar the corridor between your office and the invisible common room.

Compared to Facebook (that could improve this too) Twitter has little context: according to most client, either you enjoy going through everything someone posts or you don’t see any of their contribution coming your way; this is great with dedicated individuals or with friends close enough to make their inner questioning, restaurant reviews and issues with insurance claims relevant — but most of the time, I have to chose between having no more wide-eyed news from the Silicon Valley or going through @loic’s “Should I run or sort my desk?” moments. Of course it’s easy to over-look those — but I’d rather have a computer do it, rather then me. Loïc himself had to go through that, and decided an energetic cleaning. Hashtags and search are trying to resolve that, but mostly are good for events or breaking news — and don’t even let you opt out of an event, at least not intuitively: you can always ‘mute’ someone, but you wouldn’t know when they are back to their usual meanderings. I can’t prevent @chrismessina from sharing his interests for great bargains (I don’t live in the US, so I can’t use any of these), but I need his insights on the OpenStack.

Twitter can’t really give a pre-set list of contexts, because to some, family is work, or work is their passion; to some, music is their job, or their job is to find restaurants. However, people seem to be quite good at setting up a list of contexts that they step in: hashtags have been meant for that, and the aforementioned twitters and many others have rightfully added a #deals or a #weightloss to their blurbs. What I haven’t seen is a client able to take those away. It would then be up to followers to enforce a clear and systematic code from those seeking attention. This could be done either at the general level, or by followee; although with poster-specific hash-tags, these might not be very different.

Why negative filtering only? Because it would help greatly to balance the two positive way to get content: subscription and keyword search, and be symetric to the negative “-user:johndoe” option in search.

Facebook is going through an upheaval of their legal principles: a praise-worthy initiative, but by opening the sausage factory, they let us see how horrible the whole thing is. I already vehemently criticised the Principles; let me point at the contradiction between that rosy, fluffy document and the real lawyer thing: “We are the World, we love the children” and “You are allowed to do anything between consenting adults” turned into Nothing objectionable in any way, you comply with DMCA (even if you live in a country that hasn’t signed this madness). Let’s dive:

This text “Proposed Statement of Rights & Responsibilities” (please open next to this window: I’ll refer to it all the time) is actually a translation in English of the usual legalese that no one reads and that is generally as binding as what its own contradiction and common sense let go through a judge’s mind.

1. Privacy. Not sure why that document has to be separate. No big deal.

2. Sharing: I still hate that idea that we are ‘owner’ of content, and not its ‘authors’, but I guess this is because ‘content’ includes logs, involuntary traces of our activity.  This might explain the surprising “For content that is covered by intellectual property rights-”, although re-using our traces might need a licence too.

3. “You will not send or otherwise post unauthorized commercial communications” Authorized by whom? Isn’t there high principles in favour of ‘freedom of communication’ in the neighbouring document (that are about to be trampled)? The list includes impossible to enforce elements and broad considerations, and neglects the possibility of consensual communication. Symmetrically, the Principles ignore the need for Policing the exchange of information that is not pre-approved communication.

4. Registration: nothing to say, except that I like the: “You will not provide any false personal information on Facebook/You will keep your contact information accurate and up-to-date.” that might have some Orwellian overtones but actually challenges truth in identity; oh, and that 4.3 is surprisingly stern compared to the Principle 10 — a little bit like the poster saying: “Welcome to NY! I<3NY!” in the airport, next to the TSA agent that looks you right in the eye and announces you just won a cavity search.

5 to 8 are actually quite boilerplate. (Well, 8.2 should clarify that some links are not to be shared, and explain why and how — but once again: they do a decent job at bridging the contradictions between something legible and a court-enforceable text; although I’d always prefer the comic-book version.)

9 is fun. I’m not sure I understand 9.2.1 but it sounds like it’s in contradiction with the way all the application connecting Facebook with another site are infringing. 9.2.2 is ironic coming from Mr. Face “Black box” Book. 9.2.5 will be fun to implement for a SNS that started as a Fb application and decides to fly away: e.g. If things go sour with Fb, can Twitter erase the twits that came in from their Fb app? 9.2.7 would like to be more specific — he told me so.  The rest is understandable, but might need to lighten up the rhetoric of the principles, or at least add one element: ‘US laws apply; sorry y’all foreigners’.

10.3 must go. And ‘sponsored’ is not understood by most, so if you want to be clear and transparent, etc. use the word: “ad” and “paid by. . .”

11 sounds limiting the promise of all parties being equal. I guess I misunderstood the point of having that statement in the Principles.

14.2 needs to be there, because it’s the common ‘hosting’/safe harbor clause of the DMCA — but it needs to be made cristal clear, and probably be promoted to the Principles section: Facebook will comply with the local authorities.

14.3 CapsLock is annoying. If you want us to read something, make it short, or draw it.

As far as I can tell, the rest seems acceptable/common/obvious/boring.