I could write a Why I like Twitter article, but I’m not even sure what to put inside.

So I guess you know what is Twitter: a bunch of people who want the world to know they’re pooping. Yes, it’s useless. And don’t tell me it changed anything in Iran, because it didn’t. I’ve compiled all the stuff I hate about Twitter. Here they are, and it’s totally subjective.

First, Twitter is currently the most egocentric thing we’ve ever seen to date. Most people just want to be followed, the world has to know about their life. So followers matter, even more than the equivalent “friend-requests” on Myspace or Facebook. My point is that it’s mostly write-only: when someone follows more than 100 guys, you can be sure he/she won’t read it. You say “web 2.0” , I say “web 1.0“.

It’s useless. The retention rates (60% of Twitter users do not return after a month) are the consequences. I’m also pretty sure that the number of people using Twitter like it was originally meant for (human tweets, no feed, no massive RT) is remarkably low.

Twitter is centralized: there’s only one entry point to the network. A government could easily shut it down by filtering any request to twitter.com. And, more pragmatically, it could go down. And it went. While a decentralized system would simply prevent a small portion of users from accessing the whole network, a centralized one just crashes. We can find many decentralized systems out there, starting with email, which is a robust technology.

I want Twitter to die. And I want it to be resurrected with modern technologies based on rock, like email addresses. Yes, your email address could have been your identity if the network was decentralized. I’m expecting a lot from StatusNet, and even more from Google Wave (BTW, if you have an invite…), which seems to provide a more general approach than Twitter based on email look-alike XMPP . Twitter is a burden for the web. Since everybody use it, you have to use it, even if it’s terribly designed.

About design:

• Load balancing, have you heard of it?
• Why the fuck did you choose “pull” architecture instead of “push“? Simplicity‽
• Reply to tweets instead of people, that hard?
• Tweet edition, can’t do?
• Nobody use SMS to tweet. You should remove the 140-character limit.

It’s slowly replacing RSS. What a terrible news since RSS was also designed to be decentralized. Moreover, a RSS stream can embed full article, while Twitter is just indirections and redirections. Everything depends on another service. For instance, if bit.ly goes down, Twitter dies. Talking about shorteners: URLs used to have a meaning, a value, but they ruined it — stupid limit.

It is said to be a network of people and sharing. So why the fuck did brands and spammers invade it? Ok, I know why, no need to tell me. What I need to know is how many accounts are really useful for the network? Lay off any aborted account, any porn spammer, any auto-following brand, any “@naruto666 lol” user, any “Hey, did you know Disney bought Marvel?” retweeter. Seriously, how many? 1%? Less? How can investors trust such a fake network? Too many questions… The whole point is that they want to make money out of something which worth nothing. And one day, they’ll find out that Twitter advertisement is just stupid and unadapted.

The information is duplicated (or “retweeted”), because after all, Facebook, Twitter, Tumblr, blogs, whatever… are the same. They’re just a stream of (dis)information from people you trust (or not). I can’t stand it. Twitter is the opposite of what we’re trying to build with the Semantic Web. Yes, we want the web to come to us instantly, but does it provide access to news by topic, efficient and global search, or simply identity? No, the only thing you can do is search by keywords, not even semantic field: we need filters that Twitter doesn’t have (neither any other social network). Data used to be easily accessible, now it’s just mess and noise. Like IM, it’s not meant to be persistent. But it is. Therefore, we have to filter content ourselves, while it’s the machine’s job!  Today, we’re just using stupid ‘#’ because the service can’t provide proper indexing. Don’t forget that instant is not the opposite of tidied.

It’s fun to see reactions on Twitter, it’s like flame wars. You don’t know anything about what just happened but you still give your say. Hiding in the mass of instant reactions, it’s too easy to spead fake — or biaised — information. It’s also super easy to spread stupid uncritical point of views, which is a well-known “now”-ness effect. Like I read somewhere, people are ODing on digital courage.

Traveling Wilburys — Tweeter and the Monkey Man (1988):

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Anyway, I’m @WarrenSeine, and I’m a daily user of Twitter. Damn, it’s addictive.

This week’s biggest technical news is the announce of Ubiquity. It sounds like a great step for interoperability. If you have not read the news, take a look at the following video:

Ubiquity for Firefox from Aza Raskin on Vimeo.

To sum up, Ubiquity adds interactions to web pages. It takes a verb and possibly objects in natural language and actually does the action. I made a few commands related to Last.fm to try it.

As a concept program, it aims to define a new kind of HCI, maybe in order to design the future of Mozilla Firefox. You can try it. Today, it sounds more like a tool, an extension to avoid clicking everywhere and copy/pasting everything — I actually use it when linking to Wikipedia or Last.fm.

Let’s start from the beginning. Web browsing is a technique commonly used to get information, it does not directly provide a way to use information. As you would read a magazine, cut out a page that pointed out the address for a restaurant and open your map to get the directions, you often read a web page, copy the address, go to Google Map (or equivalent) and paste the address. We need to simplify these steps. They involve too much technical knowledge such as the well-known “copy/paste” and the “go to a website”. In a natural language, you would say « I want to get directions for this address ». There is no specific vocabulary in this sentence: everybody can understand what you meant.

Franz Ferdinand – What You Meant (2005):

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Ubiquity tries to understand natural language, but it is like an empty dictionary, we need to fill it up by giving it definitions, in particular, verbs. In the current Ubiquity implementation, we would ubiq the last sentence by writing « map this ». map is the verb and this a context-dependent object (what you have highlighted). Writing a command in Ubiquity means scripting the verb (which requires limited programming skills) to operate on the objects. In the previous example, this would simply open a tab in the browser which queries the map service using the highlighted address and display the resulting map. Intuitive and obvious. High level abstraction is the key to a wider public usage. From Boileau:

Ce qui se conçoit bien s’énonce clairement.

Et les mots pour le dire arrivent aisément.

Nevertheless, it does not exactly fit for the end-user at this time and may even never. Natural language is not a simple interface. If users do not want to type in an URL, they will not hit a key combination each time they want to communicate with the application. It may require a complete redesign of web browsers. It could be the only way to interact with your browser. Bye bye hot keys, location bars, “save as” actions… I am dreaming of a « What do you want to do? » operating system. One prompt to rule them all.

Nowadays, web browsers can do everything. News, mails, games, chat, music. You will not find any desktop application that you can not replace with a RIA. This is not a bad thing since it simplifies lots of  actions. However, we are reaching a point where the OS is useless, where everything depends on the browser you use. But the browser is not the OS! If Mozilla feels this way, then they should try releasing a standalone system based on their technologies. In my humble opinion, any operating system should natively provide a natural language interpreter, a kind of inter-process communication working with local and web applications. Some launchers such as Gnome Do are looking in the right direction, but still needs a real language analysis and a better context awareness.

Source: Mozilla Labs