That was a unique insight in modeling social networks that is starting to be adopted by other services as well, which in my opinion is a good thing. A conversation between Andrew Chen and Tim O’Reilly ocured on Twitter over just this issue:

Andrew Chen:

@timoreilly re:twitter, Networks are “friends only” or “strangers too” Friends only = less network effect. Twitter=friends+strangers though?

Tim O’Reilly:

@andrew_chen friends + strangers definitely the stronger model. web, twitter have this in spades.

Kevin Marks joins in on the conversation, noting:

its not so much friends vs strangers as not forcing reciprocity. “following” is a good way to express it

I agree with Marks’ sentiment. There are implications to how terms are used to describe social interactions online, and “following” is an apt term.

Terms that imply a bi-directional relationship, such as friend, family, or colleague, need to be confirmed by the individuals on both sides of the relation. However, one way relationships are equally useful, especially in “broadcast” scenarios that occur in public.

It is perfectly reasonable for two strangers to both follow and be followed by each other. Each arc remains one way, however, and the people remain strangers, despite the mutual interest in what each other is publishing.

I suspect some of the confusion stems from trying to compare Facebook with Twitter. That’s like comparing apples to oranges however, because Facebook is a private social network whereas Twitter is a public social network. That distinction is where the line is clearly drawn between friends and strangers.

The first case, interaction between people, is familiar to anyone using social networking sites like Facebook. When someone adds you as a friend, Facebook sends you a notification prompting you to confirm or deny the request. If you approve the request, that person will be allowed to view your profile.

In effect, a “contract” is established between the two people. Because Facebook knows the identity of each person, restrictions on that contract can be put in place. Privacy settings are a classic example of such a restriction. For example, you can limit who can view your contact information and phone numbers.

The second case, integration between systems, is less familiar but is getting increased attention as data portability becomes necessary. In this scenario, the authorization requirements typically involve only one person who has separate accounts on separate systems.

For example, I use Google Calendar to keep track of my schedule. When I accept event invitations on Facebook, it would be useful to have those events added to my calendar. I would like to allow Facebook to modify my calendar, but I don’t want to allow it access to my email (which is why I won’t give Facebook my Google password).

OAuth specifies a protocol for establishing system-to-system authorization. It allows you to approve access to your data on one system to another system acting on your behalf.

At the core, OAuth defines a mechanism for exchanging access tokens, which are issued during authorization. It is up to the system to implement and enforce different access rights, such as read-only vs. read-write privileges, or access to one set of data but not another.

In the above example, Facebook would be acting on my behalf, automatically adding events to my calendar so that I don’t have to add them manually. Google would recognize that the token used allows access to my calendar, but would deny any attempts to access my mail.

I had a brief conversation with Dan Mosedale, of the Mozilla Messaging crew, where I stated my stance that such a replacement is non-pragmatic. He suggested something interesting: most mail access today is not via IMAP, but via the web.

That is a perspective I hadn’t considered, but it is certainly true. Most people receive email through Microsoft, Yahoo, Google, or AOL. Through the use of Ajax techniques, much of this mail is being delivered over HTTP. It is by no means a stretch to imagine IMAP being replaced by an HTTP-based protocol, given an effort to standardize the data format transmitted in the requests and responses.

Ian Betteridge picks up on this trend, building on the following quote from an interview with Monica Guzman.

I’m convinced that newspapers need to rise up and take responsibility not just for the quality of the news, but for the quality of the conversation.

Betteridge makes a succinct observation that I find interesting as a computer engineer.

This is totally true, and actually reflects a significant change in what a news story … actually is. Rather than think of a story as a single finished article, think of it as part of a network of the story, it’s comments, and the comments and posts that it spawns across the web.

A story is no longer a single page: it is a network of pages.

Jeff Jarvis refers to the two stages in this transition content economy and the link economy. Old media companies must find a way to transition into the era where hyperlinks tie the story together. He gives the following imperatives.

1. All content must be transparent: open on the web with permanent links so it can receive links.
2. The recipient of links is the party responsible for monetizing the audience they bring.
3. Links are a key to efficiency.
4. There are opportunities to add value atop the link layer.

The self-publishing tools brought about by the Internet are incredibly empowering. Be it news and opinion on blogs to classifieds on craigslist, its never been easier for people to make themselves heard. As traditional publishers deal with these challenges, the entire system gets stronger and more accountable.

I initially wondered why Google saw fit to define a binary serialization format, when both XML and JSON are suitable and human-friendly. But it didn’t take long to see the benefits, considering the scale at which Google operates, and the sheer amount of data it wrangles.

I’ve done a cursory read through of the documentation, and while there are some features I’m not sold on, it looks solid as a whole. I’ll be experimenting further, in order to add it to my toolchain, and see where I can apply it to my projects.

This release also solves one outstanding mystery for me, which originated a couple weeks ago when Chris Messina tweeted:

Apparently XML gives Google indigestion. They convert all XML to binary but the crap it all over the web.

For example, I use http://jaredhanson.net to sign on to sites that support OpenID. Being a geek, I’m comfortable with using my URL to identify myself. However, the general public is accustomed to using their email addresses.

Toward that end, Michael Richardson of Vidoop has launched Emailtoid, a service which will allow people to sign on using email addresses rather than URLs. Chris Messina has the complete rundown.

Because I’m intrigued by the inner-workings of OpenID, I poked around a bit. Here’s what I found…

First of all, its important to realize that Emailtoid will only work when logging on to sites that support it. This means the site has to support OpenID, as well as fallback Emailtoid when and email address is used to sign on. That limits the total number of supporting sites to one, currently: erp.gobyairship.com, the example relying party. However, this is just a technology preview at this point, and you have to start somewhere.

Sign On

To sign on, simply enter your email address.

Now, you will wind up on the Emailtoid service. In order to verify that the email address you entered is actually yours, Emailtoid sends an access key to your email address.

Go to your inbox, where a new message should be waiting. Click the link in the message, or copy the access key into the page.

Once your email address has been verified, you will be logged in at the original site.

Associate an OpenID

If you already have an OpenID, you can associate that with your account at Emailtoid.

With that association in place, Emailtoid will redirect you to your OpenID provider when you sign on, where can enter your password as usual. This is more convenient that switching over to your inbox in order to get access keys.

Identifier Details

Under the hood, Emailtoid is generating OpenID identifiers for email addresses that do not already have such an identifier associated with them. When I first signed on to the relying party, my identifier was: http://emailtoid.net/i/d99a3e8e/756a31ac/

Because I’ve associated my OpenID with my Emailtoid account, that link simply redirects to http://jaredhanson.net/. An account with no such association currently gets a boilerplate placeholder.

Final Thoughts

As noted in the release announcement, this is not a final solution, but rather an attempt to start a discussion about best practices. I’ve got some ideas rolling around in my head. As I experiment, I’ll post about my discoveries. Stay tuned.

we’re going to have to face the problem of education in a world in which nearly all knowledge is available to everyone, instantly, all the time.

Take the address book of your cell phone as an example. Without it, I wouldn’t know the phone numbers of even my closest friends and family. In fact, there are times when asked for my own number that I have to consult my phone.

I often think that I would be completely ineffective in my career as a software engineer if it were not for the almighty Google search. Whenever I encounter a problem, a few carefully chosen keywords are often enough to find a solution.

Yet, each time my own knowledge is called into question, I reach the conclusion that this ultimately a good thing. The technology industry has essentially been crowdsourced, under an open source development model where anyone can contribute. This allows engineers to build better applications, and we’ve only seen the tip of the iceberg.

For example, there is advanced image processing research currently underway which can analyze batches of photos, and automatically pull out ones that are taken at the same location, based solely on the image in the picture. Furthermore, as additional photos are analyzed, they enter a feedback loop which improves the processing algorithms.

When online photo sharing sites such as Flickr are tapped, the algorithms themselves become crowdsourced, improving by the second with each uploaded photo. Eventually, this will lead to devices just like the phone of the future, where perception alone is enough to search all of the world’s information.

We’re used to thinking of media in terms of yesterday’s categories - largely focused on either inert information or passive entertainment.

But do they apply anymore?

Why is expanding the definition of media important? Because it’s only by fully understanding new categories of media that we can begin reinventing media strategies and business models.

Rather than wonder about business strategy, I’d like to look at the impact of new forms of media on another of Haque’s favorite topics: community.

In the past, if you wanted to voice your opinions to large audience, you needed column inches in a newspaper or a time slot on a radio or television station. These operations require extensive resources, thereby limiting the total number of people with access.

This limiting effect has built on itself over time, resulting in our present situation, where a handful of conglomerates control the dissemination of news, entertainment, and advertising to entire cultures. Those in power attempt to extract the greatest amount of “value” from the market, resulting in public airwaves are anything but.

Of course, this comes at a cost to the communities that are witnessing local broadcast stations disappear. It’s rare to turn on the television or radio and be presented with information relevant to daily life. Instead, programming is produced in a select few locations, often thousands of miles away, and broadcast to millions of people, all for the sake of efficiency.

However, the web and social media have had a democratizing effect that truly empowers the people. Instead of pushing a message to a “target audience,” the focus has been on building tools and systems that allow people to communicate. Now, anyone with can instantly publish an article or record a podcast, and be heard across the globe.

These new forms of media allow people to discover each other, often leading to organic, grass-roots communities formed around a common purpose or desire. Sometimes these communities are defined by geography, whereas others exist completely in virtual space.

The important point, though, is the ability for these communities to form where previously there was little to none. If I understand Haque’s thinking correctly, the fostering of community will lead naturally to the biggest economic gains.

I’ve been meaning to upgrade my server for some time, but never had the motivation until last week’s problems arose. I’m very happy with the new server. It has improved specs across the board, and costs me half as much as the previous one.

Everything appears to be in order from my perspective. Depending on where you are located, DNS updates may take a couple days to propagate. If you notice anything that is broken, please let me know.

In it’s current incarnation, the site serves as a web-based business card. I can point people to a single destination, which in turn links to the various ways to get in contact with me. One of those ways is, of course, instant messaging.

One of the handy features of IM is presence indicators, which show whether someone is available or away. I’ve brought those indicators to my identity site. If I’m online, a green dot will be displayed next to my IM address. If I’m away, a red dot will appear. The dot will remain gray when I am offline.