I want to write a bit about fragmentation because it seems to me it is one of the coming stories in web services these days. There is no reason to have so many identities scattered all over the web. In a sense, all these new services are counting people multiple times. You have accounts on all of them, and you might use all of them but they are not integrated even when they do the same thing.

Email is not a fragmented technology because of cross-server sending. but I can see that if email was invented today it would be fragmented because the companies with the servers would only want to send the email inside their company and would require people to have an account on their servers to be able to receive email.

Web. Can you imagine what would happen if web servers would only talk with clients of the companies that host them? Well, you can, just think AOL circa 10 years ago. The unifying thing for the web was the common protocol, http and the willingness of all to work with it.

This is a problem for all new services that we use. Think of twitter and buzz, think of gowalla and foursquare, think of facebook and myspace. I can't have a single identity and interact with all of these services based on that identity. I have both a twitter and a buzz account. And I don't want them. I'd like to be able to interact with buzz from my twitter account. This does not mean that I want to import my twitter feed into buzz. I don't want to have buzz at all. Strike that, my buzz should be a proxy to twitter. A non-identity service. When I post something on twitter, I post the exact same thing on buzz as well, thanks to Tweetdeck. It might seems like a unitary thing but the data lives on 2 servers. I might be able to go and edit the content on one service and not on the other and the illusion of "same" dissapears.

One such service that fragmented early and still is not integrated is IM. It's a crying shame that I have to have 6 IM accounts and I can't have just one. And there is no technical reason for this other than greed from the companies that have these services. They want the accounts, they want to be able to count these users because these users mean advertising money. Does it matter to yahoo that I only have a yahoo account to IM once in a blue moon? NO! They'll happily count me as a user, and they probably ask for 10$ for me when they negotiate their sale. Technically, there is no point why IM could not be integrated. It would require allowing adding to the contact list of emails of people that are not on your network. Then it would require a cross-service IM protocol to communicate with those users in chat sessions. As a matter of fact the technical solution is already here, jabber/xmpp would make it oh so simple. But there is no force in the world that could bring to the same table Microsoft, Yahoo, Goolge, AOL, Facebook and whoever else is out there making IM servers and end this nonsense.

An upcoming fragmented service is video-conferencing. It is similar to IM but far more complex. I remembered about this one when I heard Apple pitch their own iphone video call feature. Think about how many of these video call services there are. Skype, the nokia 3g one, now the Apple one. Add to this all the video chat options in the IMs. You can't video-call from an iPhone to a nokia phone unless you're using skype. walled-garden right there. Apple's video call works right now only between iphones. What a mess. And there are no prospects of this getting better. A funny thing about this is how skype is concerned about Apple shutting them down from the iphone to promote their own video-call feature.

Now think again about email. I can send an email to a hotmail account from my gmail account and I can receive email back. I can download all my email from gmail using imap and store it on my local linux server and can still see it when I switch from gmail to a local solution. Cross-server usage. Based on a common protocol and exchange rule. Email is full of those. MX, POP3, IMAP.

It is up to people who host services to understand that there is greater value in allowing "externals in" than in forcing externals to create an account with you. What I mean is that they should allow people who don't have an account on their system to be added as connections and be usable to people who already have an account with them. The problems are not technical but human/business. And business problems are oh so hard to solve. I believe by this point that there must be a business school in US somewhere that teaches MBAs that fragmentation is a good technique to make money; just because it is so prevalent in the modus operandi of todays companies.

Is there a solution out there?

Add: I keep finding more and more example of this cancer of fragmentation. Think the video coded problem of late. The solution to x.264 being proprietary IS NOT making a new coded that is free and splitting the market. The correct solution is making the x.264 coded free in the first place. How much money do the owners of x.264 think they'll make? For the sake of all 6 billion of us, make the damn thing free !!! DRM is another chapter. I keep buying books and music that is riddled with proprietary drm that is incompatible with other drm schemes and players. I pretty much buy useless bits that might be useless by next year. Even if these schemes survive, they fragmented me into user's of this company's drm and users of this other company's drm. For what end? To buy the same media more than 1 time?