More people on Portable Social Networks

Yesterday Pascal tipped me off about a brainstorm session by some notables about Portable Social Networks.

They’ve got the stuff on the microformats wiki page right now.

On BarcampCologne (which I’m unfortunately going to miss) there is going to be a lot of activity around this subject, seeing at least the first release of Noserub (German post) a social network aggregator.

Here in the Netherlands Robert had also whipped up an implementation of this concept and he’s planning to hold a meeting around this same subject in Amsterdam. After my talk with Willem, we had agreed that PICNIC could be a nice event in or around which to embed these kind of meetings/devhouses.

And there’s going to b a Social Network DevCamp in Richmond sometime after.

I wrote about this earlier and it’s nice to see it getting adopted all over the web.

Dutch Hyvers

On the other hand, the leading Dutch social network is hinting at opening up its information to the outside world (Dutch article). It’s one of the largest silos of presence, profile and relation information in the Netherlands and as such is incredibly valuable. For now it has been hindered by a complete lack of applications from the outside and arguably amateurish development from the inside.

Yme Bosma —the guy leading the effort— looks like a decent sort and I’m curious as to how this will work out. Things that would be nice:

1. Microformat (hCard, hReview, hResume and XFN) all content on the profile pages of Hyves members.
2. Allow people to set their presence in Hyves using an API.
3. Adopt an existing widget platform for integration in Hyves pages so the amount of widgets available will increase and developers can develop against a target.
4. A smart implementation of OpenID providership would be nice. Log onto third party sites in the Netherlands using your Hyves username.
5. Full application development inside the platform such as is possible on Facebook, judging from the current matureness level of Hyves, looks like it’s too far away to be realistic.

What’s next?

I think hResume and hReview are missing from the spec on the microformats wiki. I will add those today or tonight (I’m not completely sure what the procedure for editing is). Both should be optional but both could add much of the same information already contained on all those sites in a richer and opener fashion.

This does get my hands itching to implement some of this myself but it’s a bit of a conundrum what to implement. I know what the players already in the sphere should implement, but how do you get into the game?

(Posting has been a bit sparse lately. Summer is partly to blame for this. Expect lots of stuff from end of August onward and have a great summer.)

I just had a look at Simon Willison’s talk at Google about the implications of OpenID. Basically he covered a short FAQ list, answering some of the most common questions about OpenID. During the talk it was interesting to see how often he had to refer to the notion that OpenID doesn’t try to solve every problem, but that there is enough opportunity to build on OpenID to try and solve any other issues. This got me thinking about the problem of localization and how we could solve/aid this with OpenID

As a Dutch person I am very interested in localization as I would like to use a lot of service with my Dutch friends and family (mother). I would like to motivate many of my less tech-savvy Dutch friends to use things like Twitter, Jaiku and Flickr instead of local copycat versions. In many non-English speaking countries these services get copied, for example Zezz.nl which is a Twitter/Jaiku clone build by Alper for the Dutch market.

The problem with this copycat-approach is that you get localized silos of separated networks, with some people like me having to sign up to a localized version and an English version in order to be able to share with everyone I know. A good example is that I currently had to sign up for Facebook AND Hyves for social networking, as Hyves is far more popular in The Netherlands than Facebook. To make things worse: would I ever move to Brazil then I would also have to sign up for Google’s Orkut which is far more popular there!

So why not localize? First of all localization is a tedious job and secondly it is easily done wrong. To understand what goes wrong quite often, you should realize that in the English language we use the same terms over and over again on different sites for the same concepts. We use “save” for saving a file, “delete” for deleting it, and so we use other standard terms for other concepts. Calling the delete action “purge” and the save action “store” would still be understandable but defy expectations and possibly create confusion. This is exactly what happens in most localizations of systems, as there is no standardization of the terms used across sites.

So in comes OpenID, which when I log in to multiple sites already identifies me to every site, and is able to provide personal info using SRE. So why not extend this concept or create a new system that can enable an OpenID consumer to request what my localization should be, an possibly request translations directly from the OpenID provider. Obviously this should at first only work for basic terms like “Save”, “Delete”, “Open”, as full sentences might be a bit far fetched (for now).

I actually don’t think this is a problem though, as most people in foreign languages have some experience with the English language, but are just scared by a 100% English system. If for example Flickr would only localize their menu, I would expect that the user experience for many new Dutch Flickr users would become way easier. Funny enough Zooomr already has a localized version that is semi-translated.

Clearly it is not OpenID’s task to solve the problem of localization, but as Simon Willison stated: there is a lot of potential to build upon the OpenID idea. Combining this with the idea that OpenID providers will have to differentiate themselves by the services they offer leads me to conclude that they are all thinking of new techniques to implement.

An accessible OpenID interface to the localized verbs and nouns of the user would come in very handy here, especially for small sites that don’t want to do any localization. Making these libraries of terms user-generated would even make it a less tedious job and actually just a technical job. Does anyone know if this idea has been coined before?

A while back I read a post by Boris about how OpenID is not really easy to use (yet). He is completely right, and if Boris can’t use it, our moms definitely do not stand a chance.

I had a conversation about this with Ted Rheingold of Dogster, who was thinking of implementing OpenID for their users so people could user their Dogster logins to log in to affiliate third party sites.

A very important issue for him is that a lot of his users are not geeks and do not really want to get into the technological side of things. In most cases people will not be familiar yet with OpenID and you want to shield them from the complexities while still offering the benefits.

Do I have a OpenID?

When confronted with an OpenID login box, this is the first question that people —like Boris— are confronted with. What is this OpenID thing and do I have one or where do I get one?

Luckily more and more sites are offering hosted OpenID identities to their users. Wordpress.com does this for their blog owners and LiveJournal does this as well. Most people will probably prefer to use one of these hosted solutions offered by a third party site instead of hosting their OpenID themselves.

This way identities will be created until most people will have multiple OpenIDs. That still does not solve the problem of knowing that you have an OpenID and knowing what it is. I will propose a solution to this problem just after the next point.

URLs for what?

The whole concept that you use an URL to login —though I think it is quite elegant— will be difficult to explain to users, who already have trouble telling their login names and e-mail addresses apart. Adding another entity that you can use to login at sites, will only add to the confusion.

Signing in with e-mail addresses is firmly settled but it did take some time to get there. We may get to the same level with OpenID (and hopefully replace e-mail based logins altogether) some day but that is too distant currently. URLs are generally perceived as user unfriendly and normal users should not have to deal with them too much (yet).

Maybe i-names will be a solution to this sometime, but I don’t see it becoming mainstream any time soon.

Solve away the URLs

Taking both previous points together: most people will use a hosted OpenID solution and people do not want to type URLs, we can just abstract away the URLs completely.

When logging into an OpenID consuming site, that site can provide a selector with a couple of well known sites providing OpenIDs. This list of OpenID providers should be attuned to the target audience so they are familiar with these sites. With a fairly small list of providers, you can probably cover a large part of your user base.

I have made an example login box that works this way. It gives users the choice between several well known sites or the possibility to fill in your own OpenID. This is just a mockup which you can adjust in any way you like. You could expand the different login options or present them anyway you like. A site which already takes such an approach is the site for the band Rooney. You could also display the generated OpenID to the user at some point to get them accustomed to the OpenID they will be using.

Using that selector and a textbox users can pick a site they have an account on and fill in their username. The consuming site can then construct an OpenID URL from the given username and use that to log the user on. So taking my Wordpress.com username illustir it would construct my OpenID http://illustir.wordpress.com/ automatically (see the example).

What site are you taking me to?

The step where you leave the site you are logging into for another site can be a bit distressing for users. The approach that sites such as Wordpress.com take by having their own identity provider which looks and feels familiar dampens this transition a lot.

Large sites using OpenID should generally have their own provider so that they can control and attune the experience for logging users in.

Dogster’s use for OpenID

Suppose Dogster wanted their users to be able to log into third party sites using ther Dogster login credentials. This seems like it is exactly the kind of problem that OpenID is meant to solve. Especially in the case where the login is more a dependent syndication —a third party site affiliating with a bigger site— than that it is a general login (though nothing stops it from working that way as well).

So in the Dogster case they should start their own OpenID provider and OpenID enable all their accounts which are both relatively easy steps. Then, third party sites could use a Dogster login to log onto their site by simply becoming an OpenID consumer and by constructing the correct OpenID from the Dogster login.
The only problem with the Dogster case is that they use e-mail addresses as usernames and you would have to construct an URL with the e-mail in it. You probably would not want to spread e-mail addresses in that fashion.

This approach can be taken by any big site which wants to enable its users logging in elsewhere with the same credentials.

Update: I updated the example to be more clear and more educational about the actual OpenID that is being constructed.
Besides that a lot of people are missing the point. I am completely in favor of browser integration rich identity homepages and everything. Go out and build them already, but should/could/would are not going to help us right here right now. Given Livejournal+Wordpress+AOL almost everybody already has an OpenID but most of them do not know it yet. This —admittedly trivial example— is meant to fix that.

(This post got stuck somewhere in the queue. Tomorrow my longer post about Portable Social Networks.)

Willem Velthoven was scheduled to speak about anyMeta on the second day. I was already familiar with Mediamatic and their work and Willem had already contacted me about their social networking offering but his talk provided a lot of insight anyway.

Willem on the left (picture by Julian Bleeckr)

I did not know that Mediamatic is significantly in the social networking business. They seem to implement a great number of them on top of their standard anyMeta platform. Having done this a number of times, they began to wonder if they could abstract away the commonalities to reduce the duplication of effort. Willem talked about how his personal information is duplicated on a great number of websites and how this gets tiring.

AnyMeta is also the system as it has been used for the Reboot.dk website for before, during and after the event itself. It is a structured wiki which takes some getting used to but I think is quite rich in functionality. The only thing I am missing right now is a fine grained setting to receive notifications from the system.

Willem also talked about the API which any anyMeta site exposes at a standardized URL and which provides hooks to do pretty much anything you would want to with the site.

OpenID support both ways —by which I think he means both being an OpenID, accepting OpenID logon and being an OpenID provider— is supposed to be forthcoming.

In his talk Willem outlined the use cases he envisioned a networked social networks should accommodate and what the problems would be that come up with that. He was also very curious if other people had already started doing the same so that no effort would be duplicated.

I had registered myself on the Reboot site to host a conversation about the technical aspects of implementing social networks using OpenID. I mentioned that this could be a great follow-up to Willem’s talk to first talk about the need and the use cases for an open social networking system and then talk about the technical means we already have to our disposal to realise such.

See the following post on my talk.

Boris Veldhuijzen van Zanten (organizer of The Next Web conference) has gotten into a bit of a rumble with Plazes CEO Felix Petersen. Late Thursday afternoon Felix notified Boris that he couldn’t attend The Next Web conference because of a major bug with the new Plazes and an illness of his daughter. Can you imagine how amazed Boris was when he noticed that Plazes (Felix’s own product) placed him in Kopenhagen at the Reboot conference.

TechCrunch has taken up the story, providing some more details with the title “Plazes CEO Busted By His Own Product”. I have to agree with Boris for now because cancelling to go to Reboot is regretful, but lying about it is just a shame. Let’s see if Felix will respond with his side of the story.

Cristiano and myself wanted to add avatar support to Four Starters so that people could put a face to the writer of a post or comment. There already exist some solutions for this, but something more open might be nice.

(more…)

I came across MyHippocampus.com in a post on the Google Widget Toolkit newsgroups. This is not your ordinary webapp - so let’s have a closer look:

The basic idea of myhippocampus is mindmapping; a way to dump everything you’ve ever seen, read, learned or thought about into a website, so that you can search through it, visualize the evolution of your experiences, yadayada. When you see or read something new, you add the info to the site, and you’ll get back related thoughts and materials you’ve entered before, to give you some perspective.

That’s a tough pill to swallow though; “everything ever seen, read, learned or thought about” covers rather a lot of material, and it’s rather a lot of work to write it all down, so the ‘payoff’ - the ability to visualize it all, has to be large to even try. With this much info its also difficult to come up with useful visualizations. Still a closed beta, but fortunately there are screencasts to gawk at.

Remember the xkcd comic with the map of web communities? That same kind of interface is what MyHippocampus uses to map your life:

Nifty, and on a technical note, impressively done without flash. You can zoom in and out as if it’s google maps; the more you zoom in, the more details appear. The basics of usability are there as well: simple full-text search at the bottom, and an index ‘glossary’ of sorts as well.

However, I wasn’t convinced this might just work until I saw the ways you can visualize your experiences. For example, I currently don’t track the books I’ve read nor the movies I’ve seen because writing it all down and coming up with interesting visualizations of the data is too much hassle. However, this timeline feature, which can handle as many ‘islands’ (like a tag, really) as you like, seems useful:

The app also tracks where and when entries are made, comes with bookmarklets (think del.icio.us but with a world map visualization representing your tags and your bookmarks), and automatically links any content you focus on by giving you ‘neighbours’ in the dimension of time, location (in real life), location (on the map), and stuff you manually linked.

It’s certainly hip, but will it catch on? I don’t know yet. The biggest problem as I see it is difficulty in importing stuff you’ve already written down; some import wizards to grab your delicious bookmarks and e.g. amazon book lists would be a big help. There’s also no social aspect to speak of; it would be interesting to browse mindmaps of friends or people with similar tastes, for example.

You can see some screencasts here.

Mozy just released a very cool online backup solution for Mac OS X. For $4.95 a month you can get unlimited online backup, but their free version allows 2GB online backup which is enough for your basic documents. There are some cool features for Mac users including the backup of preferences, contacts, and e-mails. Files can be encrypted and if you have their payed unlimited account you can always order a backup of your files on DVD with next day delivery.