I thought it was about time I installed an OpenID wordpress plugin for my blog and give it a whirl. So I dumped it on my server and activated it and proceeded to login using myopenid.com account (it's easier to remember than my pip.verisignlabs.com one and I forget my livejournal password… So in doing so my blog redirects to myopenid.com and I login and fill in a persona. Hit share forever or whatever its called and wham.

Bugs. Something about not being about to insert into the database. nonces.
Tried the Options/OpenID and get nothing through the admin interface. Something to look at another day.

Enter Cardspace and the Firefox xmldap.org CardSpace identity selector. I install, restart Firefox and head to Microsofts own CardSpace sandbox only for the window to flash and not much else happen. Next I headed to Kim Cameron's Identity weblog and logged in successfully.

After I've installed IE7 I'll try the sandbox again.

From what I've seen, thus far I'm not impressed with any of these systems. A lot of work still to be done. For starters who wants to type http://<yourname>.myopenid.com/ into a login box over and over even without the schema? Gah! (Are you listening Jason?) Gimme an i-name any day(if only), and as for CardSpace, well I'm coming around to the simplicity and supposedly secure space it runs on, yet it in no way meets my expectations. More on that another day.

For now I'm happy that I had at least one success!

What using all these services has highlighted for me is just how User Experience is vital to my wanting to use(and build) these services.

On that note, go and sign up here, like I have and help out. There's been interesting discussions on the logo, which sucks at 16×16 and every other size.

My pre-requesites for an identity logo; must be clearly distinguishable in monochrome. Think Apple. Think Technorati.

An area I'm been working on recently has to do with storing my own personal attention data in a way that affords me ease of management through a centralised event history.

It works in a way that is much like Dave Winers concept of a river of news. The newest events appearing at the top of the list, an area I like to call "the event horizon." Recently I read Dave's interpretation on Mobile Rivers and it had me thinking. Thinking; "thats only partly how I want to do things."

Here's an excerpt from his example:

How mobile Rivers work

A scenario
Here's a scenario. Taking a trip. The plane lands, out comes the Blackberry. Check the voicemail, check email, then… What's next? For me it would be to check the news, find out if anything happened while I was in the air. Did a scandal break? Another terrorist attack? Any interesting Apple rumors?

Not how I want to do things because if infers seperation between my attention data (phone, email, news) and their alert interfaces. I want them all in one interface. Increasingly I've begun to curse my email reader, news reader, and other seperation. The sheer volume of information to wade through is becoming unmanagable. What with the 10's of mailing lists I'm now subscribed to in addition to the feeds I read in my news reader - jogging back and forth between email, mailing lists, news, SMS, missed calls is becoming a pain to manage. I want it all in one. The problem is how to manage it together.

Increasingly I find myself gravitating to my feed reader in a river of news fashion for consuming as much of my attention data as I can. This has me thinking about the interface a lot and what I really need in there. What information is important to me and how best to present it.

The more I think about the idea of an Event Horizon extending my news reader, the more I like it.

Here's how it works.

All of a users attention data is aggregated into a single event history feed. Recent email, news feeds, missed phone calls, identity requests for link contracts, calender appointments, todo events, bill payment alerts. All in a river of news format. Newest events at the top of the list in the event horizon.

Events could be made sticky by type based on filters, additionally event types could be filtered by type through tags. eg. only listing recent email. With personal meme engines (like Sam Rubys) integrated as types of data streams and given event horizon priority aswell. Not to mention other statistical information aggregation based upon ones personal attention data should the user want that.

I see something like this integrating as a central point with a users identity broker. A users event horizon stickying recent requests for identity link contracts. An interface for sharing certain private data using Identity Rights Agreements.

The beauty for me with this system is that users could access their news reader, emails, missed calls, identity information all through the one prioritised river of news or the filters. All on a mobile device if they so choose to.

Want to approve a purchase and trust your credit card to someone whether it be a retailer or online service? Simply tell them your digital identity, an example being an i-name;

=firstname.lastname

(if only XRI would have native browser support soon

Their identity broker then sends your identity broker a payment details request that appears in your event horizon allowing you to use either your mobile device or a terminal to approve/disprove your credit card details transaction. Creating a link contract in the process based upon an Identity Rights Agreement

Perhaps using your thumbprint on a device that your identity broker has approved for thumbprint authentication.
By clicking on differnet types of requests in your aggregator, you could be either diverted to the appropriate client for that type of event, whether that be your email client, news reader, calender or Identity Rights Agreement broker. Personally I'd like them all integrated in an Ajax style interface manner whereby I click reply to reply there and then in my river of news to an email thats just come in, or an SMS that has. Or to expand to read comments on a post and reply. I'd like to see the latter done using something like Atom's Threading Extension and the Atom Publishing Protocol to post those comments.

This all requires a way to send XML data securely across networks. On that topic, Jason Kolb has been discussing his future of the internet, a discussion I have been following with interest. He has the idea of using XMPP as the messaging protocol between identity servers. It will be interesting seeing how this can be done securely. i-names, XDI is doing something silimar as I understand it, and something I'd like to understand better.

When it comes to security, I can't help but think that data-mining will surely play a role here.

In conclusion, I want one interface for events, an interface to action those events important to me.

The Event Horizon.

I've been thinking about data distribution and the open API model services like Amazon S3 offer. The more I think about it, the more I see a need for an identity system with integrated online data storage. I call this an i-drive API spec. Something any hosting provider could implement.
Here's how I see it work; first a standard form of data storage API be developed with all the bells and whistles for users to secure their data and access though a standard protocol. Possibly different levels of API security spec too. Not everyone wants to store secure data and not every provider has the resources to implement high security. I'd like the option to be able to store my photo's/music/data on my own server for example. Other sensitive data on secure services. Anyway, users are then given the option of where they choose to store that data when using identity broker services. The Yadis identity service discovery system then shares that i-drive information with i-drive enabled apps. The apps can cache if the API lets it. (much like I believe XDI link contracts work). This way, anyone can store specific data anywhere that provides an i-drive service, even multiple services for the same data, eg. backup and redundancy. Some people might want that data all in one place. Me, I like diversity and choice.

As I understand it XDI will do some(all?) of this for i-names. A full blown i-broker implementation I'm guessing? More to read…

Ideally I'd like to be able to grab some open source i-drive code, dump it on my on own web host, tell my Yadis supported identity account where my i-drive is located and what it's to be used for and all the rest between my web apps and i-drive is taken care of for me. Additionally I could then implement my own open source OpenID(or i-names) broker with i-drive on my server then and start using it and encouraging sites like flickr to support it. It could all be built into my blogware. That'd be sweet. I could even have an i-drive service running on my operating system keeping sync (if my bandwidth costs weren't so high).

Using this service I could subscribe to all sorts of data that gets added to my i-drive and sync'd to specific i-drive devices (TV/Music Player). YouTube on my real tube.

Anyone know of efforts around other than XDI looking to do something like this?

Update: Another thought. Triggered while reading Mark Cuban. Make each and everyone of these i-drives BitTorrent capable.

Update: A bare bones API extendable to include data manipulation extensions for web apps to make use of. eg. attach advert clip to a podcast, resize an image, seek to a video or comment location inside media, etc. WebDAV?

Update: I found this great post by Christine when looking for another of hers I'd read previously. It covers the need for decentralized data access. Well worth a good read.

It's been great getting some feedback on my previous post with regards to this topic.

Thanks for the feedback guys.

Daniel raises some valid points I'd like to expand upon. I'm still thinking about Jasons. And have an uneasiness about some of them

Anyway, re-reading part of i-tags, I understand that the current i-tag specification simply points to a predicate of vocabularies describing the content tagged.

I'm not very well versed on these things however what I had hoped could be achieved is the ability to include in an i-tag, an indentifier as to where an embedded objects metadata can be found. I envision this being done through say a microformat predicate class.

Here an example of me tagging myself with my hCard as metadata in my about page:

<a rel="tag" class="i-tag" href="http://depressionisms.com/tblog/i-tags/craig-hCard/">Craig</a>

Heres an example i-tag as I see it, at that link above:

<div class="i-tag">
  <!-- i-tag header metadata describing the i-tag itself -->
  <div>
    <a class="id" href="http://depressionisms.com/tblog/i-tags/craig-hCard">I-Tag</a>
    <a class="subject" href="http://depressionisms.com/tblog/about/">About Craig</a>
    <a class="author" href="http://depresssionisms.com/tblog/about/">Author: Craig</a>
    <a class="publisher" href="http://depressionisms.com/tblog/">Cataga</a>
    <a class="verifier" href="https://verify.opinity.com/">I-Tag Verifier: Opinity</a>
    <a class="date" href="http://xri.net/($d*2006-01-12T12:13:14Z)">Tag Date: 2006/03/22 08:19</a>
  </div>
  <!– i-tag body metadata describing the subject identified above–>
  <div>
    <a class="predicate" href="http://microformats.org/wiki/rel-tag">Tags</a>
    <div>
      <a class="object" rel="tag" href="http://technorati.com/tag/person">person</a>
      <a class="object" rel="tag" href="http://depressionisms.com/tblog/category/i-tags/">i-tags</a>
    </div>
  </div>
  <div>
    <a class="predicate" href="http://microformats.org/wiki/rel-license">Licenses</a>
    <div>
      <a class="object" rel="license" href="http://creativecommons.org/licenses/nd/1.0/">Creative Commons NoDerivs 1.0</a>
    </div>
  </div>

<div>
    <a class="predicate" href="http://microformats.org/</a>
    <div>
      <a class="object" rel="microformat" href="http://neuraxon77.pip.verisignlabs.com/hCard/">hCard</a>
    </div>
  </div>

</div>

The example above being my hosted hCard stored on an OpenID provider. Restricting access to that data to trusted users or sites. The ability to mark attributes in that secure data as publicly publishable or not. Like email addresses.

Instead of displaying email addresses in full view for spiders, they're replaced with a placeholder for requesting those details, the placeholder being a Yadis Identity form discovering my OpenID or i-name provider contact form. The placeholder opening say an AJAX form whereby the users name, email address, and message requesting why they want access are then transmitted to my provider, the user never having to leave my site. If I then approve the request, they get an email saying so, either with the content they requested or the original link where they were looking for it. If the link, they head to the site and on entering their email and a returned password(for email user verification) in the request information area(this, along with their details could be included in the link?), they get that information displayed. If the user has their own Yadis Identity the password wouldn't be required. To get the content they wanted, users might be encouraged to get an identity! Identity viral marketing.

Ideally, what I'm looking for is one way - a tag or identifier - to describe what the content is about on a page or section of a page and who has permission to it.

For me personally and the project I'm working on, I have the need to be able to tag an object embedded anywhere in a page, and use that tag to provide additional information on the object like licensing and privacy restrictions. Whether that be a users name, an image, part or the entire document, I want to be able to tag it on creation with the tag containing the metadata for it. Say one part of a document as copyright, another part under a creative commons license. I want to be able to distinguish those sections. An image on a page may be copyright but the text might not be.

There's a lot to learn of what's out there at the moment to do this. Right now for me it's as clear as mud. I probably should have been clearer in my previous post that I may have been thinking about i-tags outside its current scope there.

Thinking about it another way, I'm wondering if eRDF could be be used by defining my own standard rdf schema for tagging content or something. It's all very confusing at the moment because I think I'm trying to do something outside the scope of everything I'm finding. Or misunderstanding what I am. All a learning process.

Really appreciate the feedback on this though.