Hosted lifebits meets infobus

Doug Purdy is thinking out loud about the principles, scenarios, architecture, and software necessary for what he calls infobus and what I have called hosted lifebits. I started to respond in comments on Doug’s blog, but of course that subverts what I declare to be a core principle, namely syndication.

There’s a crucial difference between a) committing my words to Doug’s blog, and b) committing my words to my own lifebits stream and then syndicating them to Doug’s blog. We don’t see it very clearly yet because we lack the mechanism for b).

I can kinda get the effect of syndication by referring to Doug’s blog entry from mine, and hoping that his blog engine will notice and acknowledge. But a truly syndication-oriented mechanism would imply that I publish in my own space, and then — in Doug’s space — actively subscribe back to myself. To explicitly comment on Doug’s entry, in other words, I don’t type words into his comment form. I create a subscription associated with my identity (as a conventional comment always is) that points back to my feed.

Let’s consider Doug’s point #4: “You determine if/when/how this data is accessed, the terms of use and the revocation of the license.” If I comment on Doug’s blog, I can hope for ex post facto control of my words, but whatever agreement may be (tacitly or explicitly) in place, the architecture doesn’t support that control. I may or may not be able to revise or extend my remarks. And Doug can certainly revise, extend, or delete — it’s his blog.

If I syndicate to Doug’s blog, there is still only a hope of ex post fact control, not a guarantee. But the architecture is at least aligned in my favor. The effort I invest in writing on Doug’s blog, or a bunch of other blogs, is preserved. I can archive, organize, and search all my stuff. I don’t need to depend on services Doug’s blog may or may not offer to find out who is reading and reacting to my stuff. And if I want to withdraw my comment, I just revoke the permission I gave Doug’s blog service to syndicate from mine.

Realistically, that revocation won’t erase my contribution to Doug’s blog. My words may have been quoted there, in other comments, and the mixing process dilutes control — which I argue is a feature, not a bug. But if the default is to syndicate by reference, rather than by value, the architecture favors the kind of control we want.

To clarify what I mean by favoring the right kind of control, let’s switch to a medical information scenario. Recently I had a dental xray. The image lives on the dentist’s hard drive. I want it to work differently. When I show up at the dentist’s office, I want to give the xray technician a token that grants her machine access to my lifebits store. The machine publishes the image to my store. I, in turn, agree to syndicate the image back to the dentist — maybe to copy, but maybe only to view.

One interesting benefit of this arrangement is that I’m decoupling dental service from image storage service. Maybe I’ll just turn around and reconnect them, because maybe I’d rather just let the dentist bundle those services. But when I interpolate my lifebits store into the pipeline, I guarantee portability to another dentist.

Another benefit is clarity of ownership and syndication rights. My lifebits store will have a management service where I declare, review, and adjust all of the syndication relationships between my lifebits streams and the services they participate in. And this management service can not only implement my ownership and syndication policies, it can announce them to the world. It can be the place where I say who gets to do what with my stuff. Some of those policy assertions will be private, but many will be public. Ultimately, again, there is no guarantee of ex post facto control. But if you violate my terms, it will be easier for me, or anyone, to determine that you have done so.


PS: Coincidentally, or maybe not, Doug was my guest on last week’s Innovators show. The topic was “Oslo”. But the context was our shared passion for figuring out how computers, information systems, and networks can more easily and more faithfully express the intentions of the people who own, operate, and inhabit them.

Posted in Uncategorized

7 thoughts on “Hosted lifebits meets infobus

  1. Another interesting article. I love this topic! I think you are on the right track in asserting that the lifebits service is more about providing clarity of ownership and syndication than trying to enforce it. True enforcement would lead you down an RIAA rathole that would just irritate people and provide no real value. Clear communication of your intent regarding your content will achieve the desired affect in the vast majority of situations.

    Mitch

  2. Interesting discussion. One nitpick: Not all information is the same. I don’t think the analogy between a comment made in the context of a discussion and medical records hold. We want to be able to revoke, withdraw etc. for the latter, you are damn right that the architecture should support this. Whether this is the right model for social activities such (like discussion threads), I am not sure.

  3. > Whether this is the right model for
    > social activities such (like discussion
    > threads), I am not sure.

    There is no singular right model, but it is /a/ right model. Example why: My conversation with Timo Hannay at Nature Publishing.

    http://blog.jonudell.net/2007/07/06/a-conversation-with-timo-hannay-about-the-scientific-web/

    The issue here: Scientists would collaborate better, and science would advance faster, if scientists could participate more freely in loosely-coupled online discussion: blogging, commenting on other blogs, etc.

    Many still don’t, and observers closer to that scene than me invariably report that it’s partly because they can’t take credit for contributions in that realm, or measure the influence that accrues to them from such contributions.

    The architecture I’m proposing would support measurement of contribution and influence, so that scientists could more easily justify contributing to, and influencing, a loosely-coupled network.

  4. ehud says, “…Interesting discussion. One nitpick: Not all information is the same. I don’t think the analogy between a comment made in the context of a discussion and medical records hold. …”
    I agree and this is very very important. Who is the owner of the information. What are the conditions surrounding the information has to be taken into account. It is not clear that you (Jon) are the owner of the X-ray (probably not the physical X-ray just the part that it is of your body – complex stuff). Consider Music you have bought on CD (you, of course, do not get files from your friends ;-) ). It has a copyright agreement associated with it that must be properly handled in any transfer.
    While I believe that the Twitter type of activity has some value I do not think that it is game-changing. There is a lot more to this area than has even begun to be thought about let alone executed in code.

  5. Leaving this comment in your blog’s care, partly for my future reference …

    1) Backtype does a nice job of automagically gathering copies of the comments I leave on various blogs: http://www.backtype.com/sebpaquet

    It features (local) revocation and fake/spam marking. It links to the original conversations; sadly but obviously, that link is one-way.

    2) I haven’t seen anyone do it, but one way to get pretty close to what you describe, by hand, would be to start a “comment blog” and dual-post every comment there, making sure to include reciprocal links. Somewhat tedious…

Leave a Reply to Mitch GarnaatCancel reply