Re: call to arms

On 2010-03-30, Story Henry wrote:

> For the moment, writing your foaf profile is nearly just a write only 
> exercise. Some robots read it, but the result of doing so is not 
> immediately obvious, if at all. Once one can see what the benefit is - 
> and here I mean close to immediate feedback - such as getting access 
> to a campus, being able to chat with friends, [...]

So in essence what we need is a distributed version of Facebook: all of 
the user specific metadata stored as RDF documents on the Web and/or 
accessible via some well-defined protocol which can also do targeted 
pushes and/or pulls of triples. (I wonder whether Google's Wave stuff 
could help here; certainly some version of it could.) A UI layer on top 
of that to make it usable and beautiful. Trust, which in this case would 
probably be email authentication directly to your hosting node, with 
optional PGP crypto. Perhaps also aggregation of inter-host dataflows 
for efficiency.

The main problem I see with this sort of thing is that you'd have to pay 
for your subscription to one of the hosts (or host your own). Financing 
from ads wouldn't be easy in this sort of environment, because you could 
always roll your own open source implementation which excluded any ad 
from other people's feeds. OTOH, any particular host implementation 
could do whatever it wanted at the UI layer.

> The outcome is pretty easy to describe: you are no longer locked into 
> one provider.

Yes, and that is part of the funding problem. I mean, I've repeatedly 
asked Google to enable plain exports of data from their services. Even 
they don't go there -- lock-in is evidently part of their business 
model too.

> If you want to help people grok linked data just having them play with 
> data that is based on Linked Data won't help. After all: how would 
> they distinguish that from normal data?

One big way to make that distinction would be to build in location 
independence from the start. That is, deep support for multiple 
identifiers for the same data, means of propagating trust from one 
identifier to the next, and eventually resolution services which enable 
you to rename data to its newest URI. If this sort of thing took off, 
even the giants like Google and Facebook would be inclined to either 
implement their own resolution services or to subscribe to an existing 
one.

> Yes, the best way to do that is to start getting people to play with 
> social networks. They are bound to be interested. As they improve the 
> descriptions about themselves, they will then turn to dbpedia, 
> musicbrains, and all the other linked data sites out there.

Yes. In their most rudimentary forms, I think of such services as yet 
another app/box on your (distributed) Facebook page. In the next phase 
as integration of the basic data elements into the overall UI (i.e. when 
I state I like kittens, every box that has to do with likes actually 
links to some ontology and disambiguates the meaning using standard UI 
components and data sources). And come nirvana, PnP/DnD ad hoc 
functionality to build the boxes in the first place. ;)
-- 
Sampo Syreeni, aka decoy - decoy@iki.fi, http://decoy.iki.fi/front
+358-50-5756111, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2

Received on Tuesday, 30 March 2010 15:44:30 UTC