what's missing?
marccanter
free and open to all - open source
exchange photo albums
aggregate the aggregators - how do we do that?
get OpenSocial to go beyond "OpenWidgets"
a PeopleDNS would be nice - I believe Joel DeGan still has the domain
ever heard of ThreadsML? Danny - tell them about it.
whatever happened to the Attribute Exchange - is it just too difficult
to implement? What alternatives are there?
Josh Patterson
My group, floe.tv, is working on a media mixer. We ran into some
testers who really got high on the concept of the application being
"inherently installed" on the internet, and it being able to just
"auto discover" your data by an "openID" type id-system. Thats where
we started and where I coined the term "DHCP for data" (which loosely
fits), and then we began extrapolating a series of "what ifs" and
trying to fit existing comp sci abstractions onto the problem, as
there really isnt anything new under the sun.
Yeah, data discovery is a big deal right now, basically "run time"
discovery of data. I'm not a huge fan of trying to cache data into 3rd
party silos, but people are taking shots at that. So we said "what if
the internet really was one big machine? what if we looked at all
storage as one big disk? one big database? how do we handle those,
algorithm wise, in comp sci?" and we went from there, drawing up a
model, and building a reference abstraction stack (think like the OSI
Network Model):
http://cowbell.floe.tv/WRFS_11_20_2007.html
and then we started filling in the layers with different existing tech
(dont wanna reinvent that wheel). Our goal is to work with Faraday
(Chris, Ashley, Paul, etc) in constructing a reference implementation
that is open meets our goals in terms of the desired user experience.
Feel free to criticize, comment, discuss, etc the WRFS document, its
very much a sketch right now.
Josh Patterson
Chris Saad
On Nov 27, 12:01 pm, marccanter <marccan...@gmail.com> wrote:
> shared micro-content databases - sort of like Freebase or Twine - but
> free and open to all - open source
those databases? Semantic information about 'concepts'?
>
> exchange photo albums
I think this would need an open standard for 'photos' so we can put it
into the stack - or does someone have a better idea?
>
> aggregate the aggregators - how do we do that?
aggregators is not so necessary is it?
>
> get OpenSocial to go beyond "OpenWidgets"
DataPortability reference design we can compliment Google OpenWidget
>
> a PeopleDNS would be nice - I believe Joel DeGan still has the domain
name?
>
> ever heard of ThreadsML? Danny - tell them about it.
>
> whatever happened to the Attribute Exchange - is it just too difficult
> to implement? What alternatives are there?
Brian Suda
> > exchange photo albums
>
> I think this would need an open standard for 'photos' so we can put it
> into the stack - or does someone have a better idea?
--- i think we can look at the way we are already "consuming" photos,
via RSS feeds. I subscribe to friend's flickr photos via RSS. RSS also
has an enclosure element, so it is possible to move binary files
around with some basic metadata attached in the RSS tags, description,
title. The other information can be passed with extentions, like
GeoRSS for lat/lon.
I would equate an album to a feed, then there is apple´s failed
photocasting as well.
Also, as RSS Pull technology begins to feel the weight of bandwidth
issues when scaling, people are experimenting with PUSH. So rather
then me pulling every hour to see if a friend has updated, i
subscribe, then the system pushes/pings me with the new data. Some of
this is jabber, but could easily be the Atom Publishing Protocol.
If it was up to me, that is how i would implement any
"exchange"/notification of binary data.
-brian
--
brian suda
http://suda.co.uk
Danny Ayers
Great to see you in these parts!
On 27/11/2007, marccanter <marcc...@gmail.com> wrote:
>
> shared micro-content databases - sort of like Freebase or Twine - but
> free and open to all - open source
On the more general data-oriented side I'd point to the numerous RDF
stores (many of which are built on DBs like MySQL), e.g.
http://sites.wiwiss.fu-berlin.de/suhl/bizer/toolkits/
...on the more specific (human-readable) content-oriented side, I'd
point to Atom stores, i.e. components that support AtomPub in both
directions (not sure what there is in the way of implementations yet -
try Google :-)
Both approaches offer commodity systems, both are mostly in the open
source world. In both cases I suspect the main obstacle to widespread
deployment right now is developer-unfamiliarity with treating such
systems as building blocks, compared to say ad hoc PHP+MySQL or even
Rails. The big advantage IMHO is that such systems speak richer web
standards out of the box, rather than interconnection being a
piecemeal affair put together after core development.
<plug>
At this point I'm obliged to point out that it shouldn't be necessary
for people to host (download, install, scale, maintain...) their own
storage systems, as long as they have full control of their data,
Software as a Service is an option. The /reason/ I'm obliged to point
this out is so I can plug the Talis Platform, which in short is
comparable to Amazon S3, but much more Web-oriented (and free for
community projects). Check Ian's slides from Friday:
http://n2.talis.com/resources/swig-bristol-2007/
</plug>
> exchange photo albums
That's exactly the kind of scenario where I reckon use of a (the)
resource description framework comes in handy, rather than little
bitty extensions to something like RSS.
> aggregate the aggregators - how do we do that?
Hmm, that should be doable already, but given that not everybody has
Google scale, relevance filtering seems worth attention (!!).
> get OpenSocial to go beyond "OpenWidgets"
Right. That might be your job Marc :-)
> a PeopleDNS would be nice - I believe Joel DeGan still has the domain
Oh yeah - I'd forgotten about that.
There have been some recent semweb developments that might come in handy - e.g.
http://sindice.com
is a fast lookup (try your name as keyword).
more p2p-ish data piping should be useful to, e.g.
http://pingthesemanticweb.com/
> ever heard of ThreadsML? Danny - tell them about it.
"It was the dawn of time..." ;-)
Basically when RSS was going mega it was obvious that discussion
threading (blog comments, newsgroup threads even) wasn't properly
covered. Stuff was worked out, but I reckon it probably happened too
early for mass adoption.
But now there are a couple of approaches that look promising (heh,
Atom and RDF again):
Atom Threading Extensions : http://www.ietf.org/rfc/rfc4685.txt
- is like the minimum necessary
The SIOC (Semantically-Interlinked Online Communities) folks have been
going much richer:
http://sioc-project.org/
(I'd strongly recommend taking a look at what they've got together -
blog tool plugins etc)
though core stuff is straightforward, check this diagram in the spec:
http://www.w3.org/Submission/2007/SUBM-sioc-spec-20070612/#sec-overview
> whatever happened to the Attribute Exchange - is it just too difficult
> to implement? What alternatives are there?
Dunno, but a lot of the attribute-oriented stuff I've seen seems to
overlook the kind of connectivity the web has to offer, reimplementing
a small subset of RDF in a way that's less interesting and much
harder... (grumble, grumble)
At a bit of a tangent, although a lot of work's been done around e.g.
Jabber+FOAF, there still seems to be a long way to go before the IM
protocol stuff gets hooked up with the web in a (end-user) useful
fashion...
Cheers,
Danny.
--
Chris Messina
Chris
--
Chris Messina
Citizen-Participant &
Open Source Advocate-at-Large
Work: http://citizenagency.com
Blog: http://factoryjoe.com/blog
Cell: 412.225.1051
IM: factoryjoe
This email is: [ ] bloggable [X] ask first [ ] private
Terrell Russell
this data moving around once we're ready to move it. It's open, it's
extensible, and it's already in front of millions of people.
Terrell
On Nov 27, 10:07 pm, "Chris Messina" <chris.mess...@gmail.com> wrote:
> Speaking of Jabber, I'd add it to the stack listed on the homepage.
>
> Chris
>
>
>
>
>
>
> > Hi Marc,
>
> > Great to see you in these parts!
>
Chris Saad
Regarding the Push RSS thing - we have a solution for that - coming
soon
Chris
On Nov 27, 7:08 pm, "Brian Suda" <brian.s...@gmail.com> wrote:
> 2007/11/27, Chris Saad <chris.s...@gmail.com>:
Chris Saad
push solution I mentioned above - stay tuned.
On Nov 28, 1:07 pm, "Chris Messina" <chris.mess...@gmail.com> wrote:
> Speaking of Jabber, I'd add it to the stack listed on the homepage.
>
> Chris
>
>
>
>
>
>
> > Hi Marc,
>
> > Great to see you in these parts!
>
Mike Reynolds
distributed files and perhaps types of aggregated databases. If
Google is only giving us OpenWidget, then is there a way for people to
opt into an OpenConnectionsDB? While OpenSocial can tell things about
our friends, OpenConnectionsDB could be a meta-container that really
helps us open the social graph.
In terms of data portability, I'd want to be able to aggregate all of
my friends together (LinkedIn, Facebook, etc). This could make it
happen. We'd need strict rules for privacy/etc and I'm not sure it
would ever work, just thought I'd share.
Mike
Chris Messina
A principle for this group should be to invent as little as possible
and reuse as much as is useful.
Chris
Sent from my iPhone
Danny Ayers
>
> Any reason why XFN + OAuth isn't sufficient for this?
That's distributed files covered, publishing connection data, but
doesn't go very far as an aggregated database. For that I'd personally
use (er, am using) an RDF store - it can source data from XFN (and
other things, [1,2]), transparently merge, and allows querying, with
the results presentable however you like (e.g. as hCards [3]).
> A principle for this group should be to invent as little as possible
> and reuse as much as is useful.
Yup.
Cheers,
Danny.
[1] http://esw.w3.org/topic/CustomRdfDialects
[2] http://esw.w3.org/topic/ConverterToRdf
[3] http://dannyayers.com/2007/11/04/foaf-sample-data-and
Sent from my vacuum cleaner.
--
Chris Saad
that community to include in this DP group?
Also CTO of JanRain has expressed interest
>
> Chris
>
>
>
>
>
>
>
>
> > Hi Marc,
>
> > Great to see you in these parts!
>
>
> - Show quoted text -
Chris Messina
Chris
Mike Reynolds
interactions? I see that some blogs have Comments Feeds. Would it
make more sense for comments to be stored in XMPP?
Chris Messina
One way to think about it is to look at all those "subscribe to
comments by email" options on blogs... follow up comment notification
(and social network service messages) could actually be delivered via
Jabber, providing a great deal more control and semantic richness than
email.
Chris
Danny Ayers
>
> You could -- especially for real-time notification.
>
> One way to think about it is to look at all those "subscribe to
> comments by email" options on blogs... follow up comment notification
> (and social network service messages) could actually be delivered via
> Jabber, providing a great deal more control and semantic richness than
> email.
I like the sound of that - I bet there are quite a few other places
would be a nice fit too, Wiki RecentChanges springs to mind.
--